The Phantom SR45 Build - Snow Creek Shuttle Episode 1-A Bottlehead Saga

[Paul Birkeland]

I would also suggest using a clip lead to attach the grid of the #45 to the power supply ground to see if the voltage appearing at pins 1/4 changes at all.

[ssssly]

It always amazes me how busy retired life can be.

Quick summary since my last post.

Relabeled the 4 pin according to standard convention. A1,4 Filament, 2 Plate, 3 Grid. Hopefully this reduces vice increases confusion.

Swapped a 1.2k cathode resistor in and promptly let the magic smoke out of something, immediately turned off. Created a few new bad words. Went over everything 100 times trying to figure it out. Thought maybe there was a pulled ground or ground loop. Rewired star chassis ground and changed the signal ground to chassis ground point. Retested, no smoke but crazy voltages. A(1,4) 321v, (2)358v, 3(293) vDC measured to RCA ground lug. Swapped the 1.6k back in. Still same crazy voltages. Then when poking around to try to get a voltage reading I shocked the crap out of my pinky, which was nowhere near a bare wire or terminal. The lead to the B- side of the coupling cap was touching the metal casing of the cap.

Assumed that that was what the magic smoke was let out of and sent a couple hundred volts into my pinky. Replaced it. Made sure nothing was touching the case. Fired it up up again today and...

New measurements to RCA ground
(1.6k cathode resistors, 180k PC board resistors, 4.9k Output transformer, 40H plate choke):

DRV P     (228vDC)
KRG        (14.95vDC)
REG P     (358vDC)
KDR        (7.9vDC)
B+          (358vDC)
CK          (405vDC)

A1,4        (52.3vDC)
A2           (330vDC)
A3           (1.84vDC)

B1           (358vDC)
B2           (0.03vDC)
B3           (7.9vDC)
B4           (0.003vDC)
B5           (0.004vDC)
B6           (224.8vDC)
B7           (0.004vDC)
B8           -
B9           (15.7vDC)

A1-A4      (0.003vDC, 2.7vAC)
B4-B5      (0.001vDC, 6.5vAC)

With clip lead from A3 to PC Board G: A1,4 (50.9vDC w/ dead cold tubes immediately post leds lighting)

[Paul Birkeland]

1.8k PC board resistors

I don't recall there being a 1.8K resistor on the PC boards, which part is this?

Your voltages look a lot better IMO, you have pretty close to 275V/34mA.

[ssssly]

That

I don't recall there being a 1.8K resistor on the PC boards, which part is this?

That's because I put the accent on the wrong sy llable.

180k

Looked better to me as well. Seems much happier now that the 45 and 6CM7 aren't shorted through a capacitor case.

That and I think the signal ground from the 45 was iffy before. When I rerouted it, the wire was tarnished. Used solid silver wire for the signal path and it got pushed further into the joint while soldering than I had cleaned back to solder it.  Was nice and shiney on the outside but when I desoldered it, the inside was full of nasty.

But the numbers looked close enough to me to build the second one and tweak the cathode resistor along the way.

Anything else I should be looking at that I missed?

[ssssly]

And they live.

Was able to get the 45 measurements right with the 180k resistors on the board and a 900ohm cathode resistor. But it was driving the 6cm7 into oscillation. Dropped in a pair of 160k resistors on the PC board and a 1.6k cathode resistor and everything is happy now. 248v 1.5A at the 45 plate, no oscillation in the 6cm7, making beautiful music. They do have a touch of audible hum but I think it is either a ground location issue or the long wire from the inputs picking it up. So still a bit of tinkering to do but they sound great (just have to turn the volume up a bit).

Thanks again for all the help.

[Paul Birkeland]

Do you have a hum pot on the 45 filaments?

[ssssly]

Do you have a hum pot on the 45 filaments?

I do. It is completely independent of the of the hum pot setting. I does get louder with volume but not linearly. The hum lags behind total volume by a considerable amount.

