HRSA South East Queensland Group - Mystery Crystal Set chat
We suggest you go to the Yahoo Crystal Set Club there active crystal set chat going on there.
Re: Build the mystery Crystal Set
from
bjurin_y
23 Apr 2000
I just finished building the Mystery Crystal set. It's realy a good preformer. The selectivity is excellent. I can seperate all 5 local stations here with good volume. This one is definately a keeper.
Re: Build the mystery Crystal Set
from
sandhand (Dorado Hills, CA)
24 Apr 2000
I also tossed one together, and it performs well. I like it enough that I'm going to make a nice copy for the shelf. I lost a little volume over my single tuned loose coupler, but it is very selective and easily tuned. Kurt
'Mystery' set + Wierd DX
from
ge_whiz (Surrey, UK)
2 May 2000
Hi chaps, I know I'm as ever a fortnight behind everyone else, but I did a quick lash-up of Sable's Aussie 'Mystery' set and I have to agree, it's a real cracker. Picks up all my locals and is just what I want for my living room (while the wife enjoys her digital TV in the next room). Although my 120 foot competition aerial completely swamps it, it works fine with an 18 foot wire, albeit at the top of the house (wherein is my workshop).
While testing in the early evenings it has picked up Radio Sweden International (600kW, couldn't get it during the contest) and also Radio Netherlands International (similar power). This strikes me as exceptional, to say the least, but made me notice Norfdak's message when I got in this morning.
Also, Radio Luxembourg has been trying to get up and running again on 1440 kHz, but they have mainly been relaying stuff for other people (Swedes included). They may be pushing out 1200 kW, but also have a network of relays abroad. Could be you picked up one of them. It's getting a bit tricky to prove DX in some cases - I discovered recently that Radio Japan SW has a dedicated relay about three miles from my home! If you're looking for a good single-tuned crystal set, DO try the 'Mystery'.
Re: 'Mystery' set + Wierd DX
from
wetsteve (N.E. Alabama)
3 May 2000
Hello Ray; Nice set you've built! I would like to see a quality prefab coil marketed for this set to match a common value variable cap (365uF?). I suspect the dimensions may be critical for results, I.E. a ferrite rod version wouldn't have a sufficient progressive coupling characteristic. Interesting approach!!
Re: 'Mystery' set + Wierd DX
from
ge_whiz (Surrey, UK)
3 May 2000
Dear Wetsteve A 'prefab' coil! I'm shocked! Most of the 'gain' and joy emanating from my version of this set arose from the 'pain' of cramped fingers from winding the coil! Not to mention carrying home a 6-foot length of 68mm diameter drainpipe over my shoulder just so that I could cut 5 inches off the end for a former! ("Crystal Set Fanatic Wreaks Havoc In Town Centre".) I even had to wait until curious wives and 'helpful' little boys were out of the house before I commenced winding - you know it - "45 46 47..." "DADDY, what's that?" "DOH! 1 2 3..." And had to check the room was free of cats (they LOVE to chase the wire spools on the floor). Whatever next? You'll have us BUYING diodes from Radio Shack!
My mystery radio performs about as well as a simple loose coupled toilet paper roll radio. I could not get it to provide the mass of volume available with other designs I've used. The good news is that with one antenna setting I could tune across the whole AM band with a 17-440 pf varicap, not even using the highest capacitance. It has pretty good selectivity and volume. On the other antenna setting, I could only get the lower half of the band. 590 kHz was much louder, but there was a lot of 60 cycle static, not hum, really, but static. I get that sometimes with my other radios and I don't know what causes it. The .001 across the phones made no difference. To me, my experience is more evidence that it's the whole system that determines a radio's performance. It is a simple radio to build, but with a little more effort and not much more cost, a real performer can be had. I'd appreciate it if someone could answer the static question.
I dropped a few turns on the secondary of the Mystery set and it tunes so sharply that I found I was missing stations even using a velvet vernier. I get about half the stations I receive with the large coil loose coupler,but as you noted I can tune the whole dial with a 375 cap,get adequate listening volume, no taps to play with and it seems to eliminate the short wave stations that pop up on the loose coupler.
OK.I'll admit it.If I study crystal radios ten more years I might make a good plumber's helper.But I'm going to ask this question anyway cuz you guys know everything. If you substituted a transformer for the coil in the Mystery Radio,connecting the primary winding to the antenna circuit and the secondary winding to the ground circuit,then what would be the effect?Would the metal frame of the transformer cause problems?Or What?I know the instructions say to build it exactly as laid out.Just wondering,before I try it and waste time.....Thanks.Stan Radell
Well, since none of the "know everything experts" responded, let me take a stab at it. I hope I am not too far out of bounds to assume this, . but by "transformer" should I assume you mean an "audio transformer", because, you mention "frame" and because the Mystery Radio already has a pseudo RF transformer. I say pseudo because the tuned primary (antenna side) is not actually in a closed circuit except via the transformer's distributed self capacitance to the secondary, which is grounded. (Correct me if I am wrong, but at first glance, it seems to me this effectively uses that lossy, somewhat ramdom self capacitance as a antenna decoupling capacitance to improve selectivity, almost as if by accident?? Strange!). So, if you were to replace that pseudo RF transformer with an audio transformer, the self capacitance of the audio transformer would be so much higher as to essentially ground your signal, as you suspected, and I suspect it wouldn't work at all. An audio transformer would have far too much inductance and capacitance to work as an RF transformer. You probably know this anyway. Then, just looking at the schematic, (the self capacitance is not shown as a symbol) it's amazing that set works at all anyway, so much for analysis, huh? Regards, John
Good answer J.D. But the biggest losses would be in the core of the transformer due to the eddy currents. The inductance of an audio transformer would also be far too high to be of any use at BCB frequencies. Anyone else? John show some confidence. You are in the same league as any of the other "experts" here. This stuff isn't rocket science (but a degree in that discipline helps). hi : ).
Besides the transformer's distributed self capacitance to the secondary acting as an antenna decoupling capacitance, it also provides for two selectivity settings, depending on the relationship between the mutual inductance coupling, and the capacitance coupling. Changing the antenna connection from one end of the primary to the other reverses the mutual inductance with respect to the capactive couping, providing both a selective and a broad tuning mode. John
Hi John(& John-bta 50g); FWIW(& I am NOT an expert), your analysis of this unique ckt is pretty much identical to mine. So much for textbook engineering?? What say the formal experts here about this "distributed capacitance progressive coupling" notion? Has it been used elsewhere? Thanx; Steve
WetSteve, I could not think of any reason progressive coupling would help. What I suspect it will do is unevenly distribute the current in the coil, concentrating I*I*R losses at one end, slightly increasing overall losses. Also, additional losses are introduced by the capacitive coupling in the dielectric. I guess I believe it would be a better radio with a real (variable) capacitor for capacitive decoupling, and "purposeful" design, but that is only my arrogance talking. In reality, this is not a super high Q radio anyway, and those losses probably turn out to be insignificant. In all fairness, it works, more than can be said for some "purposeful" designs of mine. I probably would have looked at a Marconi Mag Det circuit and concluded Marconi to be an idiot. (How humbling). Philosophically speaking, I have made it a hobby to analyze "old time" circuits, looking for hidden whisdom, and found it occasionally. However, crystal radio has always been mostly an informal technology, and the practicioners, it appears to me, more often back into a design by accident, than on purpose, but they work. This circuit is outstanding as a monument to paradigm, in my view. A large number of old designs (and new too) are really not founded on what I think are sound, purposeful engineering design principles. I usually (*think*) I can see ways to improve them. Then who's to say my ideas are "sound", or would really work. I keep discovering flaws in my own "thinking" and I have come to expect that to continue. Discussing diverse points of view like this surely helps us all. And . "purposeful" design is really a matter of degree. Some of our radios are deceivingly complex. I have put all my guts into designing crystal radio circuits only to find that they fall short of the mark for some off the wall interaction like this. It is common. That only means I failed to get the job done, and learned one more detail. I suspect that the designer of this set did the same, only to find that it worked better than it "should". That's just my opinion, whats yours? Regards, John
Honestly John, this seems such an unlikely purposefully intended ckt that I could conclude that it may have been a serendipidous accident. It appears to me that it may have started as a bifilar wound tank ckt and was unintentionally grounded at the "wrong" point, whereupon it was observed to work better than when grounded "properly". However, the ability to control selectivity by choosing which end of the tank gets the antenna gives me pause. Was this intended, or observed and noted later? If intentional, the ckt expresses a unique elegance, and it's designer is a person possessing a remarkeably singular insight. Just my guess, based on the same sort of analysis you describe. Regards; Steve
If you read the original article about the 'Mystery' radio on Ray's website: http://www.clarion.org.au/crystalset/mystery.html it would seem that the design and performance came about by accident, hence the name. The author states this in the opening line. In truth, I have my doubts about the validity of the sensitive/selective options; the 'sensitive' setting (which, as Ray points out in his notes, is the opposite of that indicated in the original circuit diagram) provides better selectivity than standard top-of-tank designs, and a great deal more volume than any other design I have tried. The 'selective' setting, on the other hand, provides a lot less volume (though perfectly adequate, even on weaker locals), and I suspect that this is the old less-sensitivity-gives-less-overlap argument, which just 'feels' like 'better selectivity'. In my case, I tend to use the sensitive setting and get adequate separation of most of the locals (even above 1MHz, where some circuits show mediocre selectivity), but the 'selective' setting significantly shifts the frequency range towards the higher end, enabling me to receive my favourite MW station at 1548 MHz. All these observations, however, are definitely aerial-earth dependant, and probably my location is important, too. As others have observed, once again, the inclusion of a capacitor across the output does nothing for performance as far as I'm concerned, despite the author's assertions. Incidentally, I have e-mailed some pictures of my set to Ray, but he hasn't had time to post them yet. Keep checking the site for developments. (And the rest of you, give the circuit a try!).
