Interstage curious

[CurtZHP]

Making some progress here....

Instead of unracking and opening up my G9, I found another gadget I built a couple years ago that has a new-ish 12AU7; and that was a lot easier to get to.

Popped that into the first two stages of this project.  It's much quieter and better behaved.  Dead quiet with the gain all the way off, so that tells me the second stage and final push-pull stage are quiet.  Turning up the gain a bit with a mic plugged in, it's hard to tell if any noise I'm hearing is the circuit or just room "tone" picked up by the mic.  (It's amazing how noisy the world is...)

Also, I've been monitoring through either a Shure headphone amp or a Mackie 1202 mixer, and who knows how much noise either of them is introducing at this point?

 

[mjrippe]

Make an input shorting plug with a 150 ohm (or 220, or something close) resistor across pins 2 and 3 of the xlr.  Then you can plug it in and crank the gain without hearing room noise.

 

[CurtZHP]

Make an input shorting plug with a 150 ohm (or 220, or something close) resistor across pins 2 and 3 of the xlr.  Then you can plug it in and crank the gain without hearing room noise.

Any hazard to phantom power while doing that?

 

[CurtZHP]

Speaking of phantom power....

I read here a while back about doing phantom power a little differently.  Instead of feeding it through a pair of 6.81K resistors to pins 2 and 3 of the input, it was suggested to feed it through a single 3.4K resistor to the center tap of the input transformer primary.  I assume an LS10X will tolerate this.

What are the pros and cons?

 

[emrr]

Both are appropriate by the book methods. 

The load resistor: think about it, no place for the phantom to go.  You can use a pad barrel, same thing. 

 

[CurtZHP]

Both are appropriate by the book methods. 

The load resistor: think about it, no place for the phantom to go.  You can use a pad barrel, same thing.

Looking to clean things up a little more, so the single resistor into the center tap appeals to me.

Along those lines, I also took out the polarity switch.  Just more hardware in the signal path to break, especially considering how rarely I'd use it.  Still on the fence about the input secondary series/parallel switch.

Getting ahead of myself, thinking about bells and whistles.

 

[CurtZHP]

Spent the better part of the weekend putting this pile of guts into a proper chassis (hoping that will solve at least a couple issues...).

Got a nice used 2RU chassis with plenty of room.  Got everything in there nicely.  Everything in the audio circuit except the transformers is on a piece of perf board.  The power supply PCB, as mentioned earlier, is a repeat of an earlier build.

Had a bit of hum when I first powered it up, so took a closer look at the wiring.  Made a couple changes regarding shielding that seemed to help, but still had some hum.

Here's how it looks so far....
The input XLR has pin 1 soldered directly to the shield lug, so the pin 1 connects directly to the chassis right at the connector.  A length of Belden 8541 shielded cable connects that XLR connector to the primary of the input transformer.  The drain wire for this run is only connected at the XLR, NOT the transformer.  The shield of all three transformers is also connected to the chassis, simply by virtue of the fact that they are bolted to the chassis to keep them from moving.  I ohmed this out to confirm.

The secondary of the input transformer uses another length of 8451 to go to the circuit board, but the shield is not connected at either end.  Trying it connected didn't make any difference.

The gain pot is connected to the board using 8451, with the drain wire providing the signal ground, the black providing the wiper, and the red providing the top connection.

More 8451 makes the connection to the primary of the interstage transformer, again not shielded, and again, with no change either way.  I may go ahead with connecting these shields anyway.

The secondary of the IS transformer is connected to the input of the PP stage using 8451, with the drain wire as the ground connection (secondary center tap).

Since the output of the PP stage is also carrying the plate supply, I have two heavier gauge wires going from the board to the primary of the output transformer.  I figured I should run these twisted together, since that's what I did with the heater wires to both tubes.

From the output secondary, another run of 8451 goes to the output jack.  As with the input, the shield is only connected at the connector end.

The three separate rails from the power supply go to a terminal strip right next to the supply PCB.  From there, the 0V ground, B+ supply, phantom power, and phantom ground are accessed.  I have a twisted pair feeding the heaters from there.  A single ground wire and one B+ wire go to the board, and another single wire feeds the B+ to the center tap of the output transformer primary (for the PP tubes).

