Ludwig Phase II Clone Tech Notes

[Keppy]

Jimi & Dino, could you verify that R3 on the console board is indeed 1k? It's illegible in the schematic, and we've been assuming.

[digi2t]

Here goes, no particular order;

- R3 on console board is NOT 1K, it is 1M. Verified by color code. (Might be a biggie here  )

- Pedal toe/heel markings = Toe down = wiper to GND.   Heel down = wiper to point 4.

- Vowel switch, both contacts have no connection in the OFF position.

- Animation footswitch - to GND = OFF.   Lift is ON.

- Point 8 voltage readings - Toe = 2.62 vdc, heel = 6.35 vdc. (Taken with animation OFF, and rate at MIN. I got fluctuating otherwise)

- Point 9 voltage readings -  Toe = 6.81 vdc, heel = 3.149 vdc. (Taken with animation OFF, and rate at MIN. I got fluctuating otherwise)

Anything else you need, make a detailed list, and I'll get on it pronto. You're close man... I can taste it!

[R.G.]

- R3 on console board is NOT 1K, it is 1M. Verified by color code. (Might be a biggie here  )

Not a disaster. It's illegible in all of the scans I have. What that controls is how fast the voltage controlling the animation speed can change. 1K doesn't misbias anything or keep it from working. It just changes speed faster. But it will be necessary to get it right eventually.

Good catch!

- Point 8 voltage readings - Toe = 2.62 vdc, heel = 6.35 vdc. (Taken with animation OFF, and rate at MIN. I got fluctuating otherwise)
- Point 9 voltage readings -  Toe = 6.81 vdc, heel = 3.149 vdc. (Taken with animation OFF, and rate at MIN. I got fluctuating otherwise)

Good reference. Keppy, we need to see what yours is doing.

[Keppy]

- R3 on console board is NOT 1K, it is 1M. Verified by color code. (Might be a biggie here  )

Not a disaster. It's illegible in all of the scans I have. What that controls is how fast the voltage controlling the animation speed can change. 1K doesn't misbias anything or keep it from working. It just changes speed faster. But it will be necessary to get it right eventually.

Good catch!

- Point 8 voltage readings - Toe = 2.62 vdc, heel = 6.35 vdc. (Taken with animation OFF, and rate at MIN. I got fluctuating otherwise)
- Point 9 voltage readings -  Toe = 6.81 vdc, heel = 3.149 vdc. (Taken with animation OFF, and rate at MIN. I got fluctuating otherwise)

Good reference. Keppy, we need to see what yours is doing.

My readings overlap with Dino's. The exact readings depend on what formant switches are engaged. Dino's readings look about like my readings in counter mode. In parallel mode, both readings go from about 2.5v (heel) to 4.5v (toe). In vowel mode, they move opposite each other but at different values than in counter mode. There's enough overlap with Dino's readings that I expect to at least hear the effect of the pedal if the rest of the circuit is right. He says the toe moves the wiper to ground, though, which means I wired mine backwards, dangit!

[Keppy]

I noticed that when I hit the animation switch, Q2's emitter rises from 1.8v to 2.1v, and B2 rises from .7v to .76v. B1 stays unchanged at 2.45v. Geez I wish I understood oscillators.

[R.G.]

My readings overlap with Dino's. The exact readings depend on what formant switches are engaged. Dino's readings look about like my readings in counter mode. In parallel mode, both readings go from about 2.5v (heel) to 4.5v (toe). In vowel mode, they move opposite each other but at different values than in counter mode. There's enough overlap with Dino's readings that I expect to at least hear the effect of the pedal if the rest of the circuit is right. He says the toe moves the wiper to ground, though, which means I wired mine backwards, dangit!

I always bet that I wire **every** pot backwards, and expect to rewire half of them. 

The important thing here is that you are getting some voltage sweep of about the right amount going into the inputs for the formant filters. With that, you can start debug on the filters, whether animation is running or not.

I noticed that when I hit the animation switch, Q2's emitter rises from 1.8v to 2.1v, and B2 rises from .7v to .76v. B1 stays unchanged at 2.45v. Geez I wish I understood oscillators.

This is a slippery one. B2 and B1 are connected internally by a pair of resistances that meet at the emitter. In the 2646, these are about 2.5K and 3.5K when it's not tripped, which is when the emitter is below the voltage made by that internal resistor divider. So if the emitter is open, it just looks like two resistors, totalling about 7K. The emitter is set up with a resistance to a higher voltage and a capacitor to ground. The resistor and cap make the voltage on the emitter ramp up from zero.

