120Hz hum, new caps, old tube unit, chassis-ground questions..

[systemtruck]

Hello all,
I’ve modified an old tube oscillator to become a guitar amp. Also recapped all electrolytics and audio pathway caps and large values.
It sounds terrific except there is a distinct 120Hz audio hum at a low volume. No other noise whatsoever, and no mechanical hum, just a nice clear 120 Hz in audio at various stages throughout circuit. It’s pretty low in volume, but it ruins silent moments.
This original unit is from 1955.

I tried hooking up a lantern battery to briefly test using DC heaters instead of AC, and that did nothing. I’ve also chop sticked quite a bit with zero results.
The noise is so pure that it doesn’t seem like it’s proximity to the power transformer that’s the issue. Although, it is indeed close. But, there is no buzz or hiss at all, it’s just a perfect 120 Hz. In my experience listening to transformer proximity hum, it’s usually much messier sounding and has some high buzz along with the mains sound.

So I’m assuming it’s coming from the rectified DC signal which is the only thing i can think of that would present a perfect 120Hz signal. But it doesn’t make sense to me that the signal would be so strongly AC instead of DC since the caps are all new. So I’m a little confused about that. I tried a different 5Y3 i had on hand, which seems to work fine in other equipment, but i don’t have a tube tester on hand. The results were identical. So I’m guessing it’s not the 5Y3.

So that leads me to grounding… in all of the wiring they simply used all of the tube sockets as ground points, which are of course fixed to the chassis in various places. The first three filter caps are also connected directly to the chassis. They are in a can, which was manually wired to signal ground near rectifier but also to the chassis with an extra lead.

So there are at least like 8-10 different places that the signal ground goes into the chassis.

Would a 120Hz hum make sense as a symptom from this? Or would it mean that there is also an issue with the rectification itself? I have a couple other different rectifier tubes but am thinking they put out higher voltages. The schematic here says 5AR4, which I have too, but the unit contains a 5Y3 and is screen printed 5Y3 on chassis next to tube.

I’ve labeled 4 example spots where signal ground is connected to chassis, and attached the full schematic below those…

 

[Tubetec]

Ive recapped a HP 200cd ,
Im not sure which model you have , but it looks even older ,
You'll definately need to replace the PSU caps , C12 A,B,C
Keep the cap after the rectifier the same value as spec ,
the other two can be increased in value safely to lower hum a bit ,

I typically went with x10 times the original values.

Where are you feeding in the guitar signal ?

 

[systemtruck]

Ive recapped a HP 200cd ,
Im not sure which model you have , but it looks even older ,
You'll definately need to replace the PSU caps , C12 A,B,C
Keep the cap after the rectifier the same value as spec ,
the other two can be increased in value safely to lower hum a bit ,

I typically went with x10 times the original values.

Where are you feeding in the guitar signal ?

Ok thanks for the pointers! I’ll experiment with adding some larger caps in those spots. It would be nice if that did help!
Yeah I have already recapped the whole thing except for a few of the tiny caps. Power/filter caps are all new. So the hum seems weird.

Regarding signal input, I’m actually cheating at the moment.. using an external preamp and feeding that line level into C9. Once I get this unit hum solved I will be adding a two-tube preamp circuit using the available current since the oscillator section’s two tubes will be deleted.

 

[systemtruck]

Well I tried to measure AC ripple at C12b which is the first filter cap, using my trusty old Tektronix scope and some good HV leads with clamps, and somehow the 5Y3 quickly starting turning bright red so I killed power immediately. Yeesh. Do not understand that at all. Connections were correct. Must be something about scope input. Or maybe the capacitance of cabling gave it some kind of pathway? Scary.
Anyways it all seemed ok after, so then I simply tacked on all the HV capacitance I have on hand which is a mere 80uF, to the next cap C12C, turning that from 40uF to 120uF. Did a few A/B comparisons, a little difficult since power down is required between, and sure enough there is a very distinct drop in the 120Hz tone. It does reveal some lower tones which were under it. Just by ear balling it I would guess the 120Hz tone itself drops 8-10 dB. Pretty good news. I wonder how far it would go down with 200uF.

The exposed lower tones are starting to resemble transformer buzz. That 120Hz tone that sat on top was just a perfectly round pitch.

I wonder if it makes more sense to just add capacitance to the other filter stages, or to create additional R/C filters. Is 10X really safe? There’s only 800 ohms between the first and second cap. Can I really dump that second stage into a 400uF 450V?

I also wonder if signal ground isolation from chassis will help. I suppose I could cut all of those points that are soldered to socket fasteners and create a ground bus wire with a heavy gauge lead, and come up with a star grid somehow.

 

[CJ]

Sometimes the scope ground is connected to power line ground which can cause problems, iso xfmr is the fix,

Check the dcr on your output transformer legs to the CT. The output xfmr is where all the 120 gets canceled. Check power tube matching. Maybe install 1 ohm resistor in between cathode and ground on each tube, or if you have brass balls, you can use the shunt method, connect DC ammeter from xfmr ct to each anode and read current. I have seen huge ripple on scope and no noise due to hood tube matching and a balanced output xfmr.