From the supply PCB, the power ground connects to a chassis lug along with the mains ground.


Still getting a slight hum.  If I'm listening when I unplug it, the hum disappears even as the unit still passes signal until the supply caps run out.  So I started to take a closer look at the power supply transformers.  They are both toroidals.  They are situated at the opposite end of the chassis from the input stage.  The one thing I noticed was that the B+ output from the supply was awfully close to the second transformer (the one that provides the step-up for the B+).  Just rotating the transformer seemed to help a little.

I got a little more help by simply unbolting the transformer from the floor of the chassis and placing it vertically against the side of the chassis.  I've yet to try this with the other power transformer, but I can't see that it would hurt.  I'm also figuring on moving the terminal strip, since I parked it right near those transformers.

The odd thing is that my last build uses the same supply, and in a much tighter package; and did not give me this same problem.  On that one, the power transformers are stacked on top of each other in the bottom (horizontal), and the PCB is mounted over them in a vertical position, right smack in the middle of all the audio hardware.  (That gadget is not in an RU chassis...)

It's looking a little better, but not quite there yet.  Trying right now to hunt down any ground loops I've either missed or created.

Phantom power isn't even connected at this time.

 

[DaveP]

If this helps...............

I always mount my toroids vertically.

The shield layer should go to a chassis earth (they are a chassis extension) not the audio earth.

I always use DC for heaters.

Good luck

DaveP

 

[CurtZHP]

If this helps...............

I always mount my toroids vertically.

The shield layer should go to a chassis earth (they are a chassis extension) not the audio earth.

I always use DC for heaters.

Good luck

DaveP

Yes, DC heaters here as well.

 

[emrr]

I can never tell without hands on unfortunately.  Dave’s advice is good. Every situation is different. 

Unrelated that did jump out is the 8451 everywhere.  At the end of the day if it’s not reducing interference, replace it with bare twisted (or not) wire and see if it sounds or measures better.  That shield makes a capacitor that can take out some treble when it’s a hi-Z path, cause a resonance, etc. Not a problem when low-Z. 

 

[CurtZHP]

Unrelated that did jump out is the 8451 everywhere.  At the end of the day if it’s not reducing interference, replace it with bare twisted (or not) wire and see if it sounds or measures better.  That shield makes a capacitor that can take out some treble when it’s a hi-Z path, cause a resonance, etc. Not a problem when low-Z.

It's what I had, so in most cases it was a convenience thing.  I figured any connections between the transformers and the rest of the circuit should be shielded, but as I mentioned, it doesn't seem to matter in some instances.  The input and output certainly benefit from it, but the rest don't appear to.

 

[CurtZHP]

A stupid question regarding phantom power via the input transformer....
(and I suspect I already know the answer...)
I still need to connect phantom ground to Pin 1 of the input XLR, right?

 

[DaveP]

This will have the phantom info you need:-

https://sound-au.com/project96.htm

DaveP

 

[CurtZHP]

I managed to finally get rid of most, if not all, of the hum.

Added additional power supply decoupling, after discovering some nasty bumps on the supply rails.

Now I have a new problem.  (Which may have been there all along....)

I put a scope on the output (between pin 2 and pin 1), with the gain all the way down.  There's a nasty almost-sine wave on the output that's measuring about 2V peak-to-peak!  According to the scope, the period of this wave is 50uS, which, if the math is right, translates to 20kHz???

 

[CurtZHP]

Tracked the 20kHz schmutz to the primary of the output transformer.  At least it doesn't show up anywhere else.

It's not on the B+ supply rail to that stage.  That rail is ruler flat going into the primary center tap.  I guess that transformer is doing something nasty to it?  Or is the tube oscillating?

 

[DaveP]

You may have so much gain it's forming a feedback circuit via your wiring capacitance.  Try removing C3 and/or fit 1k grid stoppers.

DaveP

 

[CurtZHP]

I think I've made some headway here....

There's got to be a catch.  I must have done something here that's going to cost me in terms of performance.

