Negative psu for germanium preamps

[ruffrecords]

Your drawing is basically right. It just needs a few tweaks.

XLR pin 1 goes directly to chassis at the XLR and nowhere else. This applies to both the input and the output. Pin 1 is not a signal pin is is a screen pin. So the connection you have from pin 1 of the input XLR to the transformer screen should be deleted. Any shields connected to pin one should not be connected at the other end of the cable. Balanced cable screens should never be connected to anything other than chassis.

You need a separate wire from 0V on the PSU direct to chassis at the same place the mains is earthed.

If you go to the DIY page of my web site:

https://www.customtubeconsoles.com/diy

ans scroll down to the Power folder, in there you will find a document called Grounding101.pdf which explains this.

Cheers

Ian

 

[Bo Deadly]

Yes, that is correct. And yes, pin 1 must be connected to the 0V of the supply because that is how the 48V current returns to the supply.

However, you should really post a schematic of the preamp if you have it. If everything is transformer I/O it should be ok but there could be something referenced to 0V of the preamp that might assume it has a low impedance path to ground. It is very likely that if you just have a large filter cap on the PS, that will present a low Z path to ground but there might be some weird scenario where you would want to tweak the circuit.

Also, a few minor points are that your ground points should converge at the 0V of the power supply with one exception which is that pin 1 of both in and out XLRs should be connected to the chassis over the shortest possible length of wire (preferrably less than a cm or two). The length of that wire determines the wavelength of RF that can be radiated into the enclosure through that shield. So the phantom return current is through pin 1 into the chassis and then from the chassis to the filter cap ground point of the PS which is also where the preamp -48V connects and the earth ground from the mains connector.

Finally, your phantom power switching is a little odd. You should just have 48V feeding single pole switch with common connected to 100R and 100u RC filter and then two 6K8 to pins 2 and 3. That's the usual way anyway.

 

[hereforever]

Thanks, I follow all of that and the Grounding doc on your site is really helpful to understand.

I'll try to trace out a proper schematic before I get too much further.

I'm not sure why I drew the phantom switch like that.  Yes I would use a single pole in the usual way

 

[abbey road d enfer]

our phantom power switching is a little odd.

The way the OP presented it has the advantage of not loading the input when off. It may not matter much. Some claim it does.

 

[hereforever]

I spent some time tracing the circuit over the weekend and here is what I came up with

Points A & B on the schematic are wired to the i/o connector, I assume some kind of variable resistor gain control gets strapped here

 

[abbey road d enfer]

I spent some time tracing the circuit over the weekend and here is what I came up with

Points A & B on the schematic are wired to the i/o connector, I assume some kind of variable resistor gain control gets strapped here

There's a few weird things in this schemo.
Q3, with an output Z of 7k drives the 2.4k input Z of the next stage.
I don't understand the role of point B.
Q6 runs at full sail, without degenerative FB.
No global NFB.
Point A: using it to reduce gain is moronic.
Are you sure there is not NFB from the output to the emitter of Q4?

 

[hereforever]

Are you sure there is not NFB from the output to the emitter of Q4?

Checked again and that's correct, no NFB there