Hum eliminator circuit experiment

[Potato Cakes]

Hello, everyone,

I recently found this: . The guy goes over how take the hum from the heater line and inject it back into the audio path to cancel itself out. I've attached a screen shot of his final schematic. So I had to test it out myself. I have a 5W single ended tube amp that has a solid state rectifier. The heater winding does not have a center tap but the schematic uses two 100 ohm resistors to artificially create one. It was already pretty clean but there was a bit of hum in it which didn't bother me when playing through it. I made the modification as shown in video and now it is completely hum free.

So I moved on a Matchless Spitfire that I built. It has a low level of hum but it so far below any guitar signal that is connected to the input that it never mattered to me. But there is still a hum and I wanted to see if I could repeat my previous success. Without thinking I made the same connection from the center tap of the heater winding (the schematic doesn't show one but my transformer does have it) to the cathode of the two EL84's and then a 4k7 resistor in series to the phase inverter's cathode. It made a sweeping oscillating sound that was very loud. I double checked the schematic and saw that there is 63V on the cathode of the phase inverter, so I couldn't use that connection point. I then tried a number of different connection points like the cathode of V1 but nothing made the hum go away like in the first amp. The "best" result that I found was connecting to one of the grids (pin 2) of V2, which is the phase inverter. The hum was completely eliminated but so was over half of the output. The guitar signal still sounded as it should but I had to turn the volume up most of the way to get the same output. Also when driving it hard it didn't breakup as smoothly as before. So I'm wondering if there is a way to use the hum cancelling circuit as shown in the video for the Matchless circuit? I've tried a number of different connection points but none improve and some create more problems. If there is a way to add the cancelling circuit in this amp it is not obvious to me. Not a big deal as I can record with it as I have it now, but for sake of possibly learning something I'm wondering if there is way to make this completely hum free using the example in the video.

Thanks!

Paul

 

[Potato Cakes]

One other thing of note while trying to find the proper point to inject signal into the amp to cancel hum. My amp is built with a switch to disconnect the 100 ohm resistors to the screens of the EL84s, which approximately halves the output. When I do this, the noise floor from the amp (small amount of hiss, guitar noise, etc) drops as expected but the hum level is still the same. When I turn the master volume down the hum disappears. Again, the level of the hum is below what I think most people who use tube amps would consider an acceptable level. It isn't audible in comparison to the level of a guitar being played through it.

 

[Potato Cakes]

Here is the schematic that the same guy in the video posts about his Twin-8 schematic. It uses an elevated ground like in the video, but it bleeds signal to the phase inverter from the speaker, like the way a Champ does negative feedback. Although the voltage at the cathode on his schematic is 1.7V whereas the Spitfire has 62V, so I wonder if I will run into the same issue with the crazy oscillating noise like before.

Thanks!

Paul

 

[dmp]

Interesting topic. Elevating the heaters is known to reduce hum (when the hum is originating in the tubes). How much is the effect of just elevating the heaters vs also tying to the cathodes for negative feedback?
And where is the hum coming from, B+? Ground layout? EMF from the heater wiring to grid wires? Magnetic coupling from the PT? I'd like to know if this cathode feedback 'fix' is a bandaid for something that could be fixed by a better layout. I don't want to watch the whole 20 minute video so I may have missed it as a skipped through, but it seems he thinks it is coming from B+
I've been using a bench supply vs the built in power supply to try to isolate the effects... I'll do some experiments on this in the near future as well.

For the matchless, don't use the CT, add two 100ohm resistors. Then elevate it with a voltage divider from B+, and see how much hum that reduces. Then use a feed resistor to the cathod of V1, like in the video, and see the effect.

 

[Potato Cakes]

I feel very confident about the grounding and layout of this amp. I had rerouted the wiring and looked at a number of grounding schemes and ended up on using a multiple star ground where each reservoir cap is the star point for the tube circuit that is fed by it. Then those grounds tie together and connect to the chassis at input jack. The power ground connects to the chassis right by the IEC inlet. This scheme has yielded the best results thus far.

From watching it a couple of times, the whole point is to take AC from center tap heater line and feed the AC voltage into the cathode of the output tubes and then use a feed resistor to the cathode of previous stage. The signal is inverted from the previous stage and cancels with the same signal on the output tubes at amount that is determined by the value of the bleed resistor. If I went to V1 then the signal would be inverted twice which puts me back to where I was. I did try connecting to V1 and many other places as I mentioned in my first post.

If I were to do the B+ voltage divider, I'm not sure what values to use since wouldn't this lower the B+? Or are you talking about doing so on the transformer winding?

I did remove the heater CT and used the 100 ohm method and there was no change in hum volume. I reversed the polarity of the heater connections from V1 to V2 and that seemed to help a small amount. There is also some hiss and a little extra hum with the input volume turned all the way up, so I know that I also have to look at some other issues regarding V1.

On side note he has more videos and does a Bassman 10 over haul and the end result is with nothing plugged into it and both the master and input volume turned all the way up there is no hiss, buzz, or hum (). So now that I've seen what is possible, that is now the goal. Again, the amp as it is right now is perfectly fine for recording and has less noise than the Vibro Champ reissue that is in the studio. I would rate the noise performance as very good, but what I'm after is exceptional.