Following DaveP's advice somewhat (thanks, Dave!), I inserted grid stopper resistors in series with the grids in the push-pull stage, since that stage is where this HF oscillation was rearing its ugly head.  The oscillation is GONE!  The hum, near as I can tell, is GONE!  And believe it or not, the thing still passes a signal!

In a departure from Dave's advice, I used 68K resistors instead of 1K (they were handy).  I will try some lower values over the weekend to see what I can get away with.  By the way, should these resistors be closely matched?


EmRR....
I attempted to address your concerns with the overabundance of shielded cable for the interconnects.  The input-secondary-to-perf-board connection and the perf-board-to-interstage-primary connection were both replaced with just plain old twisted pairs of wire.  All is well.  I went ahead and did likewise with the interstage-secondary-to-perf-board connection, and everything went straight to Antarctica from there!  The thing sounded like my grandfather after a plate of beans, only louder.  Replaced that one with shielded cable and all is well again.

Incidentally, the phantom power via input primary test went very well, except for the toggle switch crapping out.  (Found it in a parts box.  Will buy new anyway...)

I may very well be on my way to finishing this beast!

 

[CurtZHP]

"Pride goes before destruction." - Proverbs 16:18

So much for making headway and thinking I was well on my way to finishing this project.

It's still not even close to right.  There's still a low level hum whenever a mic is connected.  It's slightly worse when using phantom power with a condenser.  I've tried different mics, and different mic cables.  No difference.  And it doesn't matter how I run the phantom power, whether via the input transformer or the other way (pair of 6.81K resistors right to the signal pins on the input XLR).

I had added a "series/parallel selector" switch to the secondary of the input transformer.  It works, but the hum is still there, and it doesn't seem to matter if the connections between the winding and the switch are shielded or not.

I replaced the tube for the first two stages with a much newer one from a known working device.  Didn't matter.

Oh, by the way, the hum actually gets worse when I put the top on the chassis!

If I remove the mic and place a 100R resistor across pins 2 and 3 of the input, the hum goes away it seems.  But then I'm treated to a significant hiss or white noise.  This is basically background noise.  I can hear whatever the mic picks up (when the mic is plugged in) well over it.  Turning up the gain makes the noise louder.  I shorted out the primary of the interstage, and the noise disappears completely.  So, it's got to be coming from one of the first two stages.  I guess I should assume that it's the first stage, since turning up the gain makes it louder, and the gain pot is between those stages.

I've tried just about every grounding and/or shielding scheme I can think of, and I just can't get this thing to be quiet.  I've decoupled the daylights out of the power supply rails.  What cracks me up is that about 75% of this thing is nearly identical to the last tube preamp I built, and that one didn't give me nearly as much trouble.  And this time I had that experience to draw on.  This thing simply refuses to cooperate.

I'm trying to figure out how I could go from having no hum to having it all the time in the space of 24 hours.  The only thing I can think of is that I'm working on this in two different places, depending on what sort of free time I have.  OK, so I'm working on it in two different places.  So what?  If the thing is only going to work in one place, there's not much point in building it.  Everything else I've ever built works everywhere it's used.

I'd hate to just give up, but I'm afraid this thing is turning into a waste of time.

What do you guys do when you get to this point?

 

[emrr]

the hum actually gets worse when I put the top on the chassis!

That's PSU and chassis related. The lid is redirecting the PT flux. 

What else changed in the last 24?  Besides moving spaces?  Distrust the outlet you are using now.  Or a light dimmer.  Etc. 

Take the PT outside and at a distance, run the AC to the PSU through 3 feet of wire and see what happens.  Or skip that and use an outboard PSU. 

What's the chassis?  Should you trust it?  I've seen chassis that proved to be the problem, no good continuity between any panels. 

 

[DaveP]

What frequency is the hum?  This gives a clue.

If it's 60Hz then it could be a hum loop.

If it's  180Hz then you have radiation from your transformers.

Think guitar pick-ups.............coils/transformers  pick up whatever is available, this is why good input transformers have mumetal cases.  Try fitting a temporary earthed steel screen in various places to see what changes.

Gain: do you have more than you need?

Is there a faulty electrolytic cap anywhere?  When we make stuff like this, we often use old parts just to try the circuit out.

Don't give up, come back to it later with a fresh mind!

Post some pics of your layout.

DaveP