• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Troubleshooting First Tube Project

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I just completed construction of my first tube amp. I got the project from an old issue of Glass Audio, 4/00, and the amp is a 20W designed by Joseph Norwood Still. The amp works and plays music, but I have a 60Hz hum in the speakers. Looking with a DMM, I am seeing about 60mV RMS in one channel and 40mV RMS in the other. I was very carefull in wire routing and grounding to avoid ground loops. I tried repositioning some circuits to keep them as far away as possible from the 60Hz lines, including the heater lines, which are twisted. In looking for the source of the hum, I found that there was about 5V RMS of ripple on the 390V from the PSU. At first I thought it might be the 100uf cap, but now I'm not so certain. When I first power the amp on, the ripple on the 390V line is 40mV or less, but after things begin to warm up, the hum shows up, and then the 5V ripple is present on the 390V line. I have a feeling that this is what is causing the hum in the speakers but I am not sure what the exact source is to fix. Anyone have any input. I have attached the scematics from Glass for both the amplifier, and the PSU. I did not use the source selector or the balance control, and have the amp hardwired for 4 ohm speaker impedance with no NFB.
 

Attachments

  • diyaudio picts.jpg
    diyaudio picts.jpg
    81.1 KB · Views: 573

PRR

Member
Joined 2003
Paid Member
Watch the voltage on the 6L6's common cathode resistor: is it rising? I think 25V-30V would be right. Voltage on pin 5 of each 6L6 must be "zero" +/-1V. A leaky tube or cap could be sending it much higher, meaning more current, more ripple (and also short tube life).

Otherwise: suspect the first 100uFd power cap. Modern caps don't fail often, but it is possible.
 
OK, 5V of ripple divided by 15 (plate resistance / transformer impedance) divided by 22 (2K : 4 Ohm) divided by 10 (output tubes matched to 10%) gives about 1.5 mV to speaker. So it's good enough for PENTODE mode. UL will be about 10X worse, since Rp is 10X lower in triode. Maybe more capacitance would help in UL, add more in parallel with C1 and see what happens.

If THAT doesn't help, disconnect the transformer center tap from ground and tie it directly to C1 (-) - add an extension wire if necessary... if hum improves, your grounding scheme needs improvement...

Second, add a 10K resistor in series with your 82K bleeder, forming a voltage divider giving about 50V. Add a cap - at least .47 uF - from the 50V point to ground. Now disconnect the filament center tap from ground and tie it to this 50V point..
 
Guys, thanks for the advice. It's great to have knowledgeable folks like you willing to share your experiences and skills.

PRR, you're right, voltage across the common cathode resistor is about 31Vdc. I don't notice it changing once the amp gets to a point of being able to reproduce music. It starts at zero then as the amp comes to life, it begins to climb, settling at 30.9Vdc, with the hum present. These resistors are 240ohm as opposed to the 250ohm as I couldn't locate 250s when I purchase the parts at the time. Would this make that much of a difference?

Pin5 shows zero volts, as you expected. I tried a new cap on for the first 100uf with no change in behavior. I also tried additional capacitance, but again, no change in the hum.

To Tom's suggestions, I tried moving the center tap to the negative of C1, with no improvement. Just to further follow your suggestion of grounding, I tried to further improve the grounding by reconfiguring some grounds, again with no success, and no change in the ripple output of the amp or the PSU.

I have not tried the 10k divider trick yet, as time has not yet permitted. Forgive for asking a dumb question, but I am not sure what the thinking is. Can you elaborate?

A couple of other notes. The 6L6s were purchased matched Quad. I have not yet tried different posistions in the amp. Would that make a difference?
Total PSU current draw on the 390V is 250mA, if that lends any clues.
 
If it's 60Hz hum, it's probably not a ripple issue- ripple hum is 120Hz and can sound a bit buzzy. 5V of ripple on the B+ feeding the output stage is pretty low, lower than I would expect for a 100u cap and as much current as you're drawing. In any case, the ripple presented to the output stage is common mode, so that would be unlikely as a source of your problem.

More likely it is either a ground loop or hum pickup from heater wiring or from transformer radiation. First area I would look is around the input tube.
 

PRR

Member
Joined 2003
Paid Member
> resistors are 240ohm as opposed to the 250ohm ... Would this make that much of a difference?

Heck, no. 6L6 were made to survive 20% resistors. 250Ω +/-20% is 200Ω to 300Ω. Even 10% tolerance allows 225Ω to 275Ω. 240Ω is exactly equal to 250Ω, for our purposes.

What SY says. Layout, grounding. Post a picture.
 
One more thing to try.

