• 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.

Problem with my amp, and I have run out of ideas

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
P.S. I really appreciate all of your FAST responses and interest. As much as this is driving me nut I am starting to remember why i use to love DIY audio.

[/IMG]

Doctorate degree drove me numb-skulled:headbash:. As a Bachelor (at the time 1971) tubes purified my ways of thinking. Keep at it.:D

As SY mentioned in post 7; a zobel network is required in the anode 1st stage to +; with an 6SN7 to start with 470pF + 10K.
Suggest the UL screen zobels typ 1nF & 1K to arrest o/p stage HF oscillation. All these components should be good quality and caps NOT ceramic types.
If you think you have the parasitic oscillation solved, that isn't good enough. Is it ? Yes/No ? It will just take an awkward load for it to re-occur. You don't want a marginally operating amp. The zobels outlined should get you going with most makes of o/p trannies......unless the nasty surprise in post 38 arises.

If you feel something repectable is working, send us a pic of a low level 1Khz square wave into a dummy load on the output and post it. It reveals alot about the status of the amp. If you don't have a signal generator, most good scopes have a 1Khz sampler on the front panel for probe adjustments. Use it VIA an pot attenuator at low level.
BTW; 85mA for starts for KT88's is a tad high at 500V. Back it to 50mA per tube whilst circuit behaves odd.
This is a good case for a variac and a separate B+ supply on the o/p stage to see how it behaves. It will also reduce the risk of ruining the screen grids by excess dissipation caused by rampant parasitic oscillation.

richy
 
Come to think of it, I'm highly tempted to fit 100 ohm resistors to the output stage screen grids to avoid straining the tubes if all isn't right. Resistors are cheaper than tubes. The o/p stage Zobels are in the right place across the UL & Anode windings. See pic regarding placement of 1st stage zobel. It's quite possible one may not need it after problem in amp has been found. example pic enclosed.
 

Attachments

  • example 1ststage Z.jpg
    example 1ststage Z.jpg
    23.5 KB · Views: 148
Well folks I have built the 2nd channel with the original schematic but with the addition of the RC network on the input stage and across the screen to anode connection on the OPT. I was able to clean up the wiring alittle this time around most notedly the heater wiring.

DSC01921.jpg


DSC01903.jpg


DSC01923.jpg
 
All the voltages test great, the stability is great, it passes a 1KHz sqaure wave super clean, it produces about 60w RMS, its 3dB cutofff is about 30KHz. So i started to listen to the setup with the JM Lab speaker in the above photo, but after about 30min of listening I developed a headache. After trying to track down the problem it has become clear that the OPT is ringing with the input signal. I replaced the speaker with a 4ohm test load and ran some tones through the input. And from about 1Khz to 20Khz I can hear the damn test tone on the OPT, and it is pretty loud. With several watts on the output I can hear the ringing OPT clear across the room. I assume there is nothing I can do about this other than replace the OPT ? If I have to replace the OPT this project is going back into the closet for another 5 years or I am trashing it.
 
Administrator
Joined 2004
Paid Member
Are the OPTS in cans or mounted out in the open? Some level of singing (or as you call it "ringing") is normal in operation at significant power levels.

I've had plenty of transformers that made audible noise when driving power resistors in a quiet room over the frequency range you mentioned. (Actually pretty much any that I can recollect that were not in a tar filled can.)

Seems like you need to improve the isolation between you and the transformer, but note also that on real music, not test tones, the amount of energy in the nuisance band you identified will be a very small % of the total being reproduced.

The headache you mentioned probably has more to do with the electrical characteristics (distortion spectrum, etc.) than it does with any noise your transformers are making. Some judicious tinkering with stage operating points, overall global feedback, and coupling capacitor choices may be in order.

And are you sure that the speaker you are using does not have ultrasonic or HF response issues that are exacerbated by the wide bandwidth of this amp. Is the speaker synergistic, i.e. a good match to the amp?

I'm assuming that everything else as you say is copacetic.. (You did take a look at the amplifier output with speaker connected, right?)

