ampex tube mixer PCB

My Am pex MX-10 developed a severe 60 cycle hum recently. I went through it with my scope and found lots of AC on DC the power rails. Star ground was good and the power transformer checked out. I replaced the selenium rectifier with diodes and I replaced the diodes on the high power portion of the PSU, I also replaced the most likely smoothing caps and checked what resistors I could in the power filter section. It didn't settle down, so I went back and checked my work and found I had damaged some of the diodes when replacing them.

I'm a novice at dealing with tube electronics and I have never dealt with point to point wiring before. I understand the basics of doing this sort of thing safely. However, working on P2P stuff is a total pain in the ass. Troubleshooting is especially hard, as it seems easy to damage components when disconnecting them to troubleshoot other components.

So, I decided to make a PCB for the power supply in my mixer. The idea is to make it much easier to troubleshoot problems and swap out components when they fail. I am trying to stick to the original design so I can use the stock power transformer, also I like the sound of the thing, so I'm not looking to change it much.

I have the PCB broken up into two parts, the HV rectifier circuit and the filter/voltage divider circuit. I haven't decided how I want to work out the heater supply. I'm thinking it would be best to keep it off of the boards.

This is my first attempt at laying out a high voltage PCB, I would appreciate any feedback on any mistakes I am making. I kept the traces at 3mm and kept the distance between traces at 3mm. This was an arbitrary decision, but I figured it couldn't hurt.

Here is the schematic I'm working from.


Here is the schematic I drew up in Eagle
rectifier


filter


Here are the PCB layouts
rectifier


filter

looks fairly clean to me... two thoughts:

1) why separate the filter from the rectifier?
2) since you have so much space, make those pads about 3 times bigger... if you ever have to rework something or replace components later, you'll be less likely to rip a pad when desoldering.

As far as the heaters go, it just seems like they were cutting corners by using one of the tube's heaters as an artificial center tap.  Replace the cap there (if you haven't already) and maybe cut that ground reference, add in a 200R trim pot to balance the reference, and reference it to 1/2 HV supply.

Or just leave the heaters be...

Thanks for the feedback.

gemini86 said:

looks fairly clean to me... two thoughts:

1) why separate the filter from the rectifier?

Click to expand...

I'm working with the free version of eagle, so I couldn't get all of the components to fit in the legal area. My plan was to combine the sections in photoshop, but there isn't a lot of space in the unit, so it seemed like a good idea to separate the sections anyway. I haven't taken the measurements, but I'm guessing I'll have to mount one PCB on top of the chassis by the transformer and one in the "guts".

gemini86 said:

As far as the heaters go, it just seems like they were cutting corners by using one of the tube's heaters as an artificial center tap.  Replace the cap there (if you haven't already) and maybe cut that ground reference, add in a 200R trim pot to balance the reference, and reference it to 1/2 HV supply.

Or just leave the heaters be...

Click to expand...

I'm gonna mount a small perfboard where the selenium rectifier was with the heater diodes, so it would be easy enough to put a trimpot there as well.

Out of curiosity, would there be any benefit to sticking a 1,000 or 2,000 uF cap across each of the places where the tubes meet the HV supply?

like this...

not really...

the wiring for the heaters should be pretty beefy, and have little to no voltage drop as it goes down the chain, so just put one (or two) large caps right by the recto.

As a side note, I just downloaded the latest revision of kicad, which came out last month... it just keeps getting better and better! If you're running a pc or linux, I strongly recommend you try it.

It goes hand in hand really well with TopoRouter... (great for cramped designs)

did you recap the mixer?

Or, if you need to lay out boards with no size or pin restrictions, try FreePCB. I've had great success with it.

http://www.freepcb.com

Peace,
Paul

It didn't settle down, so I went back and checked my work and found I had damaged some of the diodes when replacing them.

Click to expand...

I'm not trying to divert the focus of your PC design but can you be clear as to whether or not the hum went away when you replaced the damaged diodes?

Thanks again for all of the feedback.

lassoharp said:

It didn't settle down, so I went back and checked my work and found I had damaged some of the diodes when replacing them.

Click to expand...

I'm not trying to divert the focus of your PC design but can you be clear as to whether or not the hum went away when you replaced the damaged diodes?

Click to expand...

I haven't replaced the diodes I damaged yet.

I went through the mixer in this order...
Looked for obvious problems (bulging caps, cooked resistors, loose ground wires)
Checked continuity on all ground connections to chassis - checked out fine
replaced the 12AU7 - no change
swapped around the EF86's - no change
replaced CR3 (the selenium rectifier) and replaced C18 (the heater smoothing cap) at the same time - no change
scoped around on the high voltage DC and found about 30vacpp throughout the high voltage power rails
replaced CR1,2 - no change
Disconnected the power transformer - vac reads fine
Hooked power transformer back up - no change
checked R39, R40, R42, R43, R44, R55 - read fine
got frustrated
checked diodes and realized I had damaged CR1, CR 2 & CR3,1. CR3,2 was fine
walked away and decided to make PCB

Like I said, I'm a total novice when it comes to tube gear. I'm just assuming the DC rails shouldn't have that much AC on them.

I also replaced the most likely smoothing caps and checked what resistors I could in the power filter section. 

Click to expand...

C12 A-D & C13?

Usually the 1st thing I do is try and isolate what stage the hum is coming from doing the quickest easiest things first.  Turn all individual channel volume controls to zero, turn up master gain and look at hum presence in A or B or both.  If strong in both I'd suspect the B+ filtering caps.  At this point you can ground the V5/6 EF86 signal grids with jumper (pin 9) and see if hum level changes.

If hum goes away when any of individual channel vol controls are turned down I'd start back at input of humming channel(s).  C3 A-D could be suspect if hum is showing on one or two channels. Try grounding grids of  1st stage EF-86 if it's showing up in select channels.

Big question: is the hum you're getting 120Hz or 60Hz?

One thing that jumps out from your measurements: 30Vac-PP is way too high.

Peace,
Paul

Thanks for all of the input.

I replaced all of the PSU smoothing caps (C12a-d, C13) and the damaged diodes. I also replaced some of the resistors because I bunged them up with my iron while changing C12). It works like a charm now. I replaced the stock caps with a bunch of discrete caps.

DC reads pretty clean, it has some ripple, but I can't hear any noise even at full gain. I am thinking of tacking a couple of .1uf 400v caps across C12d and C13, but I'm not sure that would really help.

So, I'm curious, what would it be saying if the hum was 120Hz instead 60Hz?

Adam Smith said:

So, I'm curious, what would it be saying if the hum was 120Hz instead 60Hz?

Click to expand...

120Hz hum usually indicated a power supply problem, usually in the filter caps. 60Hz hum often indicates a grounding problem.

Note: the preceding paragraph assumes a full-wave power supply, which is the norm.

Peace,
Paul