I built the amps with different input wire routing to see if they would be audibly different. They are. But they are also both too loud. So I'm going to move the signal ground so its direct to chassis at the 9 pin socket terminal strip and see what that gets me. I have fairly high hopes.

Could also just need better rf shielding than kapton that close to the output transformer. Good dielectric, but not much for shielding. And I'm fairly certain I have mylar tape someplace.

[ssssly]

Hum solved. But raised more questions.

The culprit for the loud, asymmetrical hum was the ground wire from the PCB to the IEC ground. The noisier amp had that ground wire between the PT and C-7X. I cut it a touch too short when building that amp so routed it a bit different. Rerouting it to the outside of the chassis eliminated the loud hum.

But while poking around changing and testing stuff I noticed something else. My Paramount's were picking up hum from the SR-45s. With just the Paramounts on, to get hum out of the speakers I have to intentionally set the humpot to generate hum. But with the hummy SR45s on, the Paramounts also hummed. Loud.

So it seems that the SR45s were feeding hum into the the the wall socket, that was then being picked up by the Paramounts.

Installing a pair of diodes between the IEC and chassis ground on the SR45seliminated this. But I can't help thinking there is something else going on that needs to be addressed.

Particularly that now without the hum there is a faint hiss.

They are certainly listenable now. They sound good, very good. Damn good. In fact as good or better than I've heard anything powering an Altec horn driver. They seem to roll off the highs that can make Altec horns overly bright and shrill sometimes. All previous amps I've had biamping that horn had issues with metal percussion instruments like bells, xylophones or even pianos sometimes. Where the initial attack on the instrument could be so shrill at energetic volumes it physically hurt my ears. The SR45s are revealing enough to hear both the mallet strike and full ring out of the xylophones and piano in Chick Cores's Crystal Silence without any shrillness or fatiguing qualities. The only other amp Ive had in that system that got close had glowing blue voltmeters on the front and cost more than my car. And why I rarely listened to live jazz on that system.

But something must still be amis for them to want to feed that much hum into the wall socket without blocking diodes. So still some tinkering left on the ground paths. Perhaps having the shielding ground from the AC in and heater wires tied into the signal ground too close to the tx center tap ground??? Or the safety ground bundle where it passes by the C-7X needs additional shielding???

Then there's the hiss. The character of which makes me think it's unstable capacitance somewhere. That kind of low crackly hiss you get from the blank track of a dirty record without the pops. Or a leaky cap minus the whistles. They are totally listenable with the hiss as is, it's low and obscured by the constant white noise from the stream next to my house. But I can also tell that there is better to be had by blackening the background with it's removal.

So perhaps trying bypass caps caps to see if I can track it down to an individual cap??? Or flipping the leads on non polar capacitors to see if it's related to capacitance in the internal shielding??? Or I used some Dale non inductive wirewounds for the C4S resistors instead of metal or carbon film because that's what I had at 160K---would that cause low level hiss?

[Paul Birkeland]

The problems you're experiencing likely have a solution that's more about how the amp is put together rather than a part that isn't working properly.  It's tough when an amp like this gets full of big caps, as we can't really see a lot of what's going on, and ultimately this project is custom enough that debugging procedures will end up being specific to just this amp. 

[ssssly]

Yeah. Totally.

Looking more for suggestions on methodology to track it down or start tracking it down based on experience. Not move wire from T1 to T2 to solve x problem.

If reversing or bypassing non polar caps and/or throwing shielding at ground wires/transformers doesn't work my next step, based on how I normally deal with unwanted noise in logic circuits, would be to mic the speakers to try to find a noise profile and then differentially probe each trace and component, working backwards from the output, to see where the noise is first introduced.

Or is there a better method here?

[Doc B.]

Hiss is usually noise from a tube in a low signal/high gain part of the circuit. Horns will exacerbate the issue. I would start there. Then move on to checking caps if you can't resolve it by rolling tubes.

[ssssly]

Rolling tubes doesn't affect it.

Thank you much. I shall start there.