Hello John, Here is how I think the "mystery set" works: 1. The antenna circuit is a series connection of "antenna with capacitance to ground" in series with the "parallel tuned circuit", in series with the "primary-to-secondary capacitance between the coils" which then connect to ground. This capacitance reactively reduces the antenna coupling (loading) to the tuned circuit in the same way that the Lyonodyne circuit does, except that the Lyonodydne circuit uses a variable cap at this point, not interwindinding cap as is done here. 2. The selectivity is changed by the antenna connection. When the antenna is connected to the end of the primary closest to grounded end of the secondary, the magnetic and capacitive coupling add for greater total coupling. When the antenna is connected to the end of the primary closest to the crystal-connected end of the secondary, the couplings oppose each other resulting in less total coupling, and therefor greater selectivity. I'm not sure whether the effective interwinding cap changes from one switch position to the other. Comments? Ben
Ben: much as I like your explanation of the 'Mystery' set, there is one problem; re. part 2 of your answer, Ray and I agree, (I think) that the aerial connections act in reverse to those shown in the original schematic and described by you, i.e. aerial to 'earthy' end gives greater selectivity, aerial to 'crystal' end gives greater sensitivity.
Ben & Steve & Philip, Steve & Philip, O yes! ...most likely an accident. & I think this set is a monument to the apparent success of the outrageous "design by accident" paradigm so typical of Xtal radio. Surely its principles were a "Mystery" to the originator. I think to build this set without recognition of that is missing something valuable. However, my guess is that there is a repeatability issue; compare your "eyeball" estimate of the self capacitance of Philip's gorgeous implementation with that of the student's almost scramble wound. But I guess they both work, the best probably depending on the RF environment? & Ben, I agree exactly, and would only point out one very subtle point that you probably intentionally left unsaid. The parallel LC tank combo you refer to as "tuned circuit" might also be viewed as the "parallel combo of 3 things: 1. antenna' own capacitance to dirt in series with interwinding decoupler cap, 2. tuner cap and 3. tuning inductance" where the tuning cap is both the antenna-interwinding group and tuning cap. (& left unsaid: earphone to operator cap). This subtle point would be a distinction that makes no difference except I view the so called tuning capacitor also as an additional decoupling capacitor to further unload the antenna. & FWIW, it is this defacto standard use of decoupling in the antenna circuit architecture that I think can work for single tuned sets like this, but I think gives away too much for multituned. Regards, John
Ben and Gentlemen, It looks like a gimmick to me. A gimmick is a capacitor, usually of small value, formed by twisting one insulated wire around another, or even just placing two wires in very close proximity. The gimmick is refered to as a capacitor but there is inductive coupling too, since even straight wires have inductance and when placed very close together the fields will combine. A very similar set is found in the MRL literature as MRL#39 (DP-6 and HB-17). He has the antenna and crystal hooked to the top of the detector coil . Phones hooked to crystal and bottom of det coil, no cap is used. The other winding has a varicap parallel and another section ganged Tuggle style from the bottom of the coil to ground. The coils are wound on the same form, one on top of the other. His det coil has 40 turns while the ground coil has 80. For variation, taps could be used on either winding to get different transform ratios and degree of capacitive coupling to match antenna to crystal and phones. A Navy Style (or other) loose coupling arrangement could be set up this way to get more variability. I have seen some old tube radio schematics that use a gimmick to lightly couple the radio to antenna. Blah, blah... Steve Holden
I guess it could be in the "gimmic" class, but gimmic caps I've seen are strategically orchestrated intentional components. The interesting feature of this set, that has us talking so much, is that I think it's gimmic is an accident, which un-gimmics it a bit.
Hi All, I am pleased to see the responses regarding the Mystery Crystal set design. If you would like to add your theory as to why it works so well to my www site then please email me your prepared text and I will add it to the site. http://www.clarion.org.au/crystalset/mystery.html I wish the original designer "Proton" could see the interest in his circuit 68 years after he published it Regards Ray
Hi Ray, I have been reading with great interest and belly laughs, all replys to the "Mystery Set". I think this helps lend credibility to a comment my Uncle Bob used to make to me as a youngster. Uncle Bob and my Father build custom or homebrew, if you like, crystal radios during the 1920's. My Father was an engineer and a scientist versed in the "theory of the day". My uncle was a tinkerer. He sat at a work bench, with his home made alcohol lamp heated soldering iron, a bunch of patch cable and BUILT crystal sets. My Father poured over the availble literature, of the day and designed at a drafting table. More often then I care to admit, my Uncle Bob's radios were more successful then my Fathers. Uncle Bob often said to me, "the man who has a dream, sweat on his browe and a hand full of hook-up wire often looks back at the man , sitting with his books and grins" Roger, K8CIX (I love your discovery of the Mystery set, not only does it work, its sets men dreaming)
Hi Roger - nice phillisophical touch to that - If you don't mind I'll print it and spread it around. Good one! Kev.
>>> Ray and I agree, (I think) that the aerial connections act in reverse to those shown in the original schematic and described by you, i.e. aerial to 'earthy' end gives greater selectivity, aerial to 'crystal' end gives greater sensitivity. >Yes that is the way mine works, The Earthy end connection to the aerial gived great selectivity. Greetings all; This is most interesting! I thought that this was the way it should work when I looked at the schem and not as it was labeled. My impression of it's mode of operation seems to have had some validity in practice, at least. Regards; Steve
My earlier analysis of the Mystery set (post #2132) has to be in error because it came out with the wrong results re the "S" and "B" antenna connections, and also because it is just wrong. It's simply amazing how such an apparently simple circuit can be so complex to analyze. Here's another try for the brass ring. I hope this one does better in standing the 'test of time'. (for validity). I've posted a set of schematics to help in understanding my new analysis of the Mystery set in a new album entitled "Australian Mystery Set". The Schematic set is entitled "Mystery Set Schematics" and must be referred to understand this analysis. There are two assumptions made in the analysis. They are that the distributed capacity between the two coil windings can be represented by one capacitor Cc, that is connected between the center of one winding to the center of the other. See Fig. B. The other is that the coupling between the primary and secondary windings is unity. Keep in mind that in transformers with unity coupling, the ratio of the voltage on any winding to any other is directly proportional to the number of turns on each winding. This also applies to a portion of one winding. (Just use the number of turns in that portion.) #1. Simplification and reduction of the circuit of the Mystery crystal set using the "Broad" non-earthy antenna connection: The physical circuit of the Mystery set is shown in Fig. 1 with the antenna connected to the non-earthy side of the primary. The detector and headphone load is shown as a loading resistor. The dots on the windings show the start of each winding, assuming that they are both wound in the same direction. Figures A through E show the inductive circuit through various changes as the reduction proceeds. Figure 2 shows the coupling capacitor Cc, between the two windings. It represents the distributed capacity between the primary and secondary windings. Cc is in series circuit with the antenna and ground. Fig. 3 shows it shifted up to the antenna and out of the way. No change in performance will result. The top and bottom leads of the secondary are connected (each 12.5 turns from the center), to the corresponding points on the primary (12.5 turns up and down from the center). This is shown in Fig. 4. Since the two top points have the same AC voltage on them, no current will flow through their connection and the circuit operation will be undisturbed. The same goes for the bottom connection. Figure 5 shows the resulting equivalent circuit. See Part 2 for the rest of this post.
