Interstage curious

[CurtZHP]

Got the thing back together on the bench.  Something has got to be horribly wrong here.  No matter what I do, there's a constant hum on the output.  Doesn't seem to matter how I ground things either.

I even went so far as to connect an output jack to various stages to hear where the hum first appears.  No matter where I am in the circuit, it hums as long as the power is on.  So, you'd think it was something to do with the power supply.  But it doesn't seem to matter how I run the power ground or how I ground anything else for that matter.

The hum is loudest when the gain setting is all the way down.  Turn up the gain, and the hum fades into the background and mic audio can be heard easily over it.  But it's still there.

Right now, I have all the 0V points connected together, and I have the mic input shield connected to the "chassis" along with the mains ground.  Hums like crazy.

 

[emrr]

If loudest with gain all the way down, I'd suspect either the pot wiring, or the pot ground position being dirty enough to inject junk into the grid. 

 

[CurtZHP]

If loudest with gain all the way down, I'd suspect either the pot wiring, or the pot ground position being dirty enough to inject junk into the grid.

For the pot wiring, I used a piece of shielded balanced audio cable (similar to Belkin 8451).  Red from first stage coupling cap to "top" of pot.  Black from pot wiper to second stage grid.  Shield (drain wire) from "bottom" of pot to signal ground. 

I'll double check the wiring next time I'm at the bench.

 

[CurtZHP]

Checked the wiring.  Nothing jumped out at me as odd.  Replaced the pot, on the off chance it was defective.  No difference.

Took out the input transformer and connected right to the first stage, just to eliminate the possibility of something not being right.  No difference.

I'm beginning to suspect a defect in the power supply.  For the life of me I cannot find anywhere where there might be a ground loop.

 

[emrr]

Try the supply with something else?  And try a different supply?

 

[CurtZHP]

Try the supply with something else?  And try a different supply?

I'll have to build something else to try.  (Shouldn't take long...)
The only other supply like this that I have is currently in another working preamp.  Might just have to borrow it.

 

[CurtZHP]

In the meantime, a stupid question....

Poking around with the voltmeter, I made the following observation, which may or may not be totally expected.

The B+ output from the supply, with the circuit connected is 254VDC.  The voltage measured at the plates of the first two stages is around 105VDC.  Sure, there's going to be a voltage drop across those 56K resistors, but 140V??

 

[DaveP]

This is how you check this kind of thing out.

First ohms law..........140/56= 2.5mA          OK

Second check the data sheet for that tube to see if 2.5mA is acceptable

Better still draw the loadline on the sheet to check you have a sweetspot where the  +ve and -ve movements on the line are the same.

Valve wizard site is very handy for setting up tubes
http://www.valvewizard.co.uk/

By the way, your cathode follower resistors should have been at least 10x higher, lucky you didn't melt something!

DaveP

 

[CurtZHP]

This is how you check this kind of thing out.

First ohms law..........140/56= 2.5mA          OK

Second check the data sheet for that tube to see if 2.5mA is acceptable

Thanks.  Caught that the second time around when my math was wrong on the first try.
Clearly I need to step back from this thing for a bit.  I'm starting to make stupid mistakes out of frustration with it.

I did do some poking around the power supply itself with the scope, and I have a feeling there's something up with the B+ rail.  The output of the regulator shows a pretty nasty sawtooth wave (that I'm not seeing on the filament supply regulator...).  The downstream filter caps seem to be smoothing it out, but still leaving a lot of grunge behind.  Adding another cap on the end didn't help at all.

Since this supply design clearly works well on another device, it's got to be something wrong with this particular build.

 

[CurtZHP]

Shaping up to look more like the power supply.

I whipped up a copy of a known working circuit (my last tube build), and got the same result.  Hum when the gain is all the way down, noise wherever else you put it.

Now, short of building a whole new supply, I need to figure out what's broken on this one. 

 

[CurtZHP]

So, here's a couple questions as I bird-dog this issue with the power supply....

How critical is it to keep the B+ ground and filament supply ground separate from each other?  I noticed while looking over the schematic for the Gyraf G9 that the "-" side of the B+ rectifier and the heater rectifier are separate.  (Obviously, everything ends up grounded to the chassis eventually...)  But on my last build (which uses this same power supply) they are connected.  No problems with the last build, so maybe I was getting away with murder??

Also, on this one, the primary of T102 is wired in parallel (similar to the G9), while this same primary was wired in series the last time I used this supply.  I can't think of any reason why this would contribute to a noise or hum problem, but am I missing something?

(Reattaching schematic so folks don't have to search for it upthread...)

