Ampeg VH-140C - channel selection and power supply Q's

[aion]

I'm working on an adaptation of the Ampeg VH-140 preamp and had a few questions about the channel selection.

Here's the schematic:


Each of the channels has 3-4 JFETs that act as switches, muting the inactive channel at various points. The JFETs are controlled by an op-amp comparator (IC7A, middle left) that appears to have a ~1V reference voltage, so it passes +15V or -15V depending on whether pin 2 is grounded. One channel uses P-channel JFETs for switches, driven by the negative supply, while the other uses N-channel and the positive supply.

In my adaptation I'll be using a Traco DC-DC converter to get +/-15V from a 9V input. There are only 5-6 opamps so the current consumption will be much lower than the rated maximum.

My questions:

1. Is the JFET channel switching overkill? I like the idea of muting at several points to control noise, and PCB space isn't an issue, but if this is all going inside a fairly small box with no need for remote switching then maybe it's better to simplify it?

2. Why are they pulling the comparator's reference voltage from +44V through a 2W resistor when +15V is one pin away (middle-left)? Is there an issue with sharing the supplies? I won't have +44V in mine, but trying to understand why they did what they did here.

3. Is the op-amp comparator method the best way of achieving this switching, or could a CMOS flip-flop work better? P-channel JFETs aren't as common as N-channel, so it would be easier to build if I could use the J112 for everything and not get the negative supply involved.

4. Assuming the op-amp is a suitable method, would it be it OK from a noise perspective to use half of an op-amp as a comparator and the other half for signal? IC6A (FX return buffer) seems redundant to IC5B since I won't have an effects loop, but if I had a spare op-amp from the comparator then I'd add it in - unless there would potentially be noise issues from mixing functions on one chip.

5. Lastly - are there any other considerations relative to using the DC-DC converter vs. the power supply in the original amp? Will the comparator and LEDs potentially inject any noise to the supply when they're changing channels that wouldn't be an issue with a transformer? I think of DC converters as being less robust, and it requires that less filtering be used, so I get worried about current spikes being more likely to make it into the audio - but maybe there's no real difference in this application.

I appreciate the insight!

[anotherjim]

My guess is that way the "comparator" input is going to see the true state of the A/B switch before everything else settles on switch-on, because the Zener clamped supply will come up before the bipolar one and drop out last on power-off.
It seems to require 44v mainly to supply the reverb tank driver which may be a pre-existing design and ready stocked reverb tank.
I don't know if R31/44 in the amp supply feeds are meant to provide some sag effect? Its unusual to provide series  resistors without a filter cap following. If they are for sag then you don't want any action in the unselected channel which might explain all those mutes. Also, it looks like there are no local supply filter caps for the circuits which would leave them open to interaction. Very odd.

[aion]

It seems to require 44v mainly to supply the reverb tank driver which may be a pre-existing design and ready stocked reverb tank.

Yeah, +/-44V is the raw supply coming from the transformer which is then dropped to +/-15V with regulators. It just seemed impractical to me to use 44V through a high-wattage resistor instead of 15V with a smaller resistor so I thought there must be a good design reason for it.

I don't know if R31/44 in the amp supply feeds are meant to provide some sag effect? Its unusual to provide series  resistors without a filter cap following. If they are for sag then you don't want any action in the unselected channel which might explain all those mutes. Also, it looks like there are no local supply filter caps for the circuits which would leave them open to interaction. Very odd.

It's drawn confusingly in the schematic, right-to-left: V1 and -V1 are the straight +/-15V supplies coming from offboard (not shown in this image). They aren't filtered locally on the preamp PCB, but both rails do have 10uf capacitors off 7815/7915 regulators on the power supply PCB.

[anotherjim]

You're right! There are 3 degrees of supply filtering scattered about. Direct C filtered, RC filtered and resistor fed. I'd have to sit down with a printout and use colour highlighters to make sense of it. It's not unusual to have separate RC filters to supply input stages, but I'm not sure if that's what they've done. For instance, IC4 has +V1 but -V2!

[iainpunk]

i believe that if you put it in a pedal, the JFET switching is overkill.

the muting of the inactive channel is a bit overkill as well, since the ''problem'' can be fixed with a 100nF film cap over each op amp's pins for filtering (place cap as close to chip as possible.)

if you are going with the Jfet switching anyways, i recommend redesigning the switching opamp to 4000-series Cmos, since its original scheme a bit odd with that 2W to 44V resistor. TTL needs a 5V supply instead of the 15v the preamp uses, Cmos can handle way higher voltages.
in this case, you can add an extra inverter for the jfets that are now Pcannel to work correctly when replaced with Nchannel

cheers, Iain

[Rob Strand]

Yeah, +/-44V is the raw supply coming from the transformer which is then dropped to +/-15V with regulators. It just seemed impractical to me to use 44V through a high-wattage resistor instead of 15V with a smaller resistor so I thought there must be a good design reason for it.

The amp has three foot-switches (from two sockets):  A/B, Reverb,  Chorus.

Chorus uses the 15V rail, Reverb uses the 15V rail with slightly different circuit than the zener.    The A/B switch is the odd man out.  It's also pushing a lot more current into the zener than the chorus.

The only things that makes sense to me are:
- using 44V was a last minute fix to off-load some current from the 15V rail.  (Speculation only - I haven't done a current budget.)
- they wanted to run higher switch current down the A/B switch for higher reliability as it provides a higher wetting current.
- extremely speculative:  it changes the switch timing by pulling up the comparator quicker.

Clearly the Chorus circuit is an example that it can work using 15V.

[duck_arse]

might want to edit this, 7400 is restricted to 5V supply, cmos up to 18V.

[iainpunk]

if you are going with the Jfet switching anyways, i recommend redesigning the switching opamp to 7400-series TTL, since its original scheme a bit odd with that 2W to 44V resistor. cmos needs a 5V supply instead of the 15v the preamp uses, TTL can handle way higher voltages.

might want to edit this, 7400 is restricted to 5V supply, cmos up to 18V.

yes, sorry, i was confused a bit, my head is somewhere else today, i thought electronics might distract me from that today, but nothing seems to go right today

[aion]

The only things that makes sense to me are:
- using 44V was a last minute fix to off-load some current from the 15V rail.  (Speculation only - I haven't done a current budget.)
- they wanted to run higher switch current down the A/B switch for higher reliability as it provides a higher wetting current.
- extremely speculative:  it changes the switch timing by pulling up the comparator quicker.

Clearly the Chorus circuit is an example that it can work using 15V.

Good thoughts, I hadn't even looked at the chorus comparator but yes - the same thing but done with a different supply for the voltage. I won't worry about it too much then.

For instance, IC4 has +V1 but -V2!

Yep, I noticed that and chalked it up to a mistake in the schematic. I think it's a safe assumption.

[Derringer]

FWIW

I built both channels into a box and wired the stomp switch to switch between channels - no bypass.
I ditched the jfet switching entirely and there is no oscillating or any other issues. I'm pretty sure I just made the switching as simple as possible which was just a toggle between outputs of the channels. It worked so I was good with it.


For power, I just used an LT1054 to develop the bipolar 15V supply.