Part 2 Since all portions of the winding are assumed to be unity-coupled to each other, performance will not change if the tuning capacitor C1 is connected as shown in Fig.6, as long as its value is changed appropriately. C1 is connected across 50 turns of the inductor. C2 is connected across 37.5 turns. The inductance of a unity coupled coil is directly proportional to the square of the number of turns. The number of turns across which C2 is connected is 3/4 of the number of turns turns across which C1 is connected, therefore, the inductance across which C2 is connected will be 9/16 the inductance across which C1 is connected. C2 must be 16/9 times the value of C1 for the circuit to work the same. The bottom portion of the coil in fig.6 can be eliminated since nothing is connected to it. The final result is the equivalent circuit shown in Fig. 7. Here we see a conventional crystal set circuit with the antenna-ground components connected directly across the full tank, with isolation from full antenna loading supplied by the capacitor Cc. The detector load is taped in at 2/3 of the tank voltage. That's it for the non-earthy "Broad" antenna connection. #2. Simplification and reduction of the circuit of the Mystery crystal set using the "Selective" earthy connection. Review figures 8 through 14 to see the simplification and reduction of the circuit. It proceeds in an manner similar to the one for the "Broad" connection. Now look at Fig. 14. The values of Cc and C2 are unchanged. The antenna-ground components and Cc are now connected across only 1/3 of the tank instead of the full tank. The detector load is still taped in at 2/3 of the tank voltage. That's it for the earthy "Selective" antenna connection. Conclusion: When using the "Broad" antenna connection, the antenna-ground components are connected through Cc across the full tank. This arrangement puts a certain amount of antenna resistive loading on the tank. The loading results in specific sensitivity and selectivity characteristics depending on the diode load, value of Cc, unloaded tank Q, L/C ratio and antenna characteristics. See Fig. 7. When using the "Selective" antenna connection, the antenna-ground components are connected through Cc across only 1/3 of the tank coil turns. This results in a reduction by 8/9 of the resistive loading of the antenna on the tank, compared to the loading in the "Broad" connection. See Fig. 14. The reduced loading increases the circuit Q, and hence selectivity. The ratio of loaded to unloaded Q is increased towards unity. This increases circuit loss thus reducing sensitivity. Ben
Thanks, Ben, for taking the time to post your explanation of this little jewel of a Mystery Radio. This helps explain why the fourth grade classroom version does not tune across the whole band, like my first Mystery Radio does. The kid's version has both windings close wound, the short one centered over the long one. It performs well, anyway.
Hi Ben, Since I had posted the same theory as your earlier analysis of the Mystery set, I was of course disappointed that you have decided it was in error, but I guess the march of progress can't be stopped. I wonder though, if a little of both theories might not be having an effect? I had done a spice simulation of the circuit, which indicted that the combination of capacitive coupling, combined with the magnetic coupling could create the broad and narrow tuning effect, although with the parameters I choose, the difference was not great. In my analysis, the Broad and Narrow responses were reversed from the letters in the schematic, just as everyone is reporting. My analysis showed greater sensitivity for the narrow connection though. I think your new analysis also reverses the connections from the schematic which shows the broad setting as the "earthy" end of the coil, while if I followed your argument correctly, you said the "earthy" end was the more selective connection? My main argument with your new analysis is your assumption of unity coupling between the primary and secondary of the coil. I don't have a good feel for what the actual coupling might be, but it is clear from the coil geometry that it is far from unity, given that the primary extends far beyond the secondary. Even if the two windings were bifilar all the way, I believe the coupling would still be far from unity. Could you examine your unity coupling assumption again, and give some discussion of why you think it is a valid assumption? If the coupling is significantly less than unity, your simple decomposition of the circuit would no longer hold, and your following argument would not be entirely valid, although your final result might still be correct. Also note that the "earthy" end of the coil is labeled "B", while the non "earthy" end is labeled "S". Regards, John Byrns
Hi John, First off, I missed your post #2122 on how the Mystery Xtal Set may work. I should have referenced it in my post #2132. I appologize for not doing so. One thing about this Mystery set bothers me. I believe reports of current builders that they find that the earthy antenna connection gives the greatest selectivity. I find it hard to doubt the statement of the original builder that the most selectivity occurs when using the non-earthy antenna connection. I wonder if the discrepancy is caused by a difference in thickness of the DCC insulation used in the 1932 build and that used in modern builds. Different insulation thicknesses will cause different copper spacings and affect the capacitive coupling as well as the magnetic coupling. The color of the wire in the picture of the contest winning build by Ray Creighton suggests that he used film insulated wire, not DSC wire. If you are reading this Ray, what kind of wire did you use? Getting back to your post, my recent analysis indicates that the most selectivity should occur with the antenna on the earthy end of the primary. re coupling between the coils: To get a better feel of the coupling, I used a coupled-coil computer program for two coils axially spaced by 0.0205 inches, one of 12.5 turns and the other of 25 turns with axial lengths of, respectively, 0.447" and 0.895". The computer program requires both coils to use the same ga. wire so I used wire of the average ga. (27 ga.) I could not find out the insulation thickness of double silk wraped wire. The insulation thickness I used was the average of insulation thicknesses of double nylon wraped and double celenese wraped 24 and 30 ga. wire. The results were: L1=60 uH, L2=190 uH and k=coefficient of coupling=0.49. I used the same number of turns per inch for each coil because the program requires this. I know that the Mystery set has a greater tpi for the end coils than the bifilar section etc... . I tried some other approximations and still got a k of about 0.49. Maybe this "not-unity coupling will reverse the effects of the B and S connections and maybe not. You got me there. To get a "feel", maybe? of the possible value of k between the bifilar section windings, I used the coil program to calculate the k of two non-bifilar coils, axially next to each other and of very short axial length. Wire dia.=0.001". Turns spacing=0.002". Turns=25. Coil lengths=0.05". Results: Each inductance=100.442 uH. k=0.93. I would expect the k of the real world bifilar section to exceed this. 0.93 is a pretty high k and, I think, justifies using an assumption of unity coupling for the bifilar coil section. I labeled the earthy end of the coil "S" because I found that connection to give more selectivity, not realizing the confusion that would be caused because the original article labeled it "B". I'll correct this later when I transfer this whole shootin' match to an Article on my Web Site. Can' o worms, tying to analyze this Mystery Set! Message to anyone who has built the Mystery Set and has a dual-channel scope good to above 1.7 Mhz: Please connect a 1.0 MHz sine wave between the B terminal and the G terminal. Tune it in. Determine the relative phase of the signal across the secondary, compared to that of the applied signal. Do the same using the S terminal. Make sure that the scope does not noticeably load the secondary. If possible, repeat at about 550 and 1650 kHz. Let us know the results. This should help very much in determining how the Mystery Set works. Ben
Hi Ben, Yes I used enameled winding wire, of the gauges specified, could not find any DSC wire, Regards Ray
Aaah, Ben, you clever fellow - I may be a bit of a slouch when it comes to the theory, but you know I can't resist the challenge of an experiment! So, I did as you requested (I think) and here are my results. Either they may reconcile your views and mine (and Ray's), or I may just have made the 'Mystery' set even more mysterious. Read on, gentle reader... First of all, let me cover myself: (i) My beloved old (VALVE) RF signal generator has a dodgy attenuator that makes setting low output levels rather difficult, and; (ii) My 'scope probes significantly loaded the circuit, as you warned, so I had to use the 10x attenuators, and; (iii) The 'scope does not reliably trigger with an input below about 15mV. So... ...I had to use an input level of 200mV peak-peak. I hope this is not too high so as to invalidate the experiment. I carried out all measurements with a pair of 2000 ohm headphones plugged in (couldn't get sensible waveforms without them). I think that they are 2000 ohms per phone, in series. In an attempt to reduce confusion (you'll see why later), I decline to call the two inputs B, S, or sensitive, selective; instead, I'll refer to position [1] being the aerial (signal) input at the end of the primary coil nearest to the detector end of the secondary, and position [2], the end of the primary nearest to the grounded end of the secondary. Output measured just before the diode. All with me so far? Are you sitting comfortably? Then I'll begin. (Continued in next posting...)
