• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Heater Wiring - the Good the Bad and the Ugly

I am in desperate need of some help. I would be very appreciative of anyone who can help me solve a problem with hum.

First, some background. I built the following headphone amp and it works perfectly, no hum.

Nobsound 6N5P+6N11 Vacuum Tube Headphone Amplifier Board DIY KIT Single end Class A Amp|Amplifier| - AliExpress

Next I got more ambitious and built the following headphone amp:

A Single-Ended OTL Amplifier for Dynamic Headphones. – HeadWize Memorial

This headphone is almost identical to the first. The amp part I built is exactly as in the Headwize link. The power supply for the filament is identical to the one from aliexpress. The high voltage power supply is a custom based on a bridge rectifier, followed by 47uf cap, followed by 50 ohm voltage dropping resistor, followed by 330uf cap, followed by 50 ohm voltage dropping resistor followed by 47uf cap. I have separate transformers for both filament and high voltage. The negative of both transformers is connected to circuit ground. The high voltage supply is on a separate board. The filament supply is on the same board, but on the other side of the board. the filament transformer is about 2 inches from the 6AS7.

The problem is that the Headwize amp has a severe hum. The hum is the same no matter what the volume is set to. The hum starts about 30 seconds after power on. If i turn the power off, the music continues to play for a few seconds without any hum. Also, I tried using 4 D batteries for the filament supply and the hum went away.

I have tried everything I can think of to solve this problem with no success, so as I said in the beginning I would greatly appreciate any help I can get. I used high end parts in the Headwize amp, hoping to have a better sound than the aliexpress one.
 
Hello Steven. I am no valve expert, but had the same issue because my filament circuit was acting as a capacitor. There are many threads on this. I think the problem is that your filament circuit and HT circuit are grounded together. The recommendation is to elevate the ground of the heaters to about 1/3 of the HT, using a voltage divider, and connecting the heater ground to that. Should get rid of that loud hum.
 
Hi Georgema,
Thanks for the suggestion, but can you be a little more specific? Do you mean taking the positive and negative of the filament DC output and putting two resistors across it and connecting the wire between the two resistors to the ground? Also how do I compute the value of the resistors?
Thanks,
Steve
 
So if the noise disappears with batteries the hum is likely due to the heater and cathode interacting, apparently acting as a capacitor. See this article on hum.

https://dalmura.com.au/static/Hum article.pdf
Page 3 is relevant.

What is recommended is to run the DC heater above the ground potential. You have connected the grounds together of the two circuits. Instead of doing this build a voltage divider using 2 resistors, say 120k and 30k or close to those values. The 120k is connected to the 150V then the 30k and in series these are connected
To ground. Only a tiny amount of current is lost. Then the heater circuit negative is connected between the resistors, and operates 30V above gnd. The heaters still see 6.3v dc, but the heater circuit is 30v above the cathode, and for some reason this reduces the loud hum you are hearing. There are other options to reduce the hum you are hearing, but this is cheap and easy. I can hand draw a diagram if you like. What I don’t understand is why other havent described this problem with the same circuit.
 
PROBLEM SOLVED !!!!!!!!!!!!
Thanks for all the suggestions. Based on your input this is what I did. I took the complete filament circuit and separated it from the amplifier circuit so that it is only connected to the amplifier ground in one place. Before I had the filament circuit connected to the amplifier ground at several places. For some reason the hum completely disappeared. This is beyond my knowledge of electronics to understand why that fixed the problem, but I am happy that it is fixed and I have learned an important lesson.
 
So if the noise disappears with batteries the hum is likely due to the heater and cathode interacting, apparently acting as a capacitor. See this article on hum.

https://dalmura.com.au/static/Hum article.pdf
Page 3 is relevant.

What is recommended is to run the DC heater above the ground potential. You have connected the grounds together of the two circuits. Instead of doing this build a voltage divider using 2 resistors, say 120k and 30k or close to those values. The 120k is connected to the 150V then the 30k and in series these are connected
To ground. Only a tiny amount of current is lost. Then the heater circuit negative is connected between the resistors, and operates 30V above gnd. The heaters still see 6.3v dc, but the heater circuit is 30v above the cathode, and for some reason this reduces the loud hum you are hearing. There are other options to reduce the hum you are hearing, but this is cheap and easy. I can hand draw a diagram if you like. What I don’t understand is why other havent described this problem with the same circuit.

Others have described it or suggested elevating the heater wiring in the thread regarding steventrus' hum issue.