I'm going to keep doing some reading and staring at stuff. I feel that I'm no more than two resistors, a capacitor, and about 8 inches of wire from figuring this out.

Thanks!

 

[Potato Cakes]

I wound up redoing my heater wiring for both 12AX7s with solid core wire that will stay together tightly when wound with a drill. Also learned a proper way to route wiring when connecting heaters to the 12AX7s as well. Plus I found that my bypass cap for V1 had been backwards for I don't know how long. I know initially it was fine when I first built it but somewhere in the several times I've taken this apart I've managed to get it switched. However I still can't get rid of the hum. Elevated heaters on the output tube cathodes or on the B+ voltage divider with either the CT or 100 ohm resistors it made no difference. I even redid the over all grounding for fun sending everything to the electrical ground as the start point but that seemed to make it a little worse. I also noticed I am getting RF interference on V1. When I turn the tone control all the way down it filters it out, so I know it's on V1. The idle hum itself I think is on V2 because when I short the signal on the cap before V2 it is still there. I did not expect to get so far in the weeds with this amp as I have so much more I need to be doing, but this has now become my obsession to get it as quiet as the Bassman in the video above.

Thanks!

Paul

 

[dmp]

From watching it a couple of times, the whole point is to take AC from center tap heater line and feed the AC voltage into the cathode of the output tubes

Here I'm a little confused. The center tap of the H+ is typically grounded, anyway it shouldn't have AC? Each leg of the H+ is out-of-phase 6.3v around the CT, between the two legs is 12.6v. But the CT is sitting at a reference. (if you've ever mistakenly powered up a tube amp build without the heater referenced to audio ground, you'll get a HIGH level of hum, because the H+ is floating all over, with no reference to ground). Similar to how the 5v winding for tube rectifiers is floated up to the B+ by the tube rectifier. The H+ is 6.3vAC but can be at any DC reference. Maybe I'm wrong though and the H+ CT will have some AC.

If I were to do the B+ voltage divider, I'm not sure what values to use since wouldn't this lower the B+? Or are you talking about doing so on the transformer winding?

No, it is just a voltage divider from the B+ with a high enough total resistance that it doesn't load down the B+. Say B+ is 300v, a 20k and 200k would give you 30v. Tie the H+ reference to 30v, and then the H+ is 'sitting' at 30v, with 6.3v AC. The reason you need to drop the B+ for the heater reference is because tubes have a maximum heater to cathode voltage in the datasheet. 12AX7 it is 200V, other tubes are often less.

I also noticed I am getting RF interference on V1

If you have it out in the open (not closed up in a metal chassis) this could be normal, especially if you have LED / dimmer lights.

I reversed the polarity of the heater connections from V1 to V2 and that seemed to help a small amount.

Polarity should be the same for all the tubes! If I'm using grn-grn heater wire I check this with a multi-meter before connecting to the transformer. If you've just haphazardly wired them and swapped as trial and error, disconnect and check that it is right.

with solid core wire that will stay together tightly when wound with a drill.

Solid core is good, doesn't have to be extremely tightly wound, the two wires just need to be close together and perpendicular to any signal wires. Best to send the heaters down along the chassis and have the signal wires away from them.

I did not expect to get so far in the weeds with this amp

Proper ground layout and overall perfecting a build for hum/buzz is the most difficult part of tube amp builds.
Check out Merlin Blencow's book on tube amps. Chapter on grounding.

I've found taking measurements, either with a scope or a soundcard / RMAA helps.
Having a bench tube power supply has also really helped, by being able to disconnect the amp from the in-chassis PSU, power it with the 'known-good' bench supply, and take some measurements of the 'best-case' noise (i.e. you'll always have some tube noise).
Then you can turn on your chassis PSU while still using the bench supply, to check for any magnetic coupling. Next switch to the B+ and H+, etc...
Step by step isolate where the hum is coming from. If you are measuring with RMAA you'll easily see when 60Hz or 120Hz hum appears over the noise floor.

There's no way to master and really understand this stuff without spending time doing experiments, like you're doing.
AND - it is my experience that you can make a hum free, noise-minimized tube amp with Merlin's ground scheme (bus bar, fig 15-10) and AC heaters!

 

[Matador]

This confuses me also: he is showing taking a sampling of the hum from the heater CT and "sending" it to the power tube's cathode, then "feeding it back" to a preceeding stage.

There shouldn't be any "signal" to sample at the heater CT: there should be no net current flow "out" of the CT and through the cathode resistance(s). In fact, if I had to guess, what is actually happening is that hum is travelling through the output transformer, through the power tube, and "appearing" common-mode at the cathode, which is then cancelled via feedback. In fact, given the cathode is heavily bypassed, I would think that essentially the point where the hum is being "injected" is actually essentially at AC ground (or very low impedance), so I don't see how this mechanism works as described.

