So close and yet so far.... The "Heavy" preamp mess continues

[CurtZHP]

I thought you had them connected to separate secondaries - or was that a different thread???

Anyway, there is no way you can have two separate bridge rectifiers operating off the same winding - it will cause horrible hum.

Cheers

Ian

As it sits right now, I've got two 15VAC secondaries.  One is feeding a bridge rectifier for the filament supply.  The other feeding the voltage multi for the phantom AND the step-up transformer for the plate supply.

Now, I don't know if this matters or not, but the bridge rectifier for the filament supply and the bridge rectifier for the plate supply share a common ground. 

As I was drawing out the grounding scheme to better illustrate the problem, I came across something that might be the cause.  There was a jumper wire going from the ground of the phantom supply to the ground(s) of the tube supplies.
Two issues here:
1.  When I followed that path with a pencil on paper, it formed a loop.  (Not surprising...)
2.  It puts me back in the same boat I was earlier, when I had the multi on the same winding as the filament rectifier, which caused a nasty short.

I need to eliminate that jumper! 

 

[CurtZHP]

Hi...

May I ask what kind of power (toroidal?) transformer you use? I know that there are Swedish made Noratel that are being used for Gyraf G9's circuit, but they are terrible for audio since they have quite poor EMC.

John,
To briefly address your question...  (I plan to go back later today to read your post more carefully...)

The main transformer is a Triad VPT-30 and the plate supply transformer is an Amveco (exact model escapes me at the moment...)  I used similar transformers successfully in my G9 build.

 

[ruffrecords]

As it sits right now, I've got two 15VAC secondaries.  One is feeding a bridge rectifier for the filament supply.  The other feeding the voltage multi for the phantom AND the step-up transformer for the plate supply.

Now, I don't know if this matters or not, but the bridge rectifier for the filament supply and the bridge rectifier for the plate supply share a common ground. 

The heater 0V and the HT 0V have to be connected somewhere (I usually do it in the power supply)  because the heaters have to be a known dc voltage with respect to the cathodes; otherwise you get lots of hum. This is fine as long as they are derived from different secondaries.

To be honest,  I don't like these power supplies that transform down and then up again. There are far to many opportunities to get it wrong and in the long run I don't think it saves much. Don Audio does some nice toroids with separate HT, heater and phantom secondaries.

Cheers

Ian

 

[CurtZHP]

The heater 0V and the HT 0V have to be connected somewhere (I usually do it in the power supply)  because the heaters have to be a known dc voltage with respect to the cathodes; otherwise you get lots of hum. This is fine as long as they are derived from different secondaries.

To be honest,  I don't like these power supplies that transform down and then up again. There are far to many opportunities to get it wrong and in the long run I don't think it saves much. Don Audio does some nice toroids with separate HT, heater and phantom secondaries.

Cheers

Ian

I will certainly look into Don Audio.  Thanks for the tip!

Meanwhile, once I get home tonight, I will make sure that the only place Phantom 0V sees any other 0V is at the input XLR.
Because right now that is NOT the case!

By the way, I stumbled upon your pdf "Mixer Grounding 101."  (I assume that's you...)  Printed.  Highlighted.  Coming home with me!

 

[ruffrecords]

By the way, I stumbled upon your pdf "Mixer Grounding 101."  (I assume that's you...)  Printed.  Highlighted.  Coming home with me!

Yup, that's me.

Cheers

Ian

 

[CurtZHP]

And the hits just keep on comin'!!



So, while testing my theory regarding that stray jumper that was connecting the phantom ground to the rest of the supply ground at the power supply, it was not entirely surprising that I was still hearing hum when switching on the phantom power.

But now I've got a bigger problem.  This pile of crap tried to kill me!

While sniffing around for ground loops, I happened to put one finger on one of the pins for the primary of the output transformer, while my other hand was on the steel panel that's acting as the "chassis." 

YIKES!!  That didn't tickle!! 