There is an internal diode at the emitter to the resistive divider. When the emitter gets about 0.6V higher than the internal resistor divider, the diode lets current flow in from the emitter terminal. That current floods the semiconductor resistor bar with conductors and that makes its resistance drop dramatically. In particular, the resistance which used to be about 3.5K between the emitter contact and B1 drops to under 100 ohms, and starts letting massive currents in through the emitter, which makes the resistance drop even further. This keeps up until the emitter current drops below some holding current.

The resistor feeding the cap on the emitter has to be high, so it can't keep the hold current going on its own. If it's big enough, the capacitor runs out of juice, and the emitter current drops below the hold current, the internal resistances revert to full resistance, and the voltage on the internal resistor divider leaps back up to the no-emitter-current value, so the emitter current *really* goes to zero, and the cap on the emitter starts charging again.

The voltage on the emitter ramps up at a rate determined by the resistance feeding it and the cap to ground. It ramps up till it trips the emitter into conduction, at which point, a big spike of current goes into the emitter. This raises B1 in a short sharp spike of voltage, and pulls B2 down sharply. This spike is *fast* and may be over in microseconds. B2 is coupled through a capacitor in a way that trips the flipflop.

The voltages you describe indicate that the UJT is turning on once, and never turning off. B2 should be higher than B1 always, and the emitter should ramp up to about 5V by the simulation that I still don't quite trust to be perfect. But it is representative.

[Keppy]

This is a slippery one. B2 and B1 are connected internally by a pair of resistances that meet at the emitter. In the 2646, these are about 2.5K and 3.5K when it's not tripped, which is when the emitter is below the voltage made by that internal resistor divider. So if the emitter is open, it just looks like two resistors, totalling about 7K. The emitter is set up with a resistance to a higher voltage and a capacitor to ground. The resistor and cap make the voltage on the emitter ramp up from zero.

There is an internal diode at the emitter to the resistive divider. When the emitter gets about 0.6V higher than the internal resistor divider, the diode lets current flow in from the emitter terminal. That current floods the semiconductor resistor bar with conductors and that makes its resistance drop dramatically. In particular, the resistance which used to be about 3.5K between the emitter contact and B1 drops to under 100 ohms, and starts letting massive currents in through the emitter, which makes the resistance drop even further. This keeps up until the emitter current drops below some holding current.

The resistor feeding the cap on the emitter has to be high, so it can't keep the hold current going on its own. If it's big enough, the capacitor runs out of juice, and the emitter current drops below the hold current, the internal resistances revert to full resistance, and the voltage on the internal resistor divider leaps back up to the no-emitter-current value, so the emitter current *really* goes to zero, and the cap on the emitter starts charging again.

The voltage on the emitter ramps up at a rate determined by the resistance feeding it and the cap to ground. It ramps up till it trips the emitter into conduction, at which point, a big spike of current goes into the emitter. This raises B1 in a short sharp spike of voltage, and pulls B2 down sharply. This spike is *fast* and may be over in microseconds. B2 is coupled through a capacitor in a way that trips the flipflop.

The voltages you describe indicate that the UJT is turning on once, and never turning off. B2 should be higher than B1 always, and the emitter should ramp up to about 5V by the simulation that I still don't quite trust to be perfect. But it is representative.

I measured the pin resistances with the 2646 out of the circuit. I got 7k from B1 to B2, but I got 12M from emitter to either base.

[Keppy]

Another thought, R.G.: Your explanation makes it sound like the UJT is sensitive to current on its emitter, which is tied to the emitter of Q1. Doesn't that make the hfe of Q1 pretty important? Should I experiment with substitutions for that transistor?

[R.G.]

I measured the pin resistances with the 2646 out of the circuit. I got 7k from B1 to B2, but I got 12M from emitter to either base.

That's about right for when the emitter is not conducting current above its trip-over current.

Another thought, R.G.: Your explanation makes it sound like the UJT is sensitive to current on its emitter, which is tied to the emitter of Q1. Doesn't that make the hfe of Q1 pretty important? Should I experiment with substitutions for that transistor?

It doesn't, but only because the emitter of Q1 is separated from the emitter of Q2 by a big resistance. That's why I was wondering about the settings of the trimmer and the animation pot. If those are too low, the emitter can latch on.

When the emitter turns on, the current into the emitter and through B1 comes from the timing capacitor. The emitter-B2 resistance turns very small and it discharges the capacitor through the emitter terminal because that suddenly looks like a low resistance. The currents can be big, as much as nearly 1A. The emitter-B1 "on" resistance stays low until the current drops below some 'hold' current. If the resistance feeding the capacitor and emitter is big, it doesn't let current through equal to the hold current on the emitter, and when the capacitor is drained, the current drops and the emitter-B1 turns off.