 

[DaveP]

The best grounding system is still the RCA method:-

RCA Grounding Instructions
Any amplifier having two or more stages should be constructed with a straight-line layout so that maximum separation is provided between  the signal input and output circuits and terminals.
Power-supply connections, particularly those carrying ac, should be isolated as far as possible from signal connections, especially from the input connection. Signal-carrying conductors, even when shielded, should not be cabled together with power-supply conductors.
Internal wiring for ac-operated tube heaters, switches, pilot-light sockets, and other devices, should be twisted and placed flat against the chassis.
All connections to the ground side of the circuit in each unit should be made to a common bus of heavy wire. This bus should be connected to the chassis only at the point of minimum signal voltage, i.e., at the signal-input terminal of the unit.

The ground/earth wire from the wall socket should go straight to the chassis, then the busbar with all the earths connected goes to the input socket as above

DaveP

 

[CurtZHP]

The best grounding system is still the RCA method:-

RCA Grounding Instructions
Any amplifier having two or more stages should be constructed with a straight-line layout so that maximum separation is provided between  the signal input and output circuits and terminals.
Power-supply connections, particularly those carrying ac, should be isolated as far as possible from signal connections, especially from the input connection. Signal-carrying conductors, even when shielded, should not be cabled together with power-supply conductors.
Internal wiring for ac-operated tube heaters, switches, pilot-light sockets, and other devices, should be twisted and placed flat against the chassis.
All connections to the ground side of the circuit in each unit should be made to a common bus of heavy wire. This bus should be connected to the chassis only at the point of minimum signal voltage, i.e., at the signal-input terminal of the unit.

The ground/earth wire from the wall socket should go straight to the chassis, then the busbar with all the earths connected goes to the input socket as above

DaveP

Did most of that.  There's no "chassis" to speak of yet, as I'm still in the development stage.  The heaters are DC.  I'm trying to determine if the heater ground and B+ ground can be the same.

 

[emrr]

No problems here sharing A/B DC ground connection yet.

 

[CurtZHP]

No problems here sharing A/B DC ground connection yet.

Confirming my doubts, as it were, about that being the issue.  I got away with it the last time, so scratch that off the list of suspects.

Last night, while mulling this over, I recalled that at some point during the initial construction I managed to kill one of the supply resistors.  Either R103 or 104.  I can't remember which, but I replaced both with beefier ones.  Probably damaged the regulator.

 

[CurtZHP]

Naturally, my local supplier has no TL783's on the shelf, so I have to order them.

Just as well.  I need a break from this thing. 

 

[CurtZHP]

Took a break from this monster to enjoy Christmas with the family.

Yesterday, I dove back in.  Replaced the TL783 regulator on the B+ supply, thinking it was defective.  Probably wasn't.

While not as nasty as before, there is still a hum present when it's passing signal.  The good news is that it no longer hums with the gain all the way off.  I did clean up the wiring a little more, so that likely helped.  I also put back the additional B+ decoupling.  Turns out I need that.

I tried feeding the B+ for the final push-pull stage from a point downstream of that additional decoupling, and directly from the B+ supply.  No difference either way.

The hum must be coming from the first stage.  I've tried all manner of grounding to no avail.  Still a low hum under the signal.

The only thing I haven't tried so far, short of installing this whole mess in a metal enclosure, is shielding the tubes themselves.  What's the story on that?

By the way, the tube for the first two stages is microphonic as heck!  If I get close enough to it, I almost don't need a microphone!
I've heard that's to be expected with a 12AU7, but the other one is nowhere near as bad.

 

[CurtZHP]

More tinkering has isolated the offending tube.  One of them definitely is way more microphonic than the other, which leads me to believe it may be causing the hum issues as well.

Both tubes are old, salvaged from somewhere a while back.  Who knows what shape they're really in.  The "good" one is a Mullard, and the noisy one is an Arcturus.  I'll have to borrow a couple known good tubes from my G9 and try those.

The power supply doesn't seem to be the problem.  It's a proven design.
Grounding doesn't seem to be an issue.  I've done everything I'm supposed to do, and tried several other things, none of which have made a difference.
Layout doesn't seem to be the culprit, since I've tried several possibilities there.
So, let's eliminate the tubes as suspects.


(A rather humorous observation:  At one point, I unplugged it while listening on headphones, and I swear it sounded like a bathtub draining!)

 

[emrr]

Arcturus; REALLY old tube......

 

[CurtZHP]

Arcturus; REALLY old tube......

Yeah, the box I found it in (likely the original package) loses a piece every time I handle it.  The thing is starting to make funny little pinging noises occasionally without me even touching it.  The poor old thing is probably dying, and my tinkering isn't helping matters.

The tubes in my G9 were ordered from Tube Depot specifically for that project, so they're practically new compared to these.