Can't adjust DC after 1N4004 diodes. But before yes. Why?

[JW]

I have a strange symptom here. I'm building an 8 channel 312 (API) The power supply I using, a JLM power station, has 5 regulators for +/-V1 +/-V2 and 48V. I can adjust these DC voltages fine at the power supply output, however, once I'm checking voltages after the CR1&2 1N4004 diodes on the 312 boards, it's always just a steady +16.5VDC/-17VDC. This gets me wanting to adjust the voltages at the power supply so it's +/-16VDC exactly, so I turn and turn the trimmers and nothing happens when I measure after the local diodes on the 312 cards. The voltage just stays the same. Then I recheck before the diodes and it's down at 12VDC or something. I tried replacing one of the diodes, but same issue. What is going on here?

I don't have any opamps in the cards, but they're complete otherwise. So it doesn't seem to be pulling current. Does this sound like I maybe tweaked the JLM power supply trimmers too much and took out a few regulators? It's just weird that I can dial in voltages fine before the card, but it stays the same after CR1 and 2, which is where it enters the opamp.

Another thing I noticed is that when I connect the ground of either of my 4 channel 312 boards to 0V at the power supply, I no longer get the 10 ohm resistance between chassis and star ground, which is a feature of the JLM supply. Once I disconnect the 312 boards, then I get 10 ohms again between chassis and star ground. Is this normal, or does it suggest a short somewhere? Both boards behave the same though. Hmm

 

[JW]

Back in five. The input transformer's shield seems to be the culprit in shorting 0V to chassis. My transformers are mounted to the case, so I'm guessing the shield for the input transformer shouldn't be connected to the 0V at the card's input. I'm pretty sure that's not the issue, but does explain the lack of 10R between chassis and 0V

 

[Khron]

On-board zener regulation downstream of those diodes? And how relevant is adjusting supply voltages, with virtually no load connected?

 

[JW]

No zeners downstream of the 1N4004 diodes.

Are you saying I need to have some opamps plugged in before I can get appropriate voltage at the opamp pins?

 

[radardoug]

Caps after the diodes are holding the voltage up. Just adjust at the power supply output.

 

[JW]

So, let's say I adjust for +/-12V at the power supply. Are you saying the caps hold a 16V charge for awhile? I thought I could get the voltages at the opamp pins before plugging in (hazy previous build memory)

The schematic I'm using is the CAPI VP312. There is nothing downstream of the diodes except the filter caps to ground, and then the opamp voltage inputs.

 

[Khron]

The diodes are blocking the higher voltages being dragged down from the supply side, and since they likely have no bleeder / discharge resistors in parallel with them (because the opamps should / are assumed to be there as a load), yes those caps DO hold that 16v charge for a while. And your (minimum) 1megohm input impedance of your meter won't be doing much to discharge those in a big hurry.

That's partially why some meters have a "low-Z" voltmeter setting, with a ~1k input impedance (one or several of the EEVblog multimeters come to mind).

 

[radardoug]

Easy way to tell.
Adjust supply to 12 volts. Then meter at CAPI after diodes. Is it 12 volts?
Adjust supply up two volts. Is it now 14 volts at CAPI?
Adjust supply down 2 volts. What is it at CAPI?

 

[abbey road d enfer]

Back in five. The input transformer's shield seems to be the culprit in shorting 0V to chassis. My transformers are mounted to the case, so I'm guessing the shield for the input transformer shouldn't be connected to the 0V at the card's input. I'm pretty sure that's not the issue, but does explain the lack of 10R between chassis and 0V

Of course: if the shield is connected to chassis one way or another, it shouldn't be also connected to another ground.

 

[abbey road d enfer]

What Khron says. Measure with IC's/opamps in place. Now 17V, 16.5V, whatever...won't make a difference.

 

[JW]

Okay, thanks guys. I will try powering on with an opamp.

Radardoug, that procedure is basically what I was doing, up and down at the power supply, with no change at all after the diodes. It's not going to hurt the opamp with the current voltage so I'll try with one in.

 

[mjrippe]

Okay, thanks guys. I will try powering on with an opamp.

Radardoug, that procedure is basically what I was doing, up and down at the power supply, with no change at all after the diodes. It's not going to hurt the opamp with the current voltage so I'll try with one in.

No, you started at 16/17v and thus charged the caps to a higher value.

 

[JW]

Okay, with opamps in, I can adjust downstream of the diodes fine now. Sounds excellent. Everything works so far. Sorry for false alarm.

 

[abbey road d enfer]

Sorry for false alarm.

Well, I think you learned a thing or two along the way..