Testing the TK Tar Baby

[poynt99]

.99,

Is this the minimum component requirement for the oscillator?

GL.

For best results, you should have some inductance in all the places shown on the schematic.

As a minimum, you need some inductance off the Gate as shown, and some inductance in the battery lead. With enough inductance in the battery lead, you don't need any inductance in the load position in order to make it oscillate.

You only need one LED, as stated on the schematic. The two diodes were used and shown on the schematic only to represent an equivalent voltage drop of a single LED.

[TinselKoala]

Nice job TK. The wave forms look similar to the sim results.

Strange that you didn't find the results better with a LED.

I also noticed that it appears you added some resistance to the inductors. Those are unnecessary, and were only shown on my schematic to account for lead resistance, in case they make some difference in the real circuit with the real inductors. That's why I showed the resistance and inductance together in a dotted outline.

No, the supplier had two different types of 2.2uH inductors so I used one of each in series to give a nominal 4.4 uH. I understood that the resistances you put in the sim were the pure resistances of the inductors used. The two types measured a little different on my unreliable meter so I decided to use one of each type at each inductor position. Didn't the color code tip you off? Silver-red-gold-red-bigsilver? Alien resistors speaking Hungarian -- or inductors !!

The other MOSFET (Q1 as you mention) was replaced by the 1N4007, whereas the zener in the circuit replaces the FG, or bias supply.

Duh... yes I see that now. Thanks...

Anyway, I'm glad it worked. 

What is the scaling on the "VBAT" trace? What's the RJ11 jack for?

.99

RJ11 jack?  Huh? Oh, do you mean the yellow thing? That's a slug-tuned inductor in the place of the 200nH part of the CVR, turned on its side. It's probably less than 200nH but I didn't feel like winding a little coil, so I just grabbed something with a few turns and a slug in it out of the junkbox.
I thought about putting BNC jacks in but decided it was too expensive and complicated. The probe points are just simple loops of copper wire.

Vbatt is shown at 10 V/div and CVR is at 0.05 V/div, and the CVR is 1.1 Ohm metal film + the slug-tuned inductor. Tuning the inductor makes a tiny difference in some details of the waveform... I probably could use more inductance here. I'll have to check the vertical amp at that setting, it's the most sensitive range (0.005V/div and using the 10x atten probe) and I'm not absolutely sure about the voltage values indicated. But that probe (a new one for me, a Velleman 100MHz probe), I just realized, can be switched to 1x. Trying now...
OK, with the CVR probe at 1x and the scope atten at 0.05 V/div (three knob clicks up) , I see the same numbers for the peak voltages, and I trust the scope at that atten setting.

The circuit did a really strange thing though. When I got it together it didn't oscillate. I checked the build carefully, I tried different 2n7000s, different supply voltages, nothing. By putting a pot across the resistor in the gate circuit to give the mosfet gate some positive drive directly, I was able to get the LEDs to light up but still no oscillations. I probed around in the circuit with some jumpers, shorting individual components with a 100R or a bit of wire, and I discovered that by "tickling" the junction of the zener and the rectifier with a single wire end, I could get a brief dim flash of the LEDs. I messed around and messed around, and one time the LEDs came on for a half second. Then I kept messing and they stayed on for a full second. Then for two seconds, then four,  then they remained on whenever power is applied, whether with varying voltages from PS (7 to 11 V as you said) or with the 9V battery. It was as if something in the circuit had to be "conditioned" somehow. I thought maybe the cap needed some charge buildup before the circuit would oscillate, but that doesn't seem to be the cause. I can't get it to "stop" oscillating now whenever the switch is on and the battery is in place. I've shorted the cap to get rid of residual charge and it still oscillates immediately on applying power, and it worked first thing this morning too. It's working now, radiating zipons into the thermosphere as we converse.

The difference in the LED and the resistor load traces is mainly in the size of the peak on the Vbatt and the depth of the trough on CVR. It looks to me like the resistor will give the greater negative mean product, but I'm too lazy to want to do two mean power calcs by hand from photographs (although if it starts raining this afternoon I might do it anyway.)

Grounding the circuit to the box makes a big difference too. For now I'm leaving the box disconnected from the circuit ground.

[TinselKoala]

(replying to Groundloop)
For best results, you should have some inductance in all the places shown on the schematic.

As a minimum, you need some inductance off the Gate as shown, and some inductance in the battery lead. With enough inductance in the battery lead, you don't need any inductance in the load position in order to make it oscillate.

You only need one LED, as stated on the schematic. The two diodes were used and shown on the schematic only to represent an equivalent voltage drop of a single LED.

I put in a little socket for the "load". So I can put 2 LEDs in there, or one, or a 100R resistor, or whatever. The single LED makes the largest magnitude spikes by a large margin, and will likely make the largest negative mean power product, but the waveform is messier than that with two LEDs or the load resistor. So I've been running with the two tiny LEDs in series, but yes, it definitely makes bigger "noise" when just a single one is used.

[poynt99]

TK,

Ah, the two series LED's (3V to 4V drop) explains it then I think. It looks like you are getting about a 25V swing on "VBAT". With the two series diodes (1.4V drop), I am seeing about a 60V swing.

[TinselKoala]

I'm a little surprised that the polarity of the LEDs can't be reversed and have them still light up. Since the circuit works and oscillates with just a 100R in there, the diode action isn't required, I think...

Here's a funny: If I put the LED in backwards it doesn't light and of course there are no oscs. But if I then "short" across the reversed LED with another 2.2uH inductor, the LED lights up! And the oscs return but with an extra, large, positive going peak in the CVR trace.
If I have the LED in the "right way around" as shown in the diagram and glowing well,, shorting across it with the inductor kills the light and the oscs.