Routing AC heaters on a PCB.

Are there any tips and tricks to doing this with minimum interference from the heaters to other sensitive parts of the circuit?

Most of what I have read just say’s either:
A) Don’t rout it on the PCB and just use twisted pairs of wire coming away from the board.
or
B) Use DC heaters.

I would like to try routing the AC heater lines directly onto the PCB though. My instinct is to have the two traces routed side by side, or even along the same line but different sides of the PCB. However I noticed that DRIP takes a different approach, running the traces far away form each other (see the yellow highlighted traces in attached picture) which I could also imagine would work well since nothing else really seems to run along side each heater trace.

Any advice?

electric_diaries said:

Are there any tips and tricks to doing this with minimum interference from the heaters to other sensitive parts of the circuit?

Most of what I have read just say’s either:
A) Don’t rout it on the PCB and just use twisted pairs of wire coming away from the board.
or
B) Use DC heaters.

I would like to try routing the AC heater lines directly onto the PCB though. My instinct is to have the two traces routed side by side, or even along the same line but different sides of the PCB. However I noticed that DRIP takes a different approach, running the traces far away form each other (see the yellow highlighted traces in attached picture) which I could also imagine would work well since nothing else really seems to run along side each heater trace.

Any advice?

Click to expand...

The problem with the  DRIP approach is that the electric field, although reduced by the distance, is allowed to cover a much more important area, thus increasing the risk of polluting components.
I would opt for the "one trace on each side" method. Indeed, you must make sure that it does not cross sensitive paths.Let's say you have to cross a plate resistor; run the traces as close as possible to the HT side, not the plate side. If you have the choice, run them across the cathode resistor rather than the plate resistor (but not across the grid res).

I would do the AC heater tracks from one side as shown in the picture. Entrance of the power should be placed in the middle of the board. You should do also some calculations about minimal thickness of traces depending of the maximum heater currents (about 1,5 A in this case).

When I started learning to design tube mic pres about 10 years ago I had a lot of trouble with hum from AC heaters and quickly decided to use only dc heaters. Fast forward 10 years and the EZTube mixer designs all use dc heaters. However, I have recently been conducting some experiments using ac heaters with these boards for the lunch box projects. I fully expected a lot of hum to be induced into the signal path, especially into the front end of mic pres where the total gain available could be as much as 70dB. In practice there was little or no hum due to using ac heaters. There was quite a bit of hum due to the mains transformer being in close proximity to the mic pres and surprisingly a lot of this was picked up by the balanced mic input leads. Yesterday I tried a new mains transformer that is fully enclosed in mu-metal and there was no trace of hum using ac heaters.

I am not sure why the results are so good but I think the following factors are important.

1. There is no stray field from the mains transformer
2. I have a 0V plane on the PCBs
3. Elevating the heaters or not made no difference to hum
4. Using a 12AX7 with a spiral heater element made no difference.

My suspicion is that there is very little H field coupling in heater wiring (transformers are the more likely culprit) and the E field can be minimised using a ground plane.

Cheers

Ian

Thanks guys! A double sided pair actually works a lot better for me, and allows me to keep the heater traces short and to one side of the board.

Interesting results Ian! I have also found that if done properly, using AC heaters makes no noticeable difference to noise (when compared with DC)

ruffrecords said:

3. Elevating the heaters or not made no difference to hum

Click to expand...

Also interesting as this is one thing that has worked for me before, but it potentially was just compensating for another problem.

ruffrecords said:

2. I have a 0V plane on the PCBs

Click to expand...

Do you just connect all Signal 0V connections to this point? Most of the PCBs I have make are solid state low impedance circuits which work great with a common 0V signal plane. I would love to do the same with this design but I don’t know if it is as simple. Even though I am trying to keep the heaters far away, wouldn’t it be better if I could make sure that the return to 0V path didn’t run near them by using individual traces? Anything else to be aware of with the 0V plane? I have read your Grounding101 by the way, Great recourse, thanks!

ruffrecords said:

3. Elevating the heaters or not made no difference to hum

Click to expand...