(...continued) Experiment #1: 1MHz, 200mV pk-pk. Position [1]: Output exactly in phase with input, with 65% of signal amplitude. Position [2]: Output 180 deg out of phase with input, with 240% of input amplitude. Yes, that's right, under these test conditions, position [2] showed (and sounded) a higher amplitude tone when the 1MHz signal was modulated with 1kHz audio. Now, the two input settings do not yield the same tuning ranges (which to me is another advantage of this set!). The high frequency limit is set by position [1], the low frequency by the response of position [2]. These considerations dictated the choice of the next two test frequencies. Experiment #2: 1.4MHz, 200mV pk-pk Position [1]: In phase, 65% of input amplitude Position [2]: 180 deg out of phase, 270% of input amplitude. (In other words, very similar to Experiment #1 above). Experiment #3: 550kHz, 200mV pk-pk Position [1]: In phase, 65% of input amplitude Position [2]: 160 +/- 10 deg out of phase, 300% of input signal amplitude. (Input waveform becomes severely distorted in this experiment, which makes analysis difficult.) Conclusions, then: (a) Well, phase results, as requested, are presented above. (b) I thought I was going mad when I got the amplitude results. Aerial in position [1] is that which Ray and I normally associate with higher sensitivity, position [2] with higher selectivity. Yes, I can feel you all grinning. But I checked and double-checked everything, then I put the aerial back on and listened. On a strong local, the AUDIO (demodulated) output was about 200mV pk-pk with the aerial connected to [1], and about 50mV pk-pk at [2]. I even tried experiment #1 above again with the aerial connected at the same time. Same results at 1 MHz as described above. (It was midnight UTC by then, so I don't think my unauthorized broadcasts will have seriously inconvenienced the neighbours.) So, it seems that we are BOTH right, just that the experiments (at least as far as amplitude goes) don't accord with the results of listening to radio stations! Over to you, Ben - what do you want me to do next? Footnote #1: Congratulations, Ben, on theoretically predicting the contents of my workshop, viz. dual-beam 'scope, signal generator, Mystery radio. Your model only failed to predict the fine detail of the eight foot gap 'twixt 'scope and sig.gen., the rubbish that fills the gap, and the fact that the Mystery set normally resides two floors lower down.
To Ge_Wiz, In my recent reply to your latest post I forgot to ask you about the wire you used in the inductors. Did you use film insulated wire of some sort of fabric wraped wire? Ben
Gee whiz, Mr. Whiz, Those experiments seem to show some surprises, so let's try to explain them, but first, thanks very much for performing the tests. In order now, 1. The 200 mV peak-to-peak input is fine. 2. We need a load on the detector. I showed a resistor. You used phones. If I can later convince you to try the experiment again, please put a parallel RC network in series with the phones. This will equalize the DC and AC impedances of the headphone load and reduce diode loading on the tuned circuit by reducing the diode current to the correct value for low distortion detection. The proper value for the resistor is about five times the DC resistance of the headphone set. The cap should be large enough to bypass the lowest audio frequency of interest. That frequency is usually 300 Hz, so a cap of about 0.2-0.5 uF would be OK. 3. I'll guess that the input and output waveforms were not very good sine waves. My experience is that harmonic wave form distortion is usually noticeable with Serviceman's type generators. I don't know what kind of generator you have. Since the harmonics are filtered out by the Mystery set's selectivity when measuring the output, but are present at the input, peak-to-peak measurements can introduce some error. The output wave form was probably distorted by nonlinear diode loading. Using a parallel RC in series with the headphones, as suggested in point 2. will reduce the RF harmonic distortion at the output and minimize the confounding effect of those harmonics, since they change the peak-to-peak voltage. 4. Experiment #1: Two thirds voltage and in phase-would be expected on position (1) since the diode is taped at the 2/3 turns point. This is good corelation between theory and practice. When using position (2), 200% would be expected along with 180 degrees out of phase. The amplitude discrepancy may be due to harmonics changing the peak-to-peak measured value from what it would be if no harmonics were present. The discrepancy is probably caused by an RF impedance transformation voltage step-up from Cc acting on the lower part of the tank inductance below the gound connection. See Fig. #7 and #14 in my post about the Mystery Set. 5. Here is the reason that position (2) tunes lower than position (1). The signal generator is a low impedance source (probably 50 Ohms) and it is in series with Cc. That places Cc directly across the tank, thus limiting how low the total tuning capacitance can go when the variable cap is set at minimum. This limits how high a frequency can be tuned in position (1). In position (2) Cc is tapped down onto 1/3 of the tank, thus reducing its effect on the total tuning capacitance to 1/9 of what it was in position (1). The low end limit tuning frequency should be about the same in positions (1) and (2). See Fig. #7 and #14.
6. Experiment #2: Same comments as for Experiment #1.
7. Experiment #3: Same comments as for Experiment #1 except this: At resonance the antenna terminal in position (2) is really looking into the equivalent of a series tuned circuit (a circuit zero, as we engineers say). This series tuned circuit is made up of Cc and the tank. Somewhere above resonance the input impedance looks like a parallel tuned circuit (engineers call this a pole). The low input resistance at resonance acts with the output resistance of the generator as a voltage divider at the resonant frequency, thus reducing the fundamental sine wave voltage that you can measure. At frequencies above resonance (550 kHz) the input impedance goes up and I would guess that the frequency of the pole is somewhere around the second or third harmonic of the input signal (1100 of 1650 kHz kHz). Since the input impedance at this frequency is high comparared to 50 Ohms, the voltage divider effect is negligible. The effect of greater harmonic and less fundamental voltage could explain the wave form distortion you observed. Probably the pole frequency did not coincide with any harmonics in the other tests.
8. Conclusion (b): Your observations are explainable. In the lab tests the source impedance is a 50 Ohm generator. The source impedance with an antenna is usually estimated to be a 25 Ohm resistor in series with a 200 pF capacitance and a 20 uH inductance. (see "Radio Engineer's Handbook" by F. E. Terman, First Edition, footnote on page 973.) Let's assume that your antenna-ground setup has the same impedance. If you try a lab experiment again with the generator connected to the antenna terminal through a 200 pF cap in series with a 20 uH inductor, I believe that you will measure less sensitivity and more selectivity when using position (2) than on position (1). Be sure to shunt the actual generator output with a 50 Ohm resistor in all tests. Thevinin's theorem tells us that this will change the source resistance to 25 Ohms and cut the generator's output voltage in half. BTW, the inductor will have little effect on these measurements and can be eliminated. It has its greatest effect at the high end of the band. I think that these new tests will correlate well with real world reception of broadcast signals.
I would appreciate it of you would repeat the measurements with the RC in series with the headphones, a 50 Ohm resistor loading the generator, and a 200 pF cap in series with the generator (If you have a relatively low loss 20 uH inductor, add it). When measuring the input voltage, measure the generator terminal voltage, not the voltage on the Mystery Set Terminal (1) or (2).
Another thing: With your test equipment already set up, you have an excellent opportunity to show the benefit of the parallel RC network in series with the headphones. A good test would be to drive the Mystery Set with a 1000 kHz RF carrier modulated by an audio sine wave. Observe the wave form of the envelope at the generator output. Observe the demodulated audio wave form at the output of the Mystery Set. Short out the R of the RC. Note the distortion increase in the demdulated wine wave. Note also the reduction in RF selectivity. The benefit of the RC is greatest for strong signals and at high modulation percentages. If you like what you see, replace the fixed R with a pot and find the optimum value for the least distortion. The RC has little effect at low signal levels.
Ben
Ben: okay, last offering of the day. Last things first (before I forget), I wound the primary close-wound (except for the bifilar bit) with 22 SWG polyurethane enamelled wire, and the secondary with 30 SWG ditto. Point (2): Yep, I'll do the parallel RC bit, but it will have to wait to the weekend - and it's my son's eighth birthday, so no promises! Been meaning to try that for a while, now. Point (3): No, the input and output waveforms were pretty darned good from my trusty AVO SG. A little bit of distortion, only noticeable at position [2]. Experiment #3, position [2] caused the biggest problems. This may have been intermodulations from broadcast stations (oh yes, guys, I forgot to mention; at night my Mystery picks up stuff (probably Deutsche Welle, SW) with NO AERIAL! Point (4): Thinking about it, the primary is short of a couple of turns at the 'earthy' end to get the tuning range right; this may partly explain the secondary signal exceeding 200% amplitude. Point (5): Ooops, my clumsy description; position [1] tunes lower than position[2]. [1] covers about 540kHz to 1400kHz; [2] covers 600kHz to 1600+kHz. Bang goes another explanation! (Continued in next post...)
(Continued...)
Point (8): Yes, I thought source impedance might come into it, but didn't expect such a large effect. I'll rig the RCL at the input as well as the RC at the 'phones and have another go.
Well, that's the weekend booked. Great! Sorry I can't reproduce the scope traces in these pages, or even for my own notebook. Now, our Applied Physics Department is in the process of closing... I wonder if they have any old 'scope cameras needing a new home...