 

[Potato Cakes]

This is the information from which I am working. It is from Designing Valve Preamps for Guitar and Bass. I am using the multiple star ground method as is recommended. The HT CT is connected directly to the reservoir cap and balanced as also shown in this chapter.

When I rewired my amp last night all of the polarity of the V1 and V2 heaters were put back in phase.

In the Bassman 10 video, he has no noise with the amp out of the cabinet and still getting no noise.

I will look over Merlin's grounding info. Thanks for sending it.

All heater wires are away from anything audio and are tight against the back of the chassis.

I was really hoping catching the backwards cathode bypass cap would have solved all my problems but to my surprise it didn't make any difference at all.

I'll keep poking around. I'm going to order some Y2 caps that are shown in the Bassman 10 video. I may also add an RF filter on the input.

Thanks!

Paul

 

[abbey road d enfer]

Here is the schematic that the same guy in the video posts about his Twin-8 schematic. It uses an elevated ground like in the video, but it bleeds signal to the phase inverter from the speaker, like the way a Champ does negative feedback.

Sorry to disappoint you, but some of these schematics are pure nonsense regarding potentially eliminating hum, because although it may work somewhat, in all cases it depends on parasitics that can't be transposed from one type of amp to another.
For example, in the Twin 8, the signal that is "fed back" is taken from a point where there is no legit signal. there is only a signal that is transmitted via the parasitic capacitance of the xfmr. Using another transformer or wiring it differently may well increase hum instead of decreasing it.
Regarding the Spitfire amp, the presence of a Master volume makes hum elimination valid at only one setting.

 

[Potato Cakes]

Well, that wouldn't be the first time the internet led me astray. However in the Hum Eliminator video the first steps before you get to the circuit is all about looking for parasitic noises first.

Still determined to get this one amp dead quiet.

 

[Potato Cakes]

Proper ground layout and overall perfecting a build for hum/buzz is the most difficult part of tube amp builds.
Check out Merlin Blencow's book on tube amps. Chapter on grounding.

Just realized that is the same information I am using for grounding.

 

[Potato Cakes]

So I changed the grounding again. I had been also looking at a number of repair videos primarily for Fender products which had grounding/hum problems and most what I found is that they resorted back to the separated ground points for the different stages. I isolated the preamp ground to the input jack and the phase inverter, power tubes, and speaker common connection to a separate ground point. I am using the multiple star grounding point method as laid out in the grounding chapter in Merlin's book mentioned above. I also elevated the heater CT to the cathode of the output tubes. I have before elevated to a voltage divider on the B+ and have also used the 100 ohm simulated heater CT method. This amp has been taken apart several times and had the bus and single ground point to chassis method used. I've tried everything including some things that were pointed out to be essentially snake oil on this thread. Still, the hum will not go away. Again I point out that this is a very low hum and I am using this amp to cut guitars for a client's record, so the level of noise is not an issue. It is the fact that there is noise and I see other post on the interwebs (which has become incredibly unhelpful and misleading) that this circuit and others tube amp circuits are being built with "zero noise" is fueling this crazed pursuit of mine. I am beginning to wonder if claims of amps being noise free that is equivalent to "zero latency" in the digital world. This is some hum but either the person making the claim can't hear it because of hearing loss or it being too noisy of an environment, or it's too low even be a consideration when actually playing guitar through it.

I will say that the hum is isolated to the phase inverter part of the circuit. When I turn the volume down (post V1) the hum is still there. When I turn down the master (post V2 - PI) the hum goes way. When I just short the signal from one side of the phase inverter to the output tubes the hum goes away.

This build has a switch that disconnects voltage from the screen grid on output tubes to halve the power of the amp and reduce breakup when using hotter output pickups. The curious thing is that the level of the hum is the same regardless of it being full or half power.

Also, even with the chassis in the head cabinet that has a shielding plate built into it, I still get the radio interference with the volume wide open, so I'm going to add an RC filter at the input and see if that makes a difference. However, if you plugged in a guitar and played at full open volume you won't be thinking about an RF problem as this amp is very loud. The Y2 caps i put at the power inlet did nothing for filtering RF that I could hear.

I have another push-pull circuit I'm going to build see if I can achieve my goal of a no noise/hum amp with the volume wide open and nothing plugged in (using shorting input jacks of course). After I see what level of success I have with this new build I will revisit thing. As is, it works and I can keep tracking guitars with this amp.

Thanks!

Paul

 

[abbey road d enfer]

Have you tried moving/rotating the PT away?

 

[Potato Cakes]

I did think of doing this. It would require me to drill new holes. The chassis was already predrilled for the transformer. You can see in the photo that on the far right side there is a hole which has gray, pink, blue, etc cables coming out of it. These are the primary windings. The secondaries come out under the board where you see the B+ caps which is where the brown twisted pair originates. My thought was to rotate the transformer 90 degrees so the secondaries where much closer to the tube rectifier and where the heater lines should be. This will take some doing and I don't want to keep putting off the other things I have to finish building.

Thanks!

Paul

 

[abbey road d enfer]

You don't need to do it permanently. Just for ascertaining the problem is related or not.