Hum happens because the elevated voltage of the cathode's inner surface relative to the heater creates a space charge that establishes conduction between the two; any AC voltage on the heater will be transmitted to the cathode, albeit attenuated.
The cases where the cathode voltage is high are typically cath-follower, White or SRPP. These stages generally operate at high level, which renders them impervious to interference.
In most cases, low-level stages operate with very low cathode voltage, so this issue is not to be feared.

4. Using a 12AX7 with a spiral heater element made no difference.

Click to expand...

The heater being encloded in the cathode,  the latter constitutes an electrostatic shield. As long as the space charge phenomenon is not allowed to happen, the shape of the filament has no effect at all.

My suspicion is that there is very little H field coupling in heater wiring (transformers are the more likely culprit) and the E field can be minimised using a ground plane.

Click to expand...

Indeed.

abbey road d enfer said:

In most cases, low-level stages operate with very low cathode voltage, so this issue is not to be feared.

Click to expand...

Agreed, but in my EZTube designs the cathode of the 12AX7 input stage site at about +90V. Because of that and also because the output stage is an SRPP, I usually elevate the heaters. Despite that, removing heater elevation did not introduce hum which I admit surprised me.

Cheers

Ian

electric_diaries said:

ruffrecords said:

2. I have a 0V plane on the PCBs

Click to expand...

Do you just connect all Signal 0V connections to this point? Most of the PCBs I have make are solid state low impedance circuits which work great with a common 0V signal plane. I would love to do the same with this design but I don’t know if it is as simple. Even though I am trying to keep the heaters far away, wouldn’t it be better if I could make sure that the return to 0V path didn’t run near them by using individual traces? Anything else to be aware of with the 0V plane? I have read your Grounding101 by the way, Great recourse, thanks!

Click to expand...

The 0V plane does several things. First it is a low impedance 0V path for signals. You need to be careful to ensure output stage 0V signal currents do not flow in the 0V plane supplying low level circuits.  That is why the output stages on my boards are near the connector supplying HT and 0V and the input circuits are further away. You also need to ensure you do not introduce loops. I try to make the overall shape of my 0V plane follow the single 0V bus bar technique RCA used in their mixers back in the day.

The second thing it does is act as an equipotential on which E fields from other tracks will terminate. By flooding it around all tracks it reduces the propagation of E fields especially from heater wiring (although I did not realise that at the time I designed the PCB!!) Also, because it is an equipotential, it acts as a screen between boards (not relevant to your application.

So, in answer to your question, yes I would recommend you fill the entire area of one side of the PCB with it and connect directly to it but bear in mind signal flows and loops.

Cheers

Ian

ruffrecords said:

abbey road d enfer said:

In most cases, low-level stages operate with very low cathode voltage, so this issue is not to be feared.

Click to expand...

Agreed, but in my EZTube designs the cathode of the 12AX7 input stage site at about +90V. Because of that and also because the output stage is an SRPP, I usually elevate the heaters. Despite that, removing heater elevation did not introduce hum which I admit surprised me.

Cheers

Ian

Click to expand...

That's interesting. The data sheet for the Mazda/Belvy 12AX7A mentions max 180V between heater and cathode. Some brands may not be so generous.

That's interesting. The data sheet for the Mazda/Belvy 12AX7A mentions max 180V between heater and cathode. Some brands may not be so generous.
[/quote]

There does seem to be a lot of variation in the specification of Vhk between different brands of the same tube. The cathode of the top triode in my SRPP sits at about 150V. The cathode of the 12AX7 sits at about 90V. My heater elevation voltage is 25% of the HT supply which is in the region of 75V. This means the 12AX7 cathode is about  15V above the heaters and the cathode of the top triode in the SRPP is around 75V above the heaters and the cathode of the bottom triode is about 75V below the heater voltage. It is easy to forget that Vhk includes the signal peaks which for the SRPP can reach another 60V.

Cheers

Ian