Details of the galena experiment will have to wait until tomorrow. Concorde has just flown past my window, on its way to annoy the good people of NYC, and I must hie away to my cold and starving wife and child, and my fully-equipped luxury hobby shop.
Phil
PS Steve, start saving. This weekend's consultancy is gonna cost you a PACKET.
Dear Ge_Whiz, Of course, you are correct in your reply to point (5). I must have been fatigued at that point. Writing these things is very hard for me. I shouldn't have said that the low end limit wouldn't change from antenna connection on position (1) to position (2). What I should have said was this: Cc, as a fixed cap across the 12.5 turn bottom of the tank reflects, by transformation, to a cap 1/9 of Cc across the full 37.5 turn equivalent tank. See Fig.14. It should have relatively little effect on the tuning range when using position (2). On the other hand, when using the antenna on position (1), the full Cc is across the tank, compared to the condition when using position (2). That is the equivalent of adding a fixed cap equal to 8/9 of Cc across the tank. Its effect will be to reduce the high end frequency limit more than the low end limit percentage-wise, and it does. The low end limit goes down by 10% and the high end limit by 13%. (I used a high end limit of 1610 kHz) Ben
After the discussion that has been held so far, is there a possiblility that the orignal circuit can be improved upon??
Can the 'Mystery' be improved upon? Probably, after all the 'mysteria' has died down :-). The question, though, is one of quantifying the performance-to-simplicity ratio. The reason I built the Mystery in the first place was because I wanted a one-dial crystal set in my living room. It was between that and the 'Tuggle'. Having tried the Mystery, I caught caught up in its beguiling charm and never Tuggled. Anybody compared these two one-dial designs yet?
Hi Ge_Whiz, The reason I went for the Mystery Crystal set was due to its "simple design" , and the claims by "Proton" that is was a super proformer. It allowed me to build a very simple yet elegant crystal set for the SEQG competition Which is where the whole Mystery thing started. Note: Proton was the writer of a radio section in the Brisbane Sunday Mail Newspaper in the late 20's and into the 1930's, Need to go back to the state libruary and read the micro film some more especily after July 32 to see if Proton published any more crystal sets, or inprovements Will have to wait till July as we have a MAJOR tax change happening here in Australia so I am flat out making my programs complient with the new GST Regards Ray
Ray, I thought both your layout and execution was exceptionally artistic and tastful. I have indeed looked your link many times, but seen no original photo. Was your implementation your idea, from your own little brain, or did you reproduce some original photo? _J_
Hi Davidson, Thanks for the comments on my consruction, It was simpley a case of looking at what old parts I had and then I built the wooden frame to suit. I wanted to be able to show non radio types the simplicty of the parts needed to receive radio so used the open base and front board consruction, It was built over several months as I thought and re-aranged the parts, made the things like the coil and fixed condensor mount. I was trying to make it look very close to the period by using square wire for the wiring and going to a bolt shop to get strait slot round head screws etc. All the info I have is on the www page. Regards Ray in Brisbane
Well with all the hoopla about this set I don't think I want to be left out. I have the parts and wire gathered up now I just have to wait for a couple of rainy days to throw it together. Hopefully before my antenna has to come down to accomodate the new roof.
Sky: Don't worry, you may not need an aerial with this set. :-) On the other hand, a twenty-foot length of wire upstairs will probably get a few locals, more if you use a REALLY good piece of McGristle.
Hi All Mustery Crystal set People. The sarga continues, Well I was near the State Libruary yesterday so I droped in for a hour which became 5 hours, I was there till I got chucked out. I searched the Brisbane Sunday Mail from July 1932 (when the current http://www.clarion.org.au/crystalset/mystery.html Mystery set articles were published) Almost every issue there was a mention of the Mystery CS in querys to "Proton" the Wireless Circle editor. The published the Mystery CS article again on 22 Jan 1933, due to popular request (they had no photo copiers then and the original issues had sold out) and A Mystery crystal set builder must have got to the copy of the paper as the page with the mystery CS article was missing on the film. In the 16th April 1933 Issue of the Sunday mail "Proton" published his Mystery Plus Crystal set. I will add this article to the www site next month when I have time. he mentions trying at least 10 suggested variations fo the Mystery CS. Here are the changes, so you can try it now. I have not had a chance yet. On the original Mystery crystal set coil Wind on 15 turns of 24 gague wire. over the middle of the existing windings. Connect the Aerial to the top end of the new winding and connect the bottom to the earth on the secondary winding. Also he mentions reducing the existing primary to 45 turns ie: only wind on 9 turns on the primary after fixing the secondary, (Previous design was 13 turns) Proton said the main advantage of the new design was selectivity. Will add the full text and circuit to the WWW when I get a chance, I am puting out requests to Aus Radio Collectors to see if I can locate any original sets, given the amount of interest at the time, we might get lucky. Regards Ray in Brisbane OZ
Down under, they may have the "Mystery Xtal Set", but I have sitting on my bench, the "Miracle Crystal Set" from Uncle Al's Radio Shop in Oakland, CA. ca 1924. I convinced the keepers at the Museum to let me take it home and "Fix it up". It was a tough job. My protege' Alex, who is 11, had the FADA crystal Detector assembly back together with a new CW and the original XTAL, and picking up stations before I could get it out the door or the Museum. This kid will restore your faith in the next generation fellas. They are still out there. You just have to keep the faith and look for 'em. I plan to get some pics of this beautiful rig, and will post a couple. It is double tuned and uses a basketweave transfer coil to change the coupling. More later. --Loose
Interesting comments about the 'Mystery', Ray. Can you get a sponsorship deal from the Brisbane Sunday Mail for all this interest?(!) I too had to drop a few turns off the bottom of the coil to get the coverage to 530-1650kHz between the two aerial settings. Mind you, I also wound to a slightly smaller diameter and just thought that my calculations were a little off. Phil
Dear Ben (and all others following this strand with bated breath), I did the further experiments, as requested in postings #2218 and #2219, last Friday. Interesting results, with some limitations, as I will explain. First, the usual experimental notes and disclaimers: (i) Aerial position [1] and position [2] are still as I defined in #2209. (ii) I made up the input LCR network as requested. Standard values (and my junk box) dictated that the actual values used were R=47ohms, C=220pF, L=18uH. As predicted, the generator output voltage was halved; all the following results arise from a generator output (measured at generator terminals, as requested) of 100mV pk-pk. All voltages quoted hereafter are pk-pk values. (iii) I made up the headphone RC combination as suggested. Headphone DC resistance was 4k exactly, so I used R=18k and C=0.4uF. I have to say that there was no audible difference in sound amplitude or quality with either of two sets of 'phones when using this arrangement, nor any visible change in sine wave amplitude or purity at the headphones; HOWEVER, there was a significant increase in voltage at the input of the diode when these components were incorporated, and a small but noticeable improvement in selectivity when aerial Pos[2] was used, as Ben predicted. (iv) Both of the modifications above were in-circuit when the following tests were carried out. However, the input LCR network severely curtailed the tuning range when Pos[1] was used, which meant that measurements could not be made to this input when f>800kHz. Results follow in next posting. Input level was 100mV pk-pk when Mystery Set was NOT tuned to resonance - see results. Input voltage measured at generator terminals. Output measured immediately prior to diode. All voltages pk-pk. (Continued in next posting...)
