[DutchGFX]

Hey guys, I come to you dejected and defeated after trying to fix the hum in my tube amplifier. This amplifier is for a school project, and I designed it myself. I have rewired it to shorten most of the wires, but I still can't solve the hum. Here are some notes that may help anyone willing to assist:
 
The hum persists when I remove the input tube
 
1. I am running the amp without a potentiometer currently, but when I added a resistor on the input to simulate a pot, nothing changed
2. Removing the B+ does not affect the hum
3. I tried using a 2.5-0-2.5 transformer to supply AC to the tubes (paralleled) but it increased the hum
4. Shorting the output eliminates hum completely (obviously)
5. Bottlehead crack plugged in to the same outlet produces 0 hum, so the issue isn't my mains line
6. I have used a scope (not a very good one, I don't think)
7. Adding/removing grid leak resistors had 0 effect, as did shorting the grids to ground
 
I am using the 21st Century Maida Regulator from Tom Christiansen for B+, and using Pete Millet Regulated Supplies for my 2A3 filaments. The amp is built with a single 6SN7 driving one 2A3 per side, it is single ended.
 
I am not currently looking for information on upgrades to my design, I just need help getting rid of the hum. I have tried searching for ground loops, changing resistors, everything I know to try, but I haven't been able to solve it.
 
 
 
Here is my schematic, yes it is pretty simple, and believe me, I am thoroughly embarrassed to be seeking help on such a simple project

 
 
 
Here is my grounding scheme
 

 

 

 
 
HERE ARE BUILD PICS, THE CYAN REPRESENTS THE GROUND LINES IN THE BOTTOM 2

 

 

 

 

 

 
 
 
Any help is greatly appreciated. I love you all. Thanks!

 

[FrankCooter]

Don't be embarrassed. Anybody that builds tube amps has faced this problem.
 
Let's start with a few basic observations.
 
Make sure you don't have a ground loop.
 
If the hum is different from channel to channel, it is probably in the amp.  Try moving the wiring gently around with a small piece of wood as the amp runs. If the hum changes, you probably have a wiring issue. As you have done, ground the grids one by one (do NOT ground a grid if it is directly connected to the plate of the tube before it), starting with the first stage and working your way to the output stage. If the hum does not change at all, and is equal in both channels, you are probably dealing with a power supply issue.
 
Power supply hum issues are best checked with a scope. How much AC is riding on your DC supplies? For both B+ and DHT filament supplies, although there is no fixed limit,  shouldn't be more than a few mV. If you don't have a scope, simply carefully use your multimeter  with the scale set on AC mV. With DHT's, turn off the B+. If you still have hum, you have an issue with your filament supplies (sounds like you may have an issue here).  On indirectly heated filaments, hum can be caused by floating supplies. For a quick check, ground one side of the filament. For a permanent solution, use a real or virtual center tap and fix the filament voltage to ta portion he B+ via a voltage divider.
 
Hum in tube amps can also be caused by magnetic interaction between power supply transformers/inductors and signal transformers. Cure for this is usually distance (6'' is usually sufficient) combined with offsetting core orientation.
 
From what you've said, it sounds like a filament power supply issue. But I'd start by tightening up your layout and organizing your wiring.
 
Never forget HV safety procedures (single hand probe with one hand behind back, etc.,)
 
Keep us posted. More help is available.
 
Good luck!
 
 
 
If

 

[DutchGFX]

Thanks so much for the help!
 
The hum is almost equal in both channels. The difference is probably due to the fact that I broke an LED and am using 2 resistors to create equal voltage drop for the CCS on one side, and they are probably not quite equal, but this will be fixed once I buy a new LED.
 
I am using separate transformers for the filaments. I'm using 3.15-0-3.15 (I think) transformers into the Pete Millet supplies. I tried grounding the CT, but it didn't help. I will try that again when I get home. Grounding the - end of the DC out of the filament supplies killed the hum when I tried it a few days ago, but the amplifier then didn't really amplify, since it's a DHT, the grounded end cause the cathode to be raised only 2.5 V which is surely too low for operation. The sound came through but it was quiet, as if no amplifier was present. I will try this again later when I get home.
 
It seems from this that the issue is with the filament supply, thought I am not positive (no pun intended). I will also test the AC from the filament supplies. The B+ disconnected doesn't change the hum, so it must be an issue with the filaments.
 
I don't see how the Millet supply could be the issue, since he is such a well regarded designer, but if there is significant AC signal from the supply, I guess I'll try to find another supply. The regulator is maybe getting too hot, since 6.3V*1.41 = 8.8V, so it drops 6.3V at 2.5A which is near 16 watts, so maybe that is the issue? I will try using only 1/2 of the transformer, so I will use one input as the CT and one as one of the secondaries, which will give me only 5W dropped.

 

[dsavitsk]

using 2 resistors to create equal voltage drop for the CCS on one side, and they are probably not quite equal, but this will be fixed once I buy a new LED.

Replacing the LED in a CCS with resistors will likely result in noise, possibly hum. I doubt it is the primary culprit, but you should bypass the resistor with a cap if you are not going to use an LED. You can also use a string of diodes, or a zener, if you have either around.

I think I'd try disconnecting the driver stage and just work on getting the output stage quiet. DHTs can be very tricky to work with.