The reason this eliminates hum (in those cases elevating the heater solves the problem) is b/c if the heater wiring potential is at the same level as the cathodes, some electrons leak from the heater to the cathode at the mains frequency (or in the fullbridge rectified case 2xmains frequency). Since the leakage is occuring at an audible frequency, we hear it. To the tube, it is merely a signal, or a fluctuating amount of electrons, appearing as a signal to the tube.

the edit: Just to clarify why this works a bit more... Electrons will travel from a negative place to a more positive one. This is how the tubes work, we have a cloud of free electrons around the cathode, the plate is positive, electrons flow from the neg cathode to the pos plate.
So the hot heater filament, since it is hot, naturally has some free electrons as well, willing to flow to a more positive electrode. If the heater wiring swings below the cthode potential, well now some electrons can leak from heater to cathode, creating that signal mentioned, at the frequency it is swinging up and down, every time it swings below the cathode potential. So, elevating the heater wiring to a potential much higher than the cathode potential, eliminates this unwanted flow of leaking electrons.
 
Last edited:
Open rectifier heater?

no element inside the tube should be left hanging not connecting to anything, i think this is the basic rule...

In my amp I have the indirectly heated rectifier 6X4 where the 6.3 v AC is not connected to anything; the 6.3V is only used for this tube. I t works. No hum.
Not connected - it is physically imho like a direct heated valve. If there is some leakage, the current cant go anywhhere so just capacitance . . .

Previously I had the 6X4 tube on a winding with the front stage driver: hum and noise. At that time I moved it experimentally to the ground, the only position without hum.​
 
Always call a photomultiplier tube a tube but anything with a heater inside is a valve to me. I remember the magical bluey/red glow from EL34's and aspire to build a valve amp (tube amp) some day soon.

Very interesting reading posts about heaters, the problems they create and established best practice, had some thoughts (knowing very little about tubes) so here's a list:-

1. I wouldn't want 50Hz or 60Hz anywhere around (both within audio bandwidth).

2. DC is quite likely to upset a lot of valve types (gut feeling).

3. Where should the mean(mid point) potential of a heater be relative to Cathode, Anode, Grid etc?

4. Can there be interaction between tubes (output to preamp) if the heater supply is truly floating?

Knowing what I know about electronics today I'm thinking about ways to power the heaters using today's electronics. Here's a list of possibilities:-

1. Make a power supply that switches from 4.45 Volts positive to 4.45 Volts negative every say 1 second, make the switching time very fast say +4.45V to -4.45V in 100nSec.

2. Make a power supply that outputs 6.3Volts RMS AC but at a lower frequency lets say 18Hz.

3. Provide a way to float the heater supply or to set the mean output firmly at some fixed voltage somewhere in relation to Anode or Cathode (presumably closer to the Cathode).

4. Consider making the RMS heater voltage variable in some way, even to the extent of pulling the heater supply a bit low when the amp isn't doing much and pulling the heater supply a bit high when the amp is working hard.

Does any of this make any sense, have I got my wires crossed ??
 
1. Heaters can and do work on 50 or 60 Hz every day without extra hum - it's about how the wiring is done to minimize stray magnetics/capacitances.

2. DC works fine for every tube I've encountered, and I usually use 12V SMPS to power 6V tubes in pairs.

3. That depends on the circuit, normally the heater is referenced to ground, sometimes it's elevated to a % of B+

4. Heaters should never float.

As far as the possibilities, they are endless. Pointless too though.

There are exceptions for directly heated tubes, but 95% of the tubes people are using are heater types, not filament types.

Koda
 
Thanks.

Yes I don't disagree with any of that, just glad to know that tubes are still being used and put into new designs.

It's a long time since I did anything with them, as a teenager I had one or two valve radios, even one of the ex-army No. 19 sets (I think it was made in Canada) complete with Carbon microphone and 12V power supply using two small rotary DC transformers, a VHF section and an 807 as the main transmitter valve. Then there were quite powerful audio amps around, its the EL34 versions I remember most.

Early TV's and especially colour TV's gave valves a bad name back then. All the tube heaters wired in series direct to the mains (via a rectifier ?) through a big hot green or grey dropper resistor!

So many types of valve bases and so many types of valves didn't help.

I can see how tubes can give excellent results today when circuit designs can be simulated in Spice and so on.
 
Member
Joined 2021
Paid Member
HF87 filament wiring

Hello!

I'm following the HF87 as a reference for my amplifier build, and I noticed the 6.3v wires do not go to the closest tube (XV7), but to the next one (XV5), is there a reason for that?
 