(Continued from previous posting...) 500kHz: Pos[1]: Input level drops to 80mV at resonance. Output 760mV, 270deg phase advance. Pos[2]: Could not be tuned to resonance. 600kHz: Pos[1]: Input level drops to 70mV at resonance. Output 880mV, 275deg phase advance. As set is tuned to <600kHz, output moves into phase with input. As set is tuned >600kHz, output moves to 180deg out of phase with input. Pos[2]: Input reduced only marginally at resonance (near enough 100mV). Output 400mV, 85deg phase advance. Tuning set to <600kHz or >600kHz has exactly the opposite effect to that described for Pos[1] above. 800kHz: Pos[1]: Input drops to 52mV at resonance. Output 860mV, 267deg phase advance. Some distortion of output waveform (this is nearly the tuning limit of Pos[1] with the setup described). Pos[2]: Input unaffected at resonance (100mV). Output 520mV, 90deg phase advance. Effect of detuning set either side of 800kHz exactly as per 600kHz above. 1MHz: Pos[2]: Input 100mV. Output 640mV, 94deg advance. 1.2MHz: Pos[2]: Input drops to 90mV at resonance, accompanied by significant 2nd harmonic distortion. Output 810mV, 74deg advance. All measurements were made from an oscilloscope screen, so voltages are probably good to about +/- 5mV, and phase angles to about +/-5deg. Observations: As you predicted, Ben, Pos[1] now shows greater sensitivity and Pos[2] greater selectivity with this setup, in accordance with observations of actual radio reception with this set. Over to you! Phil
I finaly completed my mystery set yesterday. All I have left of my antenna is a 15 foot piece of lead in with one end dangling from the eaves. The new roof is almost finished and the new antenna will follow shortly. Even with this antenna remnant I was able to easily pull in all of the locals and a few neighboring stations within 100 Km. I must say I am impressed by the tuning capabilities of this circuit so far. In the selextive position the tuning is very sharp and any overlaping of the strong signals is almost non existant. I can't say if this won't change once I get the anteena back up. It is also a real treat to be able to tune in the entire BCB without having tapped coils or coupling to contend with. My next experiment with this cct will be to wind the coil using 21/44 litz wire. BTW Dr. Whiz: Can I have a credit for at least spellink your name right? re's Sky. To Sablesoft: In your web pages you mention that the "mystery set came in third in the building contsest" Just out of curiosity are there any details available on the two sets which came in ahead ?.
mystery set observation mjlee105 6/27/00 10:42 am
while playing with my set, i noticed that when i move the antenna from a1 to a2 my tuning order (not sure i am saying this right) reverses) in a1 right side of dial is 650 going to left is 950 etc. when in a2 650 on left 950 on right am i making sense, does anyone else see this, have i miss wired.
The "Mystery Set" is really not so mysterious. Refer to a copy of the Xtal Set Society Newsletter about late 1998 or early 1999. Look at the article: "Closet Cathedral" and you'll find this circuit was used in this project. I was first intoduced to the circuit about 12 to 14 years ago by an aquaintance through Popular Electronics. Wind an 80 turn, 22 ga. coil on a 2" form, then wind a 20 turn secondary of the same gauge, in the middle of the primary and on top of the primary windings, both windings in the same direction. Use a 400pf. variable to tune the primary, then use a 39" antenna and a ground on the primary coil. Use the secondary as the detector circuit. Use a standard headphone configuration and you'll have amazing results at tuning the locals in an area that has lots of nearby local AM stations. This is an amazing circuit and not so mysterious. Have fun and keep experimenting. Mike
Mike, I don't have that newsletter or Pop'tronics. The distinguishing features of the Mystery Set have to be the connection of ant. and gnd. to opposite sides and ends of the RF transformer, and dependance on the pri-sec capacitance of the bifilar transformer windings. This is weird, to put it mildly. I would be curious to know if you (or anyone) have seen these features before? I have seen many unbelievable designs published but never before this one. The example you cite does not seem to have these features, does it? (I do see that Mystery's equivalent circuit is quite common). _J_
John, The project was never in 'Pop 'Tronics, I said that I learned of the circuit from an individual through 'Pop 'Tronics. The "Closet Cathedral" project was in, either the late 1998 or early 1999 News letter. This little set's performance was very simple in design and performed very well. There are three main differences in the "Mystery set" and the one I describe: 1)The secondary is wound over the top of the primary, instead of bi-filar-winding it. The whole idea is to obtain as close-coupling as possible, so either way will work, but bifilar winding would probably increase the coupling, I would think, 2) The ground is at the primary end of the coil instead of the secondary side. However, the one I describe is using a 39" antenna against ground and the mystery set has the availability to use antenna on either end for a desired short or long antenna arrangement. This is why the one I describe is great for locals, densly populated, and 3)The ground being on the secondary side is similar to using a "false" ground with a variable or fixed capacitor to help the selectivety factor. The real secret to both designs is extremely close coupling of the primary circuit to the secondary circuit for maximum sensitivity and playing tricks with the antenna/ground arrangement for increased selectivity. The "mystery set" is a very neat arrangement, so try what I've described and compare the results. I think you'll be amazed at the similarity in performance and I'm of the belief the "mystery set" has superior long distance capabilities, due to the ability to move the antenna to the two different points. I will build the mystery set, as I've intended to do and see, and I'll bet the performance is very satisfacory, as I see the advantage to that set's design features.Mike
John: You keep refering to "capacitive" coupling of the pri. and sec. windings in the mystery set tuning coil. The coupling will be be mostly inductive and I beleive the capacitive aspect would be very small. The capacitance between the windings and the turns of the windings is greatly reduced by using the recomended DCC wire to wind these coils. This gives space between the turns of the windings reducing the capacitance. In a previous post I jokingly urged Prof. Phil not to bother testing the capacitance characteristics of the coil. what would be the point? Any improvement to the coil could only be made by using litz wire, altering the physical dimensions of the coil to improve the length to diameter ratio, and or removing the coil form. Re's Sky
Hi sky_wave_99 Great to hear your mystery set works well, Regards the two sets comming 1st and 2nd you can see them at http://www.clarion.org.au/crystalset/performance.html scroll down for more detailed photos both had 2 tapped/tuned coils, I have asked the builders for details but none were provided. Regards Ray
Mike, when you build it, see if you get the same result as I did. My radio with bifilar winding tuned across the whole BC band. My version with short coil wound centered over the long one cuts out at least half the BC band. Macrohenry
Okay Ben, it took me a while to catch up, but now I understand the numbers. This is why I was never any good at theoretical physics, and became an experimentalist instead! So, now we understand the calculations, you know my experimental setup, I've posted the input RF voltage and the resulting DC outputs. Can you make any sense of the Insertion Loss values that these appear to give? Phil
Phil, Congratulations. I believe your calculation is correct. Your measurements show that your Mystery set gives a reasonable insertion loss of 5.54 dB on (pos)1 and 14.2 dB on (pos)2 when driven from a 25 Ohm source having a Maximum Available Power of -49 dBw. Keep in mind this the input power level is quite high and that the loss will increase at a lower input power. If you can, how about one more measurement at as low an input voltage as you can reliably measure? What's next? Well, I have almost finished an new article that I intend to place in my web page soon. It will show a method for measuring the Insertion Power Loss and Selectivity of a crystal set with a AM signal. It deals with the ratio of output audio power to the power in the input sidebands, where the broadcast information is carried. Ben
I maintain it is worthwhile to realize why the Mystery Set is really mysterious. For me, the most mysterious thing about the Mystery Set is that it even works at all. I hope I don't sound disrespectful or step on anyone's toes about this. I do respect Proton and his creation for what it is, standing as a monument to the accessibility of crystal radio design to non-technical people, as I mentioned before (uh, post #2136). To appreciate that, one must note carefully how ridiculous this design really is. (Forgive me, Proton, RIP?) My opinion is that the Mystery Set is an entangled comedy of fatal design flaws that coincidentally cancel each other out. It's like Proton set out to design a flashlight foolin' around with a battery and bulb. He unknowingly tried connecting one end of the battery to the tip of the bulb. It lighted! It's like he didn't know he needed a complete circuit, maybe the metal table somehow completed the circuit, unbeknownst to him. He showed his circuit to "experts" who said it couldn't work, but it did. He published a writeup saying if you build it JUST LIKE me, it will work, and he was right. I imagine here is how he might have done it. First Proton apparently naively wound a RF transformer with pri and sec bifiler (wound parallel together) causing the max possible capacitive coupling, exactly what you don't want in RF transformers. It makes the transformer work as a single coil, defeating the isolation function of a transformer. He even put large wire on the antenna side when he needed it for the tuning capacitor side. That simply increased the interwinding capacitance further and spaced out the windings to reduce the inductance, hurting the coil efficiency. Then he (perhaps accidentally) connected the antenna to the top of the primary and the ground to the bottom of the secondary, leaving no closed circuit for antenna current. The set could never work like this, except that the excessive pri to sec capacitance coincidentally provided a path to his ground, on the wrong side of the "transformer". The transformer pri to sec capacitance then shows up in the antenna to ground path, serving to decouple the antenna a bit (I'm not sure the C is small enough to really do much good in that regard, wish someone could measure it, I think we wore Phil out). There are a few other things, but this is getting long. Heaven help the poor dope trying to learn how radio works from this example! No wonder Proton didn't know how it worked! I think these are the kinds of things kids do playing with inexpensive components when they don't have to worry about burning anything out. Proton's success proves the value of naive play that sets crystal radio "anti-technology" apart from other technologies. He kinda showed them' arrogant experts that you don't even have to know what you're doing to outdo them. _J_
Continued from previous post ... As we see from Ben's analysis, (+ a slight additional simplification) the radio equivalent circuit boils down to a fairly common single coil tuned LC tank circuit, with taps on the coil for antenna and detector, and an antenna decoupling capacitor in series with the antenna lead in. As luck would have it, this set architecture is particularly forgiving. The single detector tap is electrically maybe ¬ from the top of the tank and antenna taps at the top of the tank and maybe ¬ from the bottom (allowing for TBD bifilar effects). In fact, the radio should work a little better if implemented this way. _J_
John.. You may well be very close to the mark in your imaginary scenario regarding Proton's mystery set. The intitial test I did with this circuit (built to spec.) was only with a short (15') vertical antenna which was just a remnant from my old downlead. Sensitivity and selectivity were impressive although volume wasn't ear splitting. This stands to reason based on the short antenna. I did two experiments with this set today and it fell short of expectations considerably. I wound another coil using litz wire instead of the DCC magnet wire. My first thoughts on this were that I should be able to squeeze a bit more Q out of the new coil to sharpen the tuning even further. This failed drastically. The new coil although technicaly higher Q did not perform nearly as well as the spec'd coil. This was evident in the loss of sharp selectivity. Since the roofers finally finished yesterday I decided to get a new antenna up. I strung a length of #14 AWG copper stranded wire between a utility pole in my side yard and the mast I had previously used as my vertical. This now gives me an inverted L with 35 feet of downlead and about 80 feet of horizontal wire. Not shabby at all for an outdoor antenna. Of cousre I had to try this with the mystery set. Gasp!! Observation...this circuit does not cope well with a good antenna in an urban situation. The stronger locals bleed through like crazy and the BCB above 1 mHz is almost untunable no matter which switch position is used. With all due respect to Ray in presenting a very unique and interesting design I would only reccomend this set to beginers and or those who are antenna compromised. This set is not in the same league as the super sets at all. Sorry mystery folks but I calls em as I see's 'em. I would however be interested to see what other modifactions have been proposed if Ray gets the time to post them.