Grounding problems, and the proximity of transformers to other things (other transformers and tubes) are the most likely causes of hum. If you feel like a drive to up to Hartford, I am happy to help.

 

[DutchGFX]

I have now found a hum in the right channel only. I will try swapping tubes later and testing.
 
This is in addition to the hum present in both channels

 

[DutchGFX]

So, I have disconnected the filaments, and I STILL get hum. I think I oriented my transformers correctly, so I really don't know what to do at this point.

 

[DutchGFX]

With B+ and Filaments disconnected, when I touch Grid or Filaments, the him goes away. However, when I connected grid to ground, the hum remained if I took my hand off.

 

[dsavitsk]

Can you post pictures of the other side of the boards?

I don't think you are a couple of quick fixes away from getting this right -- I think you are going to have to rebuild, piece by piece, from the bottom up. I think I'd go out and pick up a couple of EL84s and try to build with those first to eliminate a couple of variables.

 

[StanD]

With B+ and Filaments disconnected, when I touch Grid or Filaments, the him goes away. However, when I connected grid to ground, the hum remained if I took my hand off.

Looks like a ground loop or unexpected path to the signal line. Without power how is there enough gain to drive hum to your speakers/headphones? The circuit should be off.

 

[DutchGFX]

I think I fixed most of it:
 
The buzz goes away when I touch the ground wires, which isn't the end of the world if I have touch the amp when I want to listen I suppose.
 
The HUM is still present, but was significantly reduced when I moved the power transformer. (I actually just disconnected the transformer and connected an equivalent that wasn't attached to the chassis so I could move it.
 
I can live with a little bit of hum, it is definitely tolerable. I haven't re attached the B+ yet since when I tried I got a bunch of sparks, probably from caps discharging, so I am waiting a little while for them to drain before soldering and actually testing. I can add pictures in a minute

 

[FrankCooter]

Those charged HV filter caps are potentially LETHAL!
 
I totally agree with Doug that you need to start this project over from scratch.
 
Before you restart it you need to get a better grasp of proper wiring technique and get a solid, reflexive, understanding of HV safety procedures.
 
To start with, your filter caps should always have bleeder resistors so that they drain themselves. And even with bleeder resistors, and the amp disconnected from all power, you should always check power supply caps with a meter to make sure they are discharged before attempting ANY electrical work.

 

[StanD]

  I think I fixed most of it:
 
The buzz goes away when I touch the ground wires, which isn't the end of the world if I have touch the amp when I want to listen I suppose.
 
The HUM is still present, but was significantly reduced when I moved the power transformer. (I actually just disconnected the transformer and connected an equivalent that wasn't attached to the chassis so I could move it.
 
I can live with a little bit of hum, it is definitely tolerable. I haven't re attached the B+ yet since when I tried I got a bunch of sparks, probably from caps discharging, so I am waiting a little while for them to drain before soldering and actually testing. I can add pictures in a minute

You didn't fix anything. Not yet.

 

[awptickes]

OK, first thing you need to do is stop touching things while the amp is plugged in. You're going to wake up across the room or wind up killing yourself. HV isn't something to play around with. It's not worth dying.
 
My background first so you understand where I come from. I've been an amateur radio operator for many years, and I repair tube RF amps, and I've built a couple guitar amps for friends. Many of these amplifiers range from a couple hundred watts to several kilowatts. Some have B+ voltages in excess of 3kV with at least 2A of current -- so instant death. You could easily stop your heart by bumping the grid connection with your finger.
 
On your project:
I see several ground loops. I think this is because you're using multi-point grounding on what appears to be MDF. Instead of daisy-chaining your grounds, tie them all to one place. Most projects accomplish grounding by using a common-chassis ground.
 
Every component's ground should run directly to the chassis ground, even if you chose to use MDF, just make sure you don't have anything daisy chained. It needs to be single-point.
 
I don't see an earth ground connection on your project. I also don't see any fuses.
 
I don't see any bypass capacitors on the grounds either, that could help your noise issue by reducing some of the harmonics.
 
 
Does the hum start immediately? Or does it take a while to build (meaning a while after the tubes are warm)?
 
 
 
Your project looks good, you've got a good start, tubes are simple once you get the basics.

 

[DutchGFX]

Is it possible you could take the image in thread and edit it to show me the ground loops? I thought since I used multiple star grounding I would have no issues.

I will add an earth connection when I get home and see if it helps.

I have bypass caps on my input tube, but didn't add them to the output because I wanted to decrease gain and eliminate capacitor noise.

I thought my ground scheme worked but maybe not...
Thanks so much for your help!

 

[awptickes]

Yeah, you need an earth. That will probably fix it, but you shouldn't stop there -- take the initiative to build a really killer amp. You've taken the first few steps and gotten yourself a working (but unsafe) amplifier. Now just make it safer by adding fusing, earth grounding, a good ground-plane, and a pretty case. Everyone likes pretty cases. Lol.
 
Your multiple-star grounding system is what's causing the ground loop. What happens is there's return on one spoke and it effects everything on that branch on its way to the hub. If everything is tied to a common ground, there's a direct path to ground, rather than your transistors feeding their return back to your tubes (where your loop most logically lies.) If you tie that second branch directly to your hub, it'll become a spoke, and your issue will probably be unnoticeable.
 
Basically, everywhere you're making a multiple-star, you're introducing a ground loop.