current source on Schmitt

[woodyc]

I couldn't find a thread but I'm sure somebody here has done this? If so, I'd be interested in seeing your version.

I was thinking about doing something like this...

[Merlinb]

I did it in a microphone preamp once. It looked virtually identical to your circuit! Works OK for linear stuff, but it's not suitable for guitar, IMO, because it freaks out if you try and overdrive it.

[woodyc]

I did it in a microphone preamp once. It looked virtually identical to your circuit! Works OK for linear stuff, but it's not suitable for guitar, IMO, because it freaks out if you try and overdrive it.

Well thats a shame, because the overdrive behavior is critical.

Merlin, any chance you could share the actual part values and voltages you were using? It would be very interesting to characterize the nature of the freak-out.

[loudthud]

I tried that in a bass amp but had other problems with the design. I didn't like that the grids drew current when the amp was on standby (no B+). Devised a circuit that cut the current off when there was no B+. Ended up just using a 30K resistor to -60V and switched it off with the standby. I liked having the grids at zero volts with Presence and Resonance controls in the feedback network.

[woodyc]

...I didn't like that the grids drew current when the amp was on standby (no B+)...

Thats funny, I didn't even notice. Its not a problem in my application, but a simple solution might be to lift the grounded side of R45 in my diagram and connect it to the cathodes instead?

[Merlinb]

Well thats a shame, because the overdrive behavior is critical.

Merlin, any chance you could share the actual part values and voltages you were using? It would be very interesting to characterize the nature of the freak-out.

It was many years ago, and the preamp is long gone, but I found the schematic. The trouble with a CCS is that when you try and overdrive and one valve is pushed to cut-off, the CCS can't maintain its desired current, so it just shuts down. You get big spikes appearing on your output signal when this happens- quite interesting to watch!


EDIT: Not necessarily the correct or finalised schematic!

[woodyc]

...The trouble with a CCS is that when you try and overdrive and one valve is pushed to cut-off, the CCS can't maintain its desired current, so it just shuts down...

I don't see cutoff as a problem unless the bias point and loadline are such that the other tube cannot carry double its idle current. Thats possible with a hot bias. Maybe that was the issue?

[woodyc]

...I found the schematic...

I took a closer look at your design and I'm pretty sure you didn't wire that up as shown because...

With your plate load the max current in either tube is about 1.7mA. Thats the highest current setpoint you could possibly use, but with the parts you show it looks like the CCS would try to pull about 5mA.

And there is a separate issue relating to the grounded emitter resistor. At your setpoint, you will have about 1.2V at the emitter. I think there will be trouble if the collector doesn't stay above that value, and I don't see how it could.

EDIT... In fact, I'm having trouble seeing how to make this work without using a negative dc supply.

[loudthud]

Good eye Woody! I ran my current source from -60V and used a 300V transistor. A TL431 kept the current rock steady. KOC uses a big negative supply and resistors in his book Principles of Power.

[Merlinb]

I took a closer look at your design and I'm pretty sure you didn't wire that up as shown because...

With your plate load the max current in either tube is about 1.7mA. Thats the highest current setpoint you could possibly use, but with the parts you show it looks like the CCS would try to pull about 5mA.

I think where the schem shows three diodes I actually used just two.

At your setpoint, you will have about 1.2V at the emitter. I think there will be trouble if the collector doesn't stay above that value, and I don't see how it could.
In fact, I'm having trouble seeing how to make this work without using a negative dc supply.

This is part of why it doesn't like being overdriven! There is not enough headroom. But this was for a microphone, remember, so headroom wasn't a great issue at the time, hence the lack of negative supply.

In general though I stand by my earlier comment- CCS aren't great for guitar use. Whether it's grid current or warm biasing, the CCS is liable to behave badly during overdrive. Maybe not in every conceivable circuit, but it's making a rod for your own back. But then, if you've got nothing better to do this winter ...

[woodyc]

I think where the schem shows three diodes I actually used just two...

Even with only two diodes, it doesn't work -- your "CCS" would try to pull 1.36mA per tube (.6 V/220 Ohm total current) i.e. bias each triode beyond clipping.

...In general though I stand by my earlier comment- CCS aren't great for guitar use. Whether it's grid current or warm biasing, the CCS is liable to behave badly during overdrive. Maybe not in every conceivable circuit, but it's making a rod for your own back. But then, if you've got nothing better to do this winter ...

How can you condem the concept and disparage people working on it (i.e. me) based on such a botched implementation? I may end up agreeing with you, but seriously, the idea is at least worth evaluating with a properly functioning CCS and a sensible bias.

[Merlinb]

Even with only two diodes, it doesn't work -- your "CCS" would try to pull 1.36mA per tube (.6 V/220 Ohm total current) i.e. bias each triode beyond clipping.

Like I say it was years ago, and that schem is just a sketch. I know it worked when I built it, but whether I ended up with different values I don't remember. The supply voltages on the schem look pretty contrived...

How can you condem the concept and disparage people working on it (i.e. me) based on such a botched implementation?

You don't need to implement everything to know how it will behave. Basic theory tells us that if the triodes are warm biased then when one cuts off, the other won't won't be able to draw enough current to keep the CCS happy. If you cold bias and use a negative supply then perhaps you can get away with it, but that's not what the OP showed in his schem.

And I didn't condemn CCS in general! I just think it's making life difficult for yourself. But like I said, if you have the time and inclination to fiddle with these things then be my guest.

[loudthud]

Opamps use the same basic circuit on the input. A diff pair with a constant current source. When you drive the input(s) beyond the input common mode range where the current source can't maintain it's current, strange things happen. Like in a TL072 the output slams the wrong rail if the inputs go too negative. Many designers overlook this when designing opamp circuits. Same applies to chip power amps. You need to clamp or limit the inputs to safe voltages.

[woodyc]

...If you cold bias and use a negative supply then perhaps you can get away with it, but that's not what the OP showed in his schem...

Merlin, The OP does show a negative 50V supply. It wouldn't work without it. But you're right about the bias in that drawing. Its not as bad as your schematic, but its clearly too hot.

I stayed up late last night and slapped together a CCS Schmitt using 6H8C tubes. Works awesome. I got all I hoped for out of the splitter and the output is very symmetric. On top of that, the overdrive is smooth and not glitchy in any way. Here's the exact setup I used and the results...



A pot was used to find the optimal bias. The way this works is you increase the input until the splitter is way past cutoff. Then walk the bias point up the loadline and watch the output increase to a plateau. If you go beyond that, the waveform begins to get asymmetric, i.e. where every other half cycle has a reduced amplitude. I forgot to take a pic of that but its not glitchy at all. It just adds even harmonic distortion. So there is definitely an optimal bias, and its better to be under than over, but its not a disaster if you're too hot.

[Merlinb]

Merlin, The OP does show a negative 50V supply.

You're right- could've sworn that wasn't there when I looked!