I use the shunt.method because I am old and if I electrocute myself I only lose like 0.01 percent of the rest of my life. And what could be bad about spending an eternity where the streets are paved with gold and the angels sing in four part harmony?And it saves the resistor hassle. You can also measure dcr of xfmr and use ohms law to figure current, but you still are gonna have probes on HV.

.

 

[systemtruck]

Hmm i have been wondering about the cancellation of the transformer.
It’s the primary winding stage that can provide some 120Hz cancellation though, as it receives the power tubes out of phase, correct?
Currently I have the bottom of the SECONDARY left dangling free since i snipped the NFB line back to cathode, which boosts gain a lot and becomes a real guitar amp. The cathode is then grounded through its 4700 resistor. It has been working great. But that leg is doing nothing. The attention circuit is providing the reference to ground for the secondary.
I’ve been unsure about the design of the transformer and a little cautious, but i suppose it is just a basic transformer and not one with tertiary / accessory windings. I will hook up the bottom leg to ground and see what that does. Should it go to chassis or to a certain signal ground? Not that that matters much since this whole unit is just a mashing of grounds.

I would love to convert this into a balanced output but I’d have to lose the attenuation, or build a matching half for other leg of secondary.

But it might be a moot point to go through all that work only to find out that the secondary, when balanced, isn’t providing any cancellation anyways and my hum is still present. Thoughts?

None the less, for now i will see what grounding that bottom leg does.

 

[systemtruck]

I connected the lower secondary leg to ground and it improved the signal to noise ratio a little bit but it also dropped the output level quite a lot. I suppose it loaded down the transformer a lot. So I put a 4700 ohm resistor to ground instead of direct to ground and it now seems to retain full volume output but also improve signal to noise ratio.
Combined with the new 120uF at C12C, the hum is very much lowered and everything is sounding quite a bit better but not perfect. The hum is still there at very low volume, and I’d love to have it as “gone” as my other amps. I will put even larger caps at C12C when I order some.
When i add a preamp circuit to this unit, it’s going to pick up whatever this hum is from and boost it. Maybe adding an extra rail filter stage down the line will be the ticket.

 

[ruffrecords]

I don't think is has been mentioned yet but heater cathode leakage in old tubes is a well known cause of hum. Lekage in old coupling caps can also seriously upset biasing which may lead to hum.

Cheers

IAn

 

[systemtruck]

I don't think is has been mentioned yet but heater cathode leakage in old tubes is a well known cause of hum. Lekage in old coupling caps can also seriously upset biasing which may lead to hum.

Cheers

IAn

Ok thanks, great to know.
On that note… how safe is it to remove the 6SN7 tube just to see if the push pull tubes hum ( or not ) without the 6SN7 installed? If the audio output still “operates” and the hum goes away, then it seems like the 6SN7 is a major culprit.
That being said, i do have another 6SN7 here, untested, and i swapped it and there was no difference. It functioned the same.

 

[Matador]

If you dial R25 all the way down to minimum (wiper at 0V), how is the hum? ?You generally need to start at the power amp and work "backwards" to the input.

R48 looks like it can cheat the bias slightly between the tubes until plate + screen currents are equal. Have you tried adjusting that?

 

[systemtruck]

If you dial R25 all the way down to minimum (wiper at 0V), how is the hum? ?You generally need to start at the power amp and work "backwards" to the input.

R48 looks like it can cheat the bias slightly between the tubes until plate + screen currents are equal. Have you tried adjusting that?

The first half, the oscillator section before R25, is killed. As is the negative feedback path from transformer, fyi. And the bottom of 4700 resistor from pin6 of V3 is grounded, and i added another 4700 resistor to ground the loose bottom leg of secondary. Those are all the mods currently.

I have external line level audio hitting C9. However, when testing for hum, i ground this point hitting C9. The hum doesn’t change when switching between grounding this point VS having line audio connected from external unit. Exactly the same sound/noise. So it’s all POST this spot, at least.

I read in service manual that R48 is for adjusting distortion. Maybe I’ll give that a shot just to see, and mark the current position. So unequal currents between power tubes can let AC ripple make it out to output? I also would like to compare the tubes, but don’t have a tube tester. Maybe it’s time.

This unit doesn’t have a “humbucker” pot for the 6.3 AC rails, but i swapped it over to DC 6 volts and everything functioned but the hum did not change whatsoever, not even the slightest. So i don’t think adding a humbucker pot on the 6.3 rails is going to do anything to remove the 120Hz.

 

[ruffrecords]

It might be worth changing C10 and C11. Old ones often leak and screw with the bias of the output stage.

Cheers

Ian

 

[systemtruck]

It might be worth changing C10 and C11. Old ones often leak and screw with the bias of the output stage.