If your Huuummm is definitely 60Hz then its most lkely a noise pick up problem in the wiring. If its 120Hz then its power supply related and may be a balance of output tube currents problem.
It may be that the matched quads of 6L6 are not matched as well as they might be so one side is pulling more current which due to the common cathode resistor means the other side is drawing less. In that situation the HV ripple doesn't entirely cancel in the output tranny.

I would try separate 470R 5W cathode bias resistors on each 6L6 instead of the common 240R - you already have 2 bypass caps in parallel across the 240R, just use one across each of the 470Rs.

You can then get a feel of each tubes actual operating point by measuring all the individual cathode voltages and can optimise by selecting tubes with the closest cathode voltages for each push pull pair. Oh - and measure these voltages with NO SIGNAL.

I know from previous posts that PRR doesn't necessarily agree BUT I have found that separate cathode resistors ALWAYS gives better results than a common cathode resistor or worse still a cathode resistor common to both channels as found in some cheaper amps.
Cheers,
Ian
 
Guys, thanks for the feedback and information. Sy, thanks for the explanation of Tom's suggestion, though I have not yet had an opportunity to try it.

I have tried re-wiring the amp, with no change. I will get picts up in a day or so. The amp is constructed using three breadboards, one for each audio channel and one for the PSU. All of the grounds for each audio channel tie back to the audio breadboard in a continuous strip, and then that ground ties to the PSU ground. I tried keeping the wiring as short a possible. I have measured the potentials between grounds and have less than 0.4mV of difference between any ground point.

I looked at the hum output using an o'scope. I am seeing a 60Hz component with at 120Hz component about half the amplitude of the 60Hz. The 60Hz looks to be about 300mVpp in the loudest channel. Not sure if this can provide any clues to origin. Unfortuately, I pulled a bonehead move and blew my scope up, as it wasn't capable of handling the voltage on the PSU lines, and I didn't realize that until it poped. Realized it was time to quit at that point.

Not sure what else to try. I thought about DC heater lines, but I don't think that will make a bit of difference. Sy and Tom, pulling the heaters up, let me make sure I have something straight. I should tie the center tap of the heater winding to the 10k divider point to provide a DC bias up above ground, correct? This was the next thing I was going to try.
 
I can not say the PS is at fault. Others are giving good advice.

Try a choke in the B+ rail in-between C1 & the lead from R1. A gound loop issue can be the problem as well.

For good design practice the SS rectifiers should be a UF4007 type for lowest noise, but this type of rectifier would not improve a PS 120 Hz (60Hz mains) hum issue.
 
As previous mentioned ...brousing over the psu schematic, unless the chassis grounds are wildly adrift, it should be really tame with all that uF. Similiar circuits in the 1960's had only a fraction of the uF on the B+ rails. Have you double-checked the heater valve holder wiring that it isn't contacting ajacent lugs ? Circuits like this should be pretty reliable.
The circuit earthing of the front end does need to be away from the power end......with the low amount of nfb with a 3 stage amp you will defintely hear some noise from the LS around -60dB. Where have you placed the front end 6SN7 ?
If the output tubes were wildly out of match, using a common cathode resistor it is kind of self balancing. As cap coupling is used for o/p tubes, an easy test for hum in this type of output stage is to yank the phasesplitter tube out and then switch-on. You should only hear o/p stage noise.

richj
 
tjotto1 said:
but I have a 60Hz hum in the speakers. Looking with a DMM, I am seeing about 60mV RMS in one channel and 40mV RMS in the other. I was very carefull in wire routing and grounding to avoid ground loops. I tried repositioning some circuits to keep them as far away as possible from the 60Hz lines, including the heater lines, which are twisted. In looking for the source of the hum, I found that there was about 5V RMS of ripple on the 390V from the PSU. At first I thought it might be the 100uf cap, but now I'm not so certain. When I first power the amp on, the ripple on the 390V line is 40mV or less, but after things begin to warm up, the hum shows up, and then the 5V ripple is present on the 390V line. I have a feeling that this is what is causing the hum in the speakers but I am not sure what the exact source is to fix. Anyone have any input.

You'll always have some amount of ripple on the first stage, but in a push pull amp most of it is self cancelled in the output transformers. It's further down the chain in the preamp section you need to be careful about ripple amounts, not so much in the power stage, at least for push-pull.

Another important point to observe is how you've done your input jacks. A lot of people will make the mistake of mounting the RCA jack on a metal chassis, which grounds the jack, and then running a ground wire back to the input stage from the RCA jack, instant ground loop! If your RCA jack is isolated from chassis then you can run a ground back to the input stage, if it isn't then run just the signal wire back to the input.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.