Connecting audio and DC heater grounds on PCB?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

rock soderstrom

Tour de France
Joined
Oct 14, 2009
Messages
3,400
Location
Berlin
I usually run my audio and DC heater grounds separately in my tube preamps. The heater is then connected to the HT+ ground at one point, usually at the output of the HT+ power supply, my central ground point, which is connected to the chassis and protective earth.

I am designing a PCB at the moment and wonder if it would be disadvantageous to consider the DC heater ground and the audio ground as one?

I am building with steel tubes, the pin 1 of these I also connect to audio ground, it would be very space saving to connect the heater ground to this as well.

Is there anything wrong doing it this way with a well filtered regulated heater voltage?
 
Last edited:
You could route pin1 of the tubes to a separate track that goes to a standoff that grounds on the chassis. you also gain extra support under the tubes which is good to have.

I wouldn't try mixing the heater ground with the audio 0V
 
Thank you for your reply and suggestions.

I think I will wire the heater again independently from the PCB with nice thick extra wires.

Whereby it is still not really clear to me what kind of problems I could get with a DC heater? More noise?

Good idea with the separate trace leading into the chassis for pin1.

How do the others handle pin1 of steel tubes?
 
I have had great success grounding the positive 6.3v of a lm317 regulated dc heater supply this allowes me to directly mount the lm 317 to the chassis without a silicone pad and achieve maximum heatsink. Fully regulated DC supply has been the only acceptable outcome for tube mic pre. I have not had heater to cathode voltage problems with a couple years of use. 70-80v cathode voltage. The regulated dc supply with a linear regulator was used for all the tube units at a certain boutique company I worked for. Very quiet very stable. Never had to fix this circuit.
 
Yes, I am aware of this. The initial question was what happens when the DC heater ground and analog audio 0V are identical on the PCB, for example in the form of a large ground plane. What specific disadvantages would this have? As I said, I'm talking about well-filtered DC voltage, not AC voltage. At first glance, this seems identical to referencing the heater voltage to the audio 0V at the power supply, as I usually do.
 
It is not identical because you have heater current flowing in the ground plane. This will cause a dc voltage drop which is probably not a problem but might possibly upset the bias of sme tubes. However, it does mean that any noise in the heater supply is directly in series with the input signal. It may only be a hundred microvolts but that is enough to degrade the S/N of a mic pre.

Cheers

Ian
 
IIs there anything wrong doing it this way with a well filtered regulated heater voltage?
That would deprive you of the possibility to run elevated heaters, if need arise.
In a decently designed system, there is very little current flowing from the heater circuit to ground, only capacitive leakage from the mains xfmr, so it can be grounded almost anywhere that's convenient, and does not need a very sturdy connection. That's why the virtual center-tap works so well.
 
Given it is a preamp, then you need to be aware that there can be leakage current from input tube cathode(s) to heater circuit, as the cathode-to-heater interface includes a resistance in parallel with a capacitance equivalent model. If the heater circuit is dc, and connected to signal ground (or an elevated and bypassed node) then the heater is not going to drive any signal through the leakage capacitance, and current through the resistance is dc. So the issue is then whether the cathode has a signal on it relative to signal ground. If the most sensitive circuits are grounded cathode, or heavily bypassed cathode bias, then there is effectively no signal to drive leakage current through the interface capacitance. So it depends on your circuitry.

The other aspect is that any such interface leakage current should preferably circulate local to input stages that generate the leakage - and hence ground the heater circuit via a single link to near the input circuit signal ground (ie. not at the power supply) at perhaps the 0V node of the input stage where the local B+ bypass cap is star grounding, as well as external signal is locally grounding.

Note that there would be noise circulating via the heater supply secondary winding parasitic coupling to other circuits (B+, etc), but that noise loop is then isolated from the most sensitive input stage 0V due to star connection.

But unless you have a low noise setup, and good spectrum analyser with low noise floor, you probably won't notice anything anyway.
 
Thank you for your input. I've been building more of my circuits on PCBs lately instead of Turret boards. In the beginning I just transferred the building method more or less, which of course doesn't really make sense, because you don't really use the advantages of a PCB. Long story short, I've changed my build method and in doing so new questions arise...I guess I'll have to try some things to see what really makes the difference.
 
Back
Top