Graham Gunderson's Energy conference presentation Most impressive and mysterious

[Spokane1]

Dear Builders and Measurement Specialists,

Here is my first pass at wiring diagrams for the Graham Gunderson Logic controller. There are two versions. One is the way things appeared to be wired during the presentation. The other is the way the system was connected a week after the convention in Graham's shop. In that version he has disabled some of the chips and appears to have attached one or two pulse generators to possibly sweep input signals.

There is a fair amount of detail in these drawings. I have included pdf versions so you can study the small text. The resistors came in pretty good, but I couldn't see what the values were for the capacitors or the trim pots.

The next step is to develop a logic diagram so that we can figure out what is happening.

Spokane1

[TinselKoala]

@Spokane1:
That's nice work! I know how tedious that kind of thing is, especially when working from photographs that may not be too clear or from good angles.

But.... It seems from your schematics that the entire bit of circuitry that includes the components associated with U4, U2 and U1 is a Red Herring that does not connect to the rest of the circuit at all, other than being powered by the Vcc supply. This is the case in both the Experimental and the Presentation schematics. I'm sure you noticed this when you were drawing them up. This simplifies the problem greatly.

So here's what I think it's doing, based on your schematics.  We appear to have the 556 timer (U3) providing a clock signal that goes unchanged to both the H-Bridge and the Synchronous Diode. And we also have in the Presentation circuit, two monostable multivibrators U5 and U7. U7 appears to be an oscillator that provides pulses to the U8 hex inverter, which go through two of U8's gates then fed to the H-bridge. But this same signal through only one gate of the U8 is then fed back to the other monostable multivibrator U5 to trigger it, I believe, and it provides a pulse that goes to the Synchronous Diode. The U8 hex inverter also passes the input signal from U7 through three gates which then goes to the H-bridge. So the hex inverter's two outputs to the H-bridge are cleaned up, squared-off and oppositely phased versions of the pulse signal from U7. The U6 is not used in the Presentation circuit. There doesn't appear to be anything that synchronizes the U3 556 clock with the U7-U5 clock system. The U3 clock is fixed frequency and pulse width but the U7 and U5 multivibrators have trimpots for adjustments of either pulse width or frequency or both.

In the Experimental version, the U7 oscillator isn't used and an external FG (or two) is presumed to provide the clock signal for the U8 hex inverter and also the feedback signal for triggering the U5 monostable multivibrator.


Did you notice that my "mockup" of the U4 circuit (the TS372 dual comparator) oscillates, when it oscillates,  at exactly 60 Hz? This is undoubtedly due to those essentially floating reference pins 2 and 5, making the thing so sensitive and unstable that it picks up EMI from the house wiring, power supplies, etc. and oscillates in step with it, when the voltage supplied to the 1k resistor is just right. That is, if I built it right to begin with and if my TL082 is behaving as the TS372 would. Of course this issue is moot because this section of the circuitry isn't actually connected to anything in both the Experimental and the Presentation schematics.

[Spokane1]

@Spokane1:
That's nice work! I know how tedious that kind of thing is, especially when working from photographs that may not be too clear or from good angles.

But.... It seems from your schematics that the entire bit of circuitry that includes the components associated with U4, U2 and U1 is a Red Herring that does not connect to the rest of the circuit at all, other than being powered by the Vcc supply. This is the case in both the Experimental and the Presentation schematics. I'm sure you noticed this when you were drawing them up. This simplifies the problem greatly.

Dear TK,

Graham said that only 25% of that circuit was being used. With 8 IC's that means only about 2 of them have any useful function, which seems consistent with your analysis.

Now, I shall take your observations and compose a logic diagram so that I can understand what is going on.  The purpose of this whole exercise is to see how the backend timing pulse is generated. I suspect that since it is so short that the start and end times relative to the input wave form are critical.  After this is done others may prefer the use of a microprocessor, but if it only takes two or three chips then why bother.

You must get up early in the morning to do this kind of review so quickly.

Thanks for the positive comments. Doing this kind of work is a walk in the park (since I took some of the photos with my sons IPhone) compared to figuring out the E.V. Gray technology from 30 year old 110 format photos.

Spokane1

[Spokane1]

Dear TK,

Here is a simplified drawing after considering your comments. This part of the system shouldn't break any bodies bank.

Spokane1

[TinselKoala]

OK, I've breadboarded the 556 clock portion (U3) of the circuit. I posted this on OUR but I suppose I should put it here too.

The only flavor of 556 that I had in my box is a TI brand NE556N, and with Vcc = 5 V, and using the timing components listed on Spokane1's schematic, it oscillates at about 292 kHz.  Going up to Vcc = 9V it goes up to about 313 kHz.

This doesn't seem to correspond to any of the frequencies in the Gunderson scopeshot though.

Changing the capacitor on Pin 8 to 10 nF instead of 1 nF brings the frequency down to about 72 kHz, and shortens the duty cycle to about 57 percent, which makes a bit more sense. Maybe.

(Morning? Is it really morning already? My how time flies when I don't sleep.)