Sky and John, I am convinced that most of what you have said is true. I think the most interesting thing about the set is the non-conductive coupling of antenna to ground. If anyone is interested in trying to make the radio work with a bigger antenna, I made a couple of suggestions back @ post 2137. When you think of the millions of people who have experimented with crystal sets over the last century or so, it is not suprising most combinations of hookups have been tried, considering the limited # of components. Steve Holden
Sky, Yout new antenna is probably picking up more signal, but since it has more capaciance, it is coupled in too tightly. Try a variable cap in series with the antenna to improve the impedance match and I think you will be very pleasantly surprised. Ben
VEEERRY INTERESTING! Thank you for the report. You have saved me a lot of trouble. I was beginning to think I was missing out on a good thing although I couldn't see it in the schematic after reading all the analysis this set has received. This reminds me of my recent attempt to use 420/46 litz wire to make a higher Q toroid ferrite coil. The measured Q was a little lower than the same form wound with enam. solid copper. I am continuing my experiments with 40/46 litz to compare. Richard
I built the Mystery Crystal Set with a 44 turn, 3.5 in. dia. over-2, under-1 basketwound primary coil containing a 22 turn interwoven bifilar secondary coil. Both windings are 100/44 litz. This set has produced some curious results, so I wonder if anyone has been keeping notes on all that has been said about the Mystery Set. Specifically, does any of the following make sense ? 1. Antenna to the stator of the primary var. cap. gives very little selectivity and tunes only the bottom part of the broadcast band (up to ca. 900 kHz). 2. Antenna to the frame of the primary var. cap. gives much improved selectivity and tunes the entire BCB. ***These two observations are OPPOSITE what the original circuit diagram says.*** 3. High impedance mag. phones, such as Trimm Featherweights (20k), and Brush Crystal phones work well with the set. 4. Sound powered phones with matching transformer (50 k-ohm primary), or S.P. phones alone, do not work. For a single-tuned set, w/o wave trap, the Mystery Set works impressively well. But, I hope I don't ruffle too many feathers in saying it won't hold a candle to a "real" crystal set. Thanks & 73, -Mike-
Hello Mike: I have not kept notes, but have read all of the recent messages on the Mystery, which has certainly got my interest. One of the messages in the last day or two mentions that with a non bifilar winding only half of the BCB band will tune. I think I will just listen for a few more messages, if they pop up, and then just build the set! 73 Paul
To Ben T. & Mike T. et al: Ben: I tried including an extra var. cap in the antenna and then the ground. This does help the selectivity but messes up the very nice single tuning aspect of this circuit. Both caps have to be adjusted in order to best receive the desired signal. That isn't so bad I suppose but If extra tuning is required I would rather go a step further to include an extra tuning circuit which provides a bit of signal gain. But then it wouldn't be a mystery set anymore. With the shorter antenna the set worked very well on its own. I would still highly recomend this circuit to urban apartment dwellers or anyone who can't do much more for an antenna than dangle a 10' piece of wire out of a bedroom window. This set may also do very well with a big antenna in a rural environment well away from strong local signals. I still have a few more tricks to try with this set particularily with the coil. Mike T: I had no problem using SP phones with the mystery set, with and without a transformer in place. I can't think of any logical reason for this not to work for you. Something strange must be happening. et al: I would only recomend building the set with the specified wire in the tuning coil unless other methods of space winding the coil are used. The cloth covering provides the required spacing between the windings of the coil maintaining a reasonalby good Q. If getting the recomended wire is a problem try using ordinary telephone hookup wire for the primary and enameled wire or litz wire for the secondary. This would give you a suitable space wound coil. re's Sky.
To some extent, we now seem to be going round in circles with the Mystery set. We pretty much agree on how it works, and that it shouldn't, and on its limitations. The things that make it attractive to me are (i) single tuning, (ii) short aerial performance, (iii) (easily) switchable selectivity, in that order. This, to me is the essence and spirit of this circuit. As Sky says, modify it and it ain't a 'Mystery' no more. While I have previously indicated that I have received DX on it, I would not advocate it as a DX set, nor will I be entering it in next year's 'Hobby' class. We all know better ways to design a DX set, and there's little point to trying to build a racing pigeon out of a budgerigar. Phil
Mr.Sky,I have the same problem here in the DFW area-a lot of stations walking over one another.What I did was to put a coil across the extra ant.tuner of the Mystery and it seems to clean up the low end quite well.The high end is still causing some problems.But the only coil I have used with sucess is the one you get with the new plastic Radio Shack xtal set.I was thinking a making a tapped coil,but as you observed,tuning is getting more complicated.....S.R.
Phil: The mystery of the mystery set is only that the coil isn't just a coil. It's an antenna coupling cap and a coil rolled into one. Aside from the inductive coupling, the primary forms one plate of the coupling cap. Electricaly the physicaly open ended primary circuit is completed by the inductive and capacitive coupling to the seconday. The close coupled secondary forms the other side of the cap. To me this is ingenious whether by design or by accident as John D. has implied. I can't say that I agree with your statement that it shouldn't work. I suspect you may have overlooked the essence of the simplicity forming the complexity of this coil design.