Attachments

  • HF87_filament_wiring.jpg
    HF87_filament_wiring.jpg
    150.8 KB · Views: 422
fighting hum with RF?

haven't noticed this sticky before. shame and ashes. some unsorted thoughts and learned-the-hard-ways:
  • AC ≠ 50/60 hz. we did shoot-outs @ ETFs, and had pete millett playing avvocato diavoli with high frequency AC heating. it should be easy to building an SMPS for e.g. 5V@1.5 A , 50khz. correctly filtered, the clean 50 khz sine wave should not IM with the signal. HF heating also takes care of the unequal voltage across the filament which is the reason why DC heating is said to kill DHTs.
  • DC isn't deadlier it's not the DC heating that kills the DHT, or at least not directly. current always impacts the thinnest and frailest part of the path. and it's not the continuous DC that kills DHTs, it's the turn-on, the inrush current with cold filaments. light bulbs always blew up at turn-on. and their filament didn't break in the middle, but always at one end where it was attached to the stem. same happens with the filaments of DHT. they blow up because the spot welded connection of the filament wire to the stems was stressed and got thinner each time the DC is turned on. until the day when the wire became so thin and so frail that it couldn't cope with the inrush … and pfff.
    using high frequency AC to heat DHTs would be one way to get rid of the IM right in the middle of the vocal range. but most of all, it offers the same advantage as 50hz AC by changing direction many times a second.
  • ramp it up! in these days and age when uinos and raspis are used for everything but walking your guinea pigs, why hasn't anybody come up with a true soft start board? first, it would ramp up the voltage of the heaters of indirectly heated tubes. ideally these tubes would be series connected and use current controlled heating. then the HF- or CCCV-filament supplies would be ramped up slowly from 0 to full voltage over, say, 2 seconds. that would prevent the huge inrush current stressing the weakest part of the chain – the connection of the filament to the stem. after the heaters and the filament have warmed up, B+ could be switched on.
  • twisting the night away. i don't know who is was, but one of the giants once told me to treat each heater winding as if it carried AC, even if in reality it carries DC. i strictly adhere to this to this day. a WE417-front end of a MC phono stage once gave me trouble. i assumed it was RF leaking and was coupled to the cathode. i remembered that many studios and broadcasters run star quad cable between microphone and preamp for more immunity of the signal. i built my own star quad and used it in the heater supply. that cured the problem.
  • PVC, PTFE, EFTE. a lot of hook up wire is teflon insulated these days. makes soldering much easier than the pedestrian PVC. and teflon seems to be ideal for tightly twisted heater wire. however, PTFE is a great insulator, but mechanically not very strong. EFTE (trademarks Tefzel. Fluon, Hostaflon, Dyneon) has similar properties but doesn't flow under pressure like PTFE, is easier to strip and mechanically more robust. some of the better wrap wire is EFTE insulated, but you can also get EFTE hook up wire. i prefer it for heater wiring because it doesn't move once twisted. and often EFTE insulation is thinner than PTFE i.e the twisted wires are closer together. (and before i forget: PTFE is highly microphonic or, more precise, piezoelectric. that's the reason why many there far less PTFE insulated microphone cables than there used to be.)
  • SRPPisssss never forget that the heater-cathode insulation is not very strong on most tubes. so often i see SRPP and other stacked circuits using one twin triode per channel i.e. the heater sits somewhere between the lower and the upper cathode potential. there are some tubes that were built for this, the 6DJ8/ECC88 being one of them. but for the majority of other small signal tubes, it's a much better idea to use the two twin triodes horizontally, i.e. one envelope for the bottom tubes left and right and the second envelope (with a separate elevated supply!) for the upper. this will prevent the premature hissing and spitting noise so often heard that gave SRPP a bad reputation. don't stress the heater-cathode insulation or you will be punished. the same also goes for cathode followers, particularly when used cathode coupled to DHT outputs – make sure not to exceed the max heater cathode voltage in the data sheet. (not directly related to heater wiring, but close by. it won't hum, but hiss.)
 
  • ramp it up! in these days and age when uinos and raspis are used for everything but walking your guinea pigs, why hasn't anybody come up with a true soft start board? first, it would ramp up the voltage of the heaters of indirectly heated tubes. ideally these tubes would be series connected and use current controlled heating. then the HF- or CCCV-filament supplies would be ramped up slowly from 0 to full voltage over, say, 2 seconds. that would prevent the huge inrush current stressing the weakest part of the chain – the connection of the filament to the stem. after the heaters and the filament have warmed up, B+ could be switched on.
That is actually something some members have been playing with. v4lve lover has made a few DHT regulators which limit inrush current in the first few seconds and let the tube warm up gently. Not as complex as you just described but pretty close to it. I have a few of his boards for my 6B4G --> GM70 build and while the amp is still in construction, me (just did some testing) and some other friends (playing amps) of him are quite pleased with them.