With a big resistance between Q1 emitter and Q2 emitter, it can't latch on and stay, no matter what Q1 is doing. If that gets too low, then yes, it can latch on and never un-latch because the current into the emitter through the timing resistor will hold it on.

[digi2t]

My readings overlap with Dino's. The exact readings depend on what formant switches are engaged. Dino's readings look about like my readings in counter mode. In parallel mode, both readings go from about 2.5v (heel) to 4.5v (toe). In vowel mode, they move opposite each other but at different values than in counter mode. There's enough overlap with Dino's readings that I expect to at least hear the effect of the pedal if the rest of the circuit is right. He says the toe moves the wiper to ground, though, which means I wired mine backwards, dangit!

I hadn't thought of taking measurements with the different formant switches on/off. I'll do that now. Be right back....

All measures = VDC, and all measures taken with Animation OFF.

All Formants OFF, pedal movement made no change; 8 = 2.225, 9 = 3.807

Parallel - Point 8 - TOE = 2.785, HEEL = 5.91
Parallel - Point 9 - TOE = 3.00, HEEL = 6.47

Counter - Point 8 - TOE = 2.611, HEEL = 6.38
Counter - Point 9 - TOE = 6.75, HEEL = 3.069

Vowel - Point 8 - TOE = 6.23, HEEL = 5.84
Vowel - Point 9 - TOE = 3.110, HEEL = 5.68

More fodder to chew 

[Keppy]

Another thought, R.G.: Your explanation makes it sound like the UJT is sensitive to current on its emitter, which is tied to the emitter of Q1. Doesn't that make the hfe of Q1 pretty important? Should I experiment with substitutions for that transistor?

It doesn't, but only because the emitter of Q1 is separated from the emitter of Q2 by a big resistance. That's why I was wondering about the settings of the trimmer and the animation pot. If those are too low, the emitter can latch on.

With a big resistance between Q1 emitter and Q2 emitter, it can't latch on and stay, no matter what Q1 is doing. If that gets too low, then yes, it can latch on and never un-latch because the current into the emitter through the timing resistor will hold it on.

So too little resistance in that spot causes Q2 to latch on, while too much prevents the timing cap from discharging, also causing it to get stuck?

[R.G.]

So too little resistance in that spot causes Q2 to latch on, while too much prevents the timing cap from discharging, also causing it to get stuck?

Here's one fairly readable write up on it.
http://www.allaboutcircuits.com/vol_3/chpt_7/8.html

Too much also keeps it from running. The example in that writeup is the 2N2647, not the 2646. but they're close.

[Keppy]

Thanks! That article helped my understanding along, though it didn't give me any ideas for this circuit.

I have another question, though. Shouldn't console board C3 block low frequencies? And isn't that right on the LFO output? What am I missing here?

[R.G.]

I have another question, though. Shouldn't console board C3 block low frequencies? And isn't that right on the LFO output? What am I missing here?

Yes and no.

C3 does block low frequencies, but no, it's not the LFO output. Console pads 25 and 26 are the LFO outputs. What C3 does is couple the very fast down-going spike from B2 of Q2 into the flipflop made by Q3 and Q4. That flipflop changes state once per spike output from Q2. The two collector outputs are the actual LFO outputs.

[R.G.]

Did a little tinkering. You should be able to trigger the flipflop manually.

If you pull out Q2, and connect one lead of a 4.7K resistor to the B2 pin, you should be able to trigger a flip by touching the other end of the resistor to the B1 pin or to ground. It should flip once per time you tag it to ground. There is some uncertainty, because you will inevitably bounce a little when you touch ground, and it may trigger more than once per touch.

[digi2t]

Hi guys,

I've managed to get my hands on another unit. I should have it sometime at the end of next week. I'll be able to do some side by side comparisons, and I'll report my findings. I'm curious to see if that piggyback resistor will be in there.

You guys need anything, don't be shy.

Dino

[Ry]

I've managed to get my hands on another unit. I should have it sometime at the end of next week.

  Really?  Where are these things coming from all of the sudden?!?  I need one...deep in my soul.

[digi2t]

I dunno, must be due to the increased solar flare activity we're having this year. 

[pinkjimiphoton]

dude, did you get that other one on ebay?

[digi2t]

Yup. Apparently, the pedal and the animation doesn't work (sound familiar?), but fuzz and formants are fine. I'll fix it up, study it for posterity, and flip it.

I had some cash to invest. At the very worst, we learn some more about these units, and I break even. Or, I find a nice Ringstinger to trade for it.