I'm measuring around 300VDC on either end of the primary of the output transformer.  As an added bonus, I'm seeing it on the secondary of the input transformer too.  And just for good measure, I looked where I had grounded the screen for the gain pot wiper.  So basically, the entire ground buss for the audio circuit is somehow getting 300VDC.  I just spend the last 30 minutes going over the power supply and the audio circuit looking for shorts, solder bridges, etc.  I even checked the decoupling caps for shorts.  There is nothing there.  It's got to be a mistake in the wiring somewhere, but I can't see it.  I followed the plate supply from the power supply section, through the plate resistors, all the way to the plates, and then through the decoupling caps.  NOTHING looks wrong.

This thing hates me.  The feeling's mutual.

 

[ruffrecords]

First thing is to make absolutely certain that the chassis is connected to the mains safety earth, preferably where the mains enters the chassis.. Second thing is to make sure the HT 0V is also connected to the chassis , preferably at the chassis. Third thing to do is to disconnect the 2Uf cap that connects the output transformer - it may be the culprit.

You already know not to touch the electronics with both hands at once - the shock passes straight through your heart. It is bad enough when you touch it with just one hand A good rule with tube circuits is never touch anything with your bare hand.

Take care and keep safe.

Cheers

Ian

 

[CurtZHP]

Reminds me of the time when I was trying to fix our old electric oven.  Had to work on it "hot" in order to pinpoint the problem.  Fortunately, the drop light I was holding sort of acted as a fuse when it popped.  Boy, did the missus give me hell!

 

[ruffrecords]

Boy, did the missus give me hell!

That was probably worse than the shock 

Cheers

Ian

 

[CurtZHP]

The plot thickens......

So, tonight I removed all the audio transformers from the circuit along with the entire front panel.  So the only things on the bench were the power supply and the basic audio circuit.  Two vacuum tubes with their respective cathode networks, plate supplies, grid resistors and coupling caps.  Nothing more.

Still seeing 350V on the ground of the audio circuit.  Careful examination of the schematic and the actual circuit show that, even if any of the coupling capacitors were shorted, even if I went ahead and replaced all of them with bare wires, there is NO path to ground through any of them.  So we can rule them out.

At some point, I started examining the wires from the tube sockets to the rest of the circuit, since there are several of them in close proximity to each other.  Perhaps something was amiss there?  No.

I put the + probe of my voltmeter to the plate pins of each tube.  ~350V, as I'd expect.  (I know this is a bit high for these tubes, but I'll correct that once I get it the hell away from ground!)  For ha-ha's I put the probe to one of the cathode pins and also saw ~350VDC.  Is this normal?? 

I managed to find another 12AX7 and tried that one and got the same results.  I'd have to dig for a spare 12AU7.

Bottom line tonight:  The problem goes away when I pull both tubes, meaning I see plate supply voltage on the correct socket pins and nowhere else.  Having either of tubes in brings back the problem.  ???

 

[ruffrecords]

THe only thing not clear to be is what are you measuring with respect to? Chassis, 0V or some other point?

Cheers

Ian

 

[CurtZHP]

THe only thing not clear to be is what are you measuring with respect to? Chassis, 0V or some other point?

Cheers

Ian

Power supply ground and "chassis" ground are connected, and that's where I'm putting the "-" probe of the meter.
(I put "chassis" in quotes because this thing doesn't actually have a chassis yet.  For now, I'm connecting the power ground to the ground lug of the mains inlet.

The audio ground WAS connected to Pin 1 of the input XLR, and that was connected to the supply ground, but all that has been removed.  But this problem was very much present with all that in place.  How this thing hasn't blown the fuse is beyond me.

I should just start all over!  :-[

 

[ruffrecords]

I should just start all over!  :-[

That might not be a bad idea. It is always best to start with a definite means of implementing your grounding strategy and wire that in place first.

Cheers

Ian

 

[CurtZHP]

That might not be a bad idea. It is always best to start with a definite means of implementing your grounding strategy and wire that in place first.

