Jack's MOSFET Crystal Set
Front of MOSFET set and tuner of plexiglass and poly-coated wood.
This MOSFET set is based on the January 2007 QST article, "High Sensitivity Crystal Set". The ALD110900A MOSFET used in the circuit is available from Mouser.
This QST article by Bob Cutler, N7FKI, has caused a lot of discussion among us crystal set enthusiasts. I built my first MOSFET set using a modified bandpass approach with toroid coils. There was so much Short Wave interference that I decided to start over and build the set in the QST article. Here are some of my initial tweaks to the QST set:
-Wired the Bogen as an autotransformer, including the pink windings, connected to a 12 position switch. Connected the Drain to the top (white) winding.
-My variable cap is one of the ceramic Russian ones with a built-in vernier, 450 pf.
-I used a 1500 pf fixed cap in series with the variable cap to match my large ferrite 300 uH rod inductor.
-The chassis is one of the plastic ones from Radio Shack. I held the inductor down with tie wraps.
-I used my old standby Sound Powered phones.
With no outside antenna connected, I can hear nothing in my rural location.
I next used an outside antenna with a series variable cap connected to the top of the detector coil, with ground; there was Short Wave interference over much of the dial. With no ground, the interference was much less. I accidentally zapped the MOSFET with these experiments...I am glad I bought more than one MOSFET!
I next assembled an outboard tuner using a Litz air core coil and a dual section Russian cap. With loose coupling between the MOSFET set and the outboard tuner, things have settled down.
Then I changed the coil on the tuner to a Litz coil on a feritte rod, mounted on a
plastic chassis to match the radio.
I next tried a number of experiments to check weak signal response of the MOSFET versus three pretty good germanium diodes. The IC socket allowed made it easy to unplug the MOSFET and plug in the diode. I set up the tuner and set a couple of inches apart and tuned in a station. I then separated the tuner and set a few inches and retuned. Here are some of the things I found:
-Tuning on the MOSFET set was noticeably sharper than on the diode set.
-The outboard tuner settings were a little different between the two sets.
-I was able to copy a station up to about 12 ½” separation between the tuner and the diode set. The distance increased to about 16 1/2” on the MOSFET set.
-With a louder station I was able to move the tuner and set apart over twenty inches with either the MOSFET or diode in place.
So is the MOSFET better than the diodes I tried? The MOSFET set seems to have the edge in sensitivity with the diodes tested. It is so selective that a high ratio vernier on the tuning capacitors would be needed to extract the most from a MOSFET set. Tests with more sensitive diodes would be useful. For some reason, the MOSFET did not work well in the toroid set. Special precautions need to be taken to insure the MOSFET is not damaged by high voltage (Bob used a shunt diode in his SW version of the set).
Is this the Silver Bullet we have been looking for in Crystal Set use? I don't know yet, but it definitely has possibilities.
Well, I have some more results from my experiments with the MOSFET set.
Strong signal handling
Last night I noticed an odd occurrence with some pretty strong signals. The signal level would suddenly drop and then stay at a lower level. Sometimes the signal would become distorted. This morning I tested the set by listening to a local powerhouse, and the effect is even more pronounced. I suspect this is a combination of internal design of the device and the .002 ufd capacitor.
Next I went back to the toroid set. It has a 10 microamp meter in the primary of the audio output transformer. On the same powerhouse station I pegged the meter using a diode detector. When I replaced the diode with the MOSFET and retuned to the station, the current went up to 5 microamps, then dropped to about 2 microamps...a kind of AVC action. The signal then sounded very distorted. I found a way around this, and that is to attenuate the strong station with a trap in the toroid set. In the MOSFET set, I simply move the set away from the tuner.
So, the MOSFET folds under strong signal conditions in the two sets I have tried. The diode detectors proved to be much more capable in handling strong signals.
Medium signal levels
I did not notice much difference between the diode and MOSFET detectors.
The MOSFET was better on weak signals, especially in the presence of other signals. The selectivity of the QST set and the toroid set was better when using the MOSFET. I was able to attenuate the power house station I mentioned with the trap in the toroid set with the MOSFET and then copy a very weak station 40 kHz away this morning. This was well after daylight when local station power levels are high.
I suspect the good selectivity is due to a couple of things. First, the full detector tank circuit (parallel coil and variable capacitor) are lightly loaded by the Gate connection of the MOSFET. Second, the switching action of the MOSFET seems to require that the tuned circuit be pretty well centered on the signal, and the sensitivity drops off quickly when tuned off the station.
Again, as noted previously, this comparison is with the diodes I had on hand. Results could vary with hotter diodes.
Note: SW interference is still a problem with the toroid set, but no problem at all with the QST set with outboard tuner.