Cheers

Ian

Yep those are already replaced. I did all elec/large caps and also all audio pathway caps. The only relevant (not in oscillator section) caps i didn’t change are C13 and C19. Maybe I’ll just do those to be sure nothing is wrong with caps. In fact C13 gets deleted by “change no. 11” in the notes at bottom of manual anyways so I’ll just chop that out.

 

[Tubetec]

You could try a 100 ohm humdinger pot in place of the two 27 ohm resistors ,
it might allow you lower noise a little further , in conjunction with the bias adjust and depending on output tube balance to some degree .

With feedback will result in better noise performance , but it will effect the feel of the amp under heavy drive ,
Ive made an NFB boost .ie you use a pot to set the maximum NFB level , then arrange a foot switch reduce the NFB for the boosted mode .
whats nice is the charachter of the amp changes , you can go from clean with an edge to a lead boost where the amp readily feedsback on any note , without any kind of pre amp drive circuit .

You could think about incorportaing a choke in the screen grid feed to the output tubes ,
in place of the 800 ohm resistor ,
but care needs to be taken with placement, orientation and screening from the mains transformer , assuming you have the extra real estate in the chassis to begin with ,

 

[Audio1Man]

About grounding problem, I would start by:

Take T2 pin 9 (the CT) and only connect it directly to the GROUND of C12
This way the CHARGE CURRENTS only go to the C12(ABC) and the DISCHARGE CURRENTS only come from C12 @ the chassis.

 

[Tubetec]

Id probably also think about a way to use the osscilator path as a tremolo , seeing as its there .

 

[systemtruck]

About grounding problem, I would start by:

Take T2 pin 9 (the CT) and only connect it directly to the GROUND of C12
This way the CHARGE CURRENTS only go to the C12(ABC) and the DISCHARGE CURRENTS only come from C12 @ the chassis.

And should the C12A/B/C ground location be considered the center point of an ideal star ground setup for whole unit?
Like this is where the mains AC cable ground wire lead should connect too?
And what about T1 ground?
And what about the speaker ground? As mentioned, I have a line to voice coil 600:8 transformer which feeds the speaker. It should be noted that there is zero difference in hum when using just line output, with the 600:8 transformer disconnected. So I don’t suspect the speaker ground (currently just audio output ground) is causing hum. But should it generally also go to the first / rectifier filter cap ground?

 

[scott2000]

So that leads me to grounding… in all of the wiring they simply used all of the tube sockets as ground points, which are of course fixed to the chassis in various places. The first three filter caps are also connected directly to the chassis.

About grounding problem, I would start by:

Take T2 pin 9 (the CT) and only connect it directly to the GROUND of C12
This way the CHARGE CURRENTS only go to the C12(ABC) and the DISCHARGE CURRENTS only come from C12 @ the chassis.

 

[ruffrecords]

I notice on your schematic there are two grounds, one called chassis ground and the other called earth ground. I notice that earth ground, which is connected to the earth pin of the mains connector does not seem to be connected anywhere else. Can you tell me where it is actually connected?

Cheers

Ian

 

[systemtruck]

The mains earth wire is mounted to a PT lug bolt, which is of course the chassis entry point for earth.
The PT 9 wire and C12 neg join together at the 5Y3 socket mount which is fixed to chassis.

FYI I deleted the front panel power switch because it didn’t work. The unit was sold as “doesn’t power up” so i got a great price, ha. So in my efforts to solve hum i figured why not delete the long AC wires running from back to front to back, for now.

The C12 caps may look messy, but it’s essentially identical to how it was from factory except they’re not in the can. Only difference is they had used a can leg as a convenient terminal for the C12 negatives to get jumpered over to the 5Y3 socket, where as i just wired the 3 negatives together and jumpered them over to the 5Y3 socket. But the factory can was loose… the 4 can legs were NOT properly soldered to chassis, so that connection could be easily broken by just jiggling the can. Thus, i didn’t worry about trying to connect C12’s neg locally to chassis, although it makes contact since the wires leans up against the chassis. So it’s effectively the same as factory. I chopsticked this to make sure it didn’t help/harm the ground and/or hum, and it has zero impact. But i could commit to one way or the other, connected locally or not, and make it isolated or soldered.

So as for the C12 negs, I’m pretty dead certain that it is connected how it was assembled in factory.

So…. Perhaps the grounding should be improved? Should there only be ONE connection to the chassis and everything from signal ground should tie back to that? As just listed, the 5Y3 socket mount brings C12 neg and PT pin 9 to chassis, and the earth mains hit unit chassis and PT case. They’re only like 3-4 inches away from each other though. Is that enough for a ground loop? I don’t know. But, on top of that, all of the tube sockets’ mounts hit the chassis and each of those have local circuit grounds connected to them. I still have the oscillator tubes active so that the overall rail voltage doesn’t rise. So that’s a whole bunch of signal-to-chassis going on around the unit which is my main reason for making this post.

Any thoughts?

Thank you!