Cheers

Ian

At this rate, I've taken so many things out, I'd practically be starting over to put it all back!

 

[CurtZHP]

Good news!!

I found the grounding problem.  Of course it was something stupid.  Bad connection that somehow escaped notice.  :-[
So basically, the entire audio circuit was floating.  Fixed it. 

Now, we're back to the original problem.  Hum when using phantom power. 

Observations.....
Plugging in a dynamic mic sounds fine. 
Plugging in a dynamic mic with phantom power switched on sounds fine.
Plugging in a condenser mic with phantom power on hums, but otherwise sounds normal.

I tried a different condenser mic, and it had the same problem.  I plugged those mics into another system (my console) and they behave.  So we can probably rule out a defect in the mic.  Tried different cables too.

Phantom power is grounded at the input XLR Pin 1.

 

[ruffrecords]

Check for ripple on the phantom output with a scope would be the first thing I would do - loaded and unlaoded.

Cheers

Ian

 

[CurtZHP]

Check for ripple on the phantom output with a scope would be the first thing I would do - loaded and unlaoded.

Cheers

Ian

Unloaded (which, I assumed to mean with no mic plugged in...) there's no ripple.  Loaded with a condenser mic, there's a nasty looking crooked square wave measuring about 20mV peak to peak.  And if I'm reading the scope correctly, the frequency of this grunge is about 66.6Hz.
(Told you this thing was possessed!)

 

[ruffrecords]

Unloaded (which, I assumed to mean with no mic plugged in...) there's no ripple.  Loaded with a condenser mic, there's a nasty looking crooked square wave measuring about 20mV peak to peak.  And if I'm reading the scope correctly, the frequency of this grunge is about 66.6Hz.
(Told you this thing was possessed!)

The tripler works at mains frequency so if there is any ripple on the output you would expect it to be around 60Hz. You could also check the level of the ripple at the input to the regulator, loaded and unloaded.

Cheers

Ian

 

[CurtZHP]

The tripler works at mains frequency so if there is any ripple on the output you would expect it to be around 60Hz. You could also check the level of the ripple at the input to the regulator, loaded and unloaded.

Cheers

Ian

Well, allowing for my reading of the scope, I wouldn't be surprised if it was really 60 on the nuts.

Anyway...
I measured the input to the regulator with phantom going to the mic.  Same ripple waveform (so, about 60Hz +/-) but much bigger.  About 1V peak-to-peak.  Tested it unloaded, both by switching off the phantom to that mic input or by just unplugging the mic, and the reading was the same.  1V peak-to-peak.  Sawtoothed looking square wave.

Going from 1V to 20mV seems like a pretty precipitous drop, like something is knocking that ripple down considerably; but I guess when you're dealing with mic input levels (that are going to be amplified like crazy...), 20mV is all it takes.

Two ideas of mine.....
1.  Not enough filtering caps south of the regulator.  There's only one, and it's a 47uF/100V.  That's seems rather anemic, especially since I changed all the multiplier caps to 100uF.  Maybe a couple more?  I'll have to find some, as I don't have any handy at the moment.

2.  Bad TL783.  This got me thinking back to when I had a similar problem with a solid state preamp I had built from a set of plans.  Spent months trying to hunt down this nasty HF grunge.  In fact, I bought my scope for troubleshooting this very problem.  Turned out to be a bad regulator.  It was regulating, all right; which made me think it was fine.  But it was crapping its bed the whole time!
Again, I'll have to find one to test that theory.

 

[PRR]

> I wouldn't be surprised if it was really 60 on the nuts.

There's no reason to doubt. Put your finger on the probe. That WILL be the local line frequency. If, at the same sweep rate, the unknown is the same wave-width, it's line-freq (60Hz for you). If unknown is half wave-width, it is double-freq (120Hz for you), typically from a full-wave rectifier. If not either, something very odd is going on.