Sky, You could well be right, it may be intentional, in which case I would think Proton a genius. But then I guess he would not have written what he did, or named it "Mystery". It has an elegance either way. I "want" to think it is by accident because of the "folklore" value of that thought, & it seems most consistent that way too. Either way, my hats off to Proton. & it indeed should work. When I say it "shouldn't work" I mean that as an anticipated comment of an "expert" upon casual observation. To hear those words, take it to your local RF circuit design weeny and watch his face when you show it to him. It would be even more elegant if the C wern't so lossy, a real capacitor design "should" work better. I have built single coil versions of this that do indeed work exceptionally well, but I have not yet built (or seen) anything quite like this, I'll fix that soon. ... and it would be even more elegant if we could show that the inter-winding capacitance is about "right", I guess about 40-80 pf. However, it "eyeballs" way too high to me. I would have thought one winding on top (over) the other might work better in that regard. (a reasonable Q cap. even variable as the needs vary across the band sould really be better). But then, as you all so astutely harmonize, it is no longer original, a big thing to me. _J_
Sky, What diode are you using with the Mystery? Ben
Hi mjlee105 I hope to put up the Mystery Plus Crystal Set next week. I will go back to the state Libruary soon as I get a chance and continue reading the sunday mail Wireless Circle to see if there are any more info published, The Libruary closed, just when I found the Mystery Plus article. Regards Ray
I don't agree that the set needs the bifilar winding to tune the entire BC band. My second Mystery set was made by winding the secondary over one of my 4"x6" coils, and it tunes the band quite nicely. Kurt
Mike, Perhaps Proton deserves "perspective" for being on a that island continent and having a young queen mother, (just kidding, when am I gonna' learn not to do that?) but let me make myself clear that I think most of all, he deserves "perspective" for a radio design that is good. Never mind that his apparent design methodology is unorthodox and makes good folklore, never mind that his schematic looks ridiculous, the set he produced is not that easy to beat. It is unlikely that the formal "accepted" orthodox design methodology could have produced a set with the advantages his has. His design is just riddled with design "errors" seem to tangle together and invert themselves into advantages. His bifilar winding of the coil actually space winds the center of the coil, improving its Q. I have noticed space winding the center half of a coil (as he effectively did) gives most of the Q advantage of space winding the whole coil. The small #30 wire on the LC circuit gets most of the advantage of being the larger #24 because it is spaced by the #24. The coil could be implemented as a single coil with taps, but with separate coils like it is, it can be made with screw connections and no soldering. And on and on. As for enhancements, almost any "enhancement" I can think of disrupts the interacting scrambled things that make the Mystery Radio work at all, not to mention well. The "enhancements" I can think of all help one feature very little at the cost of several other advantages. This radio represents a one "point" in a complex array of crystal set design tradeoff features and its hard to come up with any "points" nearby that are better in all regards. To me, the fact that it is an authentic 1930's original design and it has this "seat of the pants folklore" value are paramount. I sure plan to build one, EXACTLY like the directions, I just can't right now. _J_
i think that is what i was trying to get across. the design is good(simple, easy instructions, few parts, low cost etc), and met the needs (picked up 4 stations). heck, if we could get car companies to use the same design critera. . . (we would be all driving beetles). i built the set a few days after ray published it, and it still works for me. I still get two country stations on it. the dog still dies, the girl friend gets drunk and her mother was right, i'm no good. as a last note, i received the MRL collected works a couple of days ago in the mail,( i now have over a hundred rigs to build ), in a 1954 pamphlet, EO talks about various coil winding strategies. when he gets to bifilar his comment is, with slight paraphraseing since i am in an airport waiting to get out of st louie. " the winding is found primarly in this country on tv coils and in austrailia on radios." and best like the queen mother the design has stood the test of time.
I just finished building the Mystery Crystal Radio and it is working very good. I am receiving 6 stations in the Hartford CT area. I wound the coil using #20 enamel for the antenna tank and #30 enamel for the detector circuit. While I was listening to the radio I decided to experiment with some spider web coils that a wound a while ago. I used one of the spider web coils for the antenna tank with a 10 to 500pf variable in parallel. I used the other spider web coil in the detector circuit. Hooked the antenna to one side of the antenna tank and I connected the ground to the detector circuit. I then placed the two coils together and taped so they were very tight/close. It worked just about as good as the mystery crystal radio. The capacitance between the coils is providing the ground for the antenna tank circuit just as everyone proved. This really was a fun radio to build and with all the info presented in this club I learned a little bit more about crystal radios. I can't think of anything I've had more fun with over the years. What a hobby!!! frangolden_2000
Updated on the 28th Sep 2000 by Sable Software -
CrystalSet@clarion.org.au
I Build Mystery Radio
from
macrohenry
7 May 2000
Re: I Build Mystery Radio
from
sandhand (Dorado Hills, CA)
7 May 2000
Mystery Transformation
from
norfdak101 (Arlington,Tx.)
31 May 2000
Re: Mystery Transformation
from
davidson_j_ohn
31 May 2000
Re: Mystery Transformation
from
sky_wave_99 (Scarborough Ontario Can)
31 May 2000
Re: Mystery Transformation
from
bta_50g
31st April 2000
Re: Mystery Transformation
from
Wetsteve
31st April 2000
Re: Mystery Transformation
from
davidson_j_ohn
6th May 2000
Re: Mystery Transformation
from
wetsteve
2nd June 2000
Re: Mystery Transformation
from
ge_whiz
2nd June 2000
Re: Mystery Transformation
from
bhtongue
2nd June 2000
Re: Mystery Transformation
from
ge_whiz
2nd June 2000
Re: Mystery Transformation
from
davidson_j_ohn
2nd June 2000
Re: Mystery Transformation
from
Variocoupler
2nd June 2000
Re: Mystery Transformation
from
davidson_j_ohn
2nd June 2000
Why does the crystal set work?
from
sablesoft
2nd June 2000
Re: Why does the Mystery Crystal Set Work
from
k8cix
2nd June 2000
Re: Why does the Mystery Crystal Set Work
from
kevindunne_au
4th July 2000
Re: Mystery Transformation
from
wetsteve
4th june 2000
Operation of Mystery Set?, Part 1
from
bhtongue
6th june 2000
Operation of Mystery Set?. Part 2
from
bhtongue
6th june 2000
Re: Operation of Mystery Set?. Part 2
from
macrohenry
6th June 2000
Re: Operation of Mystery Set?, Part 1
from
bta_50g
6th June 2000
Re: Operation of Mystery Set?, Part 1
from
bhtongue
7th June 2000
Re: Operation of Mystery Set?, Part 1
from
sablesoft
8th June 2000
Re: Operation of Mystery Set?, Part 1
from
ge_whiz
8th June 2000
Re: Operation of Mystery Set?, Part 1.2
from
ge_whiz
8th June 2000
Mystery Crystal Set
from
bhtongue
8th June 2000
Reply to Ge_Whiz, part 1a
from
bhtongue
8th June 2000
Re: Reply to Ge_Whiz, part 1a
from
ge_whiz
8th June 2000
Re: Reply to Ge_Whiz, part 1a
from
bhtongue
8th June 2000
Mystery Radio _ where to now
from
mjlee105
12th June 2000
Re: Mystery Radio _ where to now
from
ge_whiz
12th June 2000
Re: Mystery Radio _ where to now
from
sablesoft
12th June 2000
Re: Mystery Radio _ where to now
from
davidson_j_ohn
14th June 2000
Re: Mystery Radio _ where to now
from
sablesoft
14th June 2000
Re: Mystery Radio
from
sky_wave_99
14th June 2000
Re: Mystery Radio
from
ge_whiz
15th June 2000
The Mystery Plus Crystal Set
from
sablesoft
17th June 2000
The Miracle Crystal Set
from
Loose_Coupler
18th June 2000
Re: The Mystery Plus Crystal Set
from
ge_whiz
19th June 2000
Mystery Set experiments, part 2.1
from
ge_whiz
19th June 2000
Re: Mystery Set experiments, part 2.2
from
ge_whiz
19th June 2000
Another Mystery set is born.
from
sky_wave_99
27th June 2000
Re: mystery set observation
from
peeblesorig
27th June 2000
Re: mystery set observation
from
davidson_j_ohn
27th June 2000
Re: mystery set observation
from
peeblesorig
27th June 2000
Re: mystery set observation
from
sky_wave_99
27th June 2000
Re: Another Mystery set is born.
from
sablesoft
27th June 2000
Re: mystery set observation
from
macrohenry
27th June 2000
Another Mystery explained
from
ge_whiz
28th June 2000
Re: Another Mystery explained
from
bhtongue
28th June 2000
Re: mystery set observation
from
davidson_j_ohn
28th June 2000
Re: mystery set observation
from
davidson_j_ohn
28th June 2000
Re: mystery set observation
from
sky_wave_99
28th June 2000
Re: mystery set observation
from
Variocoupler
28th June 2000
Re: mystery set observation
from
bhtongue
28th June 2000
Re: mystery set observation
from
rockhead_20601
28th June 2000
Mystery Crystal Radio Set
from
mike_tuggle
29th June 2000
Re: Mystery Crystal Radio Set
from
modern_radio_labs
29th June 2000
Re: Mystery Crystal Radio Set
from
sky_wave_99
29th June 2000 (
Re: Mystery Crystal Radio Set
from
ge_whiz
29th June 2000
Re: Mystery Crystal Radio Set
from
norfdak101
29th June 2000
Re: Mystery Crystal Radio Set
from
sky_wave_99
29th June 2000 (
Re: Mystery Crystal Radio Set
from
davidson_j_ohn
29th June 2000
Re: Mystery Crystal Radio Set
from
bhtongue
29th June 2000
Re: Mystery Crystal Radio Set
from
sablesoft
29th June 2000
Re: Mystery Crystal Radio Set
from
sandhand
29th June 2000
Re: Mystery Crystal Radio Set
from
davidson_j_ohn
30th June 2000
Re: Mystery Crystal Radio Set
from
mjlee105
30th June 2000
Re: Mystery Crystal Radio Set
from
frangolden_2000
30th June 2000