12/27/06, 1705 UTC
More Updated Comments
Here is the latest on my experiments with the MOSFET set.
Upgraded detector coil
I upgraded the detector coil (a commercial ferrite rod coil) to a coil with the same rod core, but wound with 660/46 Litz. Now both the detector and antenna coils are ferrite rods with Litz windings. The tuning is sharp. It is getting very difficult to tune in signals due to the increased selectivity. A greater vernier ratio is needed than my caps have (3:1). I had to move down to 570 kHz for sensitivity comparisons, so I could had a little more spread on the tuning knobs.
The audio matching has been bothering me. The volume on the toroid MOSFET set with a 100k:100 ohm transformer was louder than the QST MOSFET set with a Bogen transformer. I bypassed the Bogen in the QST MOSFET set and put in a version of Steve Bringhurst's Select-To-Match (STM) that has a 100k:100 ohm Stanley transformer (with Benny), plus a Bogen, plus selectable series caps to the headphones. This was much better! I tried just a 100k:100 ohm Stanley transformer with a Benny. This was better than just the Bogen, but not quite as good as the STM.
Bottom line: the higher impedance audio transformer or better yet the STM worked better for my set than the Bogen alone.
Using a Benny
Both the STM and the Stanley used a Benny (a parallel resistor/capacitor in the audio circuit), but I still had that distortion when I tuned right on my 850 kHz power house station. I went back to the Bogen and put in a 250k pot shunted by a .5 uf cap that I had on hand. The Benny didn't help. The distortion actually increased as the resistance of the Benny was increased. This MOSFET is not at all like a diode when it comes to strong signals.
I left the Benny in the STM for the rest of the experiments for the lower level signals.
Steve Bringhurst brought several FO-215s to today's B'ham Crystal Set lunch. I picked the best of the two I tried and used that for comparisons to the MOSFET. I checked out the audio matching for the QST set with the STM, comparing STM settings for the MOSFET and the FO-215 . The matching for this diode and the MOSFET were almost identical for the QST set. The same STM setting could easily suffice for both the diode and the MOSFET.
My approach to coils for some time has been to not use taps, so both the diode and the MOSFET Gate were attached to the top of the Detector coil. Ben, I thought about your comments on taps, but I decided to hold the detector coil a constant (no taps). I tuned the detector and antenna caps for maximum signal with the antenna matching unit and the QST set several inches apart. (The diode loaded the set more than the MOSFET. I had to reduce the set capacitance to stay on frequency with the diode). I also tweaked the STM. I moved the QST set further away from the antenna tuning unit until I could just hear the signal. I backed the set up just a bit, tweaked it and the antenna unit, then moved the set again to that hearing threshold. I did this several times for both the diode and the MOSFET in my QST set. Even at the low frequency of 570 kHz, the tuning was very touchy for these sensitivity tests!
Bottom line is that there was just a hair's difference between the MOSFET and the FO-215 from a sensitivity perspective. In my experiments today it was just too close to call a winner. Consider it a tie.
I repeated an experiment I had done this morning with another diode, but used the F O-215 this afternoon. I got right next to 850 kHz during peak power hours and tried to tune in a weaker station. With the FO-215 diode I could get to 890 kHz. If I got any closer, I could only hear hash from 850 kHz. With the MOSFET I was able to tune in 870 kHz. That was WWL in the afternoon!. There was still some hash there, but I could clearly copy WWL. The MOSFET demonstrated greater selectivity over the FO-215 diode.
I was tuning a station end at the top of the band where a station was coming in with a warble. With the FO-215, it sounded like a normal warble. I could copy it with the MOSFET, but it had a peculiar sound, sort of like a warble on top of a warble. This seems logical since we are depending on the carrier of the station to switch the MOSFET on and off.
The zero-voltage-threshold MOSFET is certainly a peculiar device. Here are observations on the MOSFET based on several experiments with my modified QST set:
-It distorts and folds under very strong signals.
-A Benny does not help strong signal distortion.
-It can be zapped with voltage that a diode could easily handle.
-It is safer in crystal set use with no direct antenna or ground connection.
-The loose coupled approach works fine.
-It requires an extra winding for the Source/Gate path.
-The MOSFET and an FO-215 are very close in sensitivity.
-The MOSFET set works better with an STM or Stanley transformer than it does with a Bogen transformer.
-The MOSFET has improved selectivity compared to the FO-215.
-Hi-ratio verniers are needed.
The MOSFET has some quirks, but does it have any advantages over the FO-215? Yes.... selectivity. If you want to wring out that last bit of selectivity to copy a station, the MOSFET could help. This device just might be useful for a Crystal Contest Set…
12/28/07, 0035 UTC
MOSFET set top view
MOSFET set bottom view
Schematic of set and outboard tuner
Back to the Birmingham Group Radios.