Please post some pix showing the output transformers and how they are mounted. I'm not familiar with what you've bought. Rubber pads, potting compound, tar, potting cans, etc could all be of use if we knew what we were dealing with. Two sources of these sorts of noise are magnetostriction in the core and windings that buzz due to the winding technique used.

Last of all, this is not the point where you should be giving up as frustrating as it may be.
 
Last edited:
Thanks Kev, your post has DONE alot to settle me but my thesis defense is in a few weeks and the smallest thing at this point is causing me to lose it :) I really just needed to vent alittle. I will get back to the amp when I can retain my composure.

The OPT is the big donut in the middle

DSC01881.jpg
 
Administrator
Joined 2004
Paid Member
Thanks Kev, your post has DONE alot to settle me but my thesis defense is in a few weeks and the smallest thing at this point is causing me to lose it :) I really just needed to vent alittle. I will get back to the amp when I can retain my composure.

The OPT is the big donut in the middle

DSC01881.jpg

Take care of the thesis first, that's the important thing now. Time for these amps later, and by the way the amp looks great!

I see that the OPT is "canned" - is it just sitting inside the can or is it actually encapsulated in place? If encapsulated it should not be making too much noise, if not then there is a lot you can do. Regardless I would get some sheet rubber and neatly cut a donut to place under the transformer to reduce its coupling to the chassis, this will at least reduce the likelihood of the chassis acting as a sounding board for transformer noise.
 
Administrator
Joined 2004
Paid Member
I just connected the amp to the radio and turned it up to what would be normal listening levels. I can clearly hear the music being played from the OPT. I can actually clearly make out the song from a few feet away.

Hopefully you didn't do this with no load connected.. Most all OPTs make audible noise when driven with no load connected, and this is also a significant risk to the health of your OPT.

Without hearing the transformers I can't say whether this constitutes something I would consider normal in operation
 
Administrator
Joined 2004
Paid Member
The pots are fully encapsulated, and the bottom of the encapsulation already had a rubber mat on it. These were $$$$ OPTs, i just cannot believe they aren't perfect.

How long ago were these purchased? I'd contact Plitron and see what they have to say about this issue.

No transformer is totally silent, but without hearing them I can't determine whether this is normal or constitutes abnormal/nuisance level behavior.

I can hear the OPTs quite clearly from as much a 6 ft away in some of my amps running them into resistive loads on the bench or with music also into resistive loads. Others I can't hear from a distance of a foot or less. I doubt it is the cause of your distress however.

Try triode connecting the output tubes and see what difference that makes to the sound with speakers. Triode operation results in significantly less open loop gain, and somewhat more linear behavior in the output devices, the distortion spectrum should be simpler and less extended than in UL mode. The much lower rp in triode mode should significantly damp any transformer parasitics and other bad behavior that may be occurring. The reduction in output power is ~3dB, and is a lot less significant than you probably appreciate. It would be one of the first things I would try.

Perhaps someone with more experience with toroids based OPTs and the mechanical noise levels they generate can weigh in.
 
Administrator
Joined 2004
Paid Member
Ok same conditions as before I measured 55-60dB coming from the OPT from 1ft away. The OPTs were purchased many many years ago, they were even discontinued a few years ago.

So now we have a level which sounds really excessive to me. The one other thing you need to establish is at what output power level this noise level was occurring - if at full power this might not be the issue it seems, but at 1W it is clearly far too much. I was expecting something on the order of 20 - 30dB at that distance. If possible try pink noise at an average level of 1Wrms and see how much noise you get. (Use an averaging voltmeter and aim for an average of 2Vrms across 4 ohms for several minutes and then measure the noise level)

I assume there are only 5 primary leads coming out of this thing that are clearly marked.. The secondary again is also clearly marked. Loosening or tightening the mounting bolt has no effect, spacing the transformer slight above the chassis on something like match sticks has no effect..

Once you have ascertained what the power input is for a given noise level I would contact the manufacturer and explain the circumstances and find out what if any recourse you have - if they are not helpful I would then make sure that you prominently but very accurately and factually make the issue known on every audio forum known to man.
 
Last edited:
Ill redo the noise measurement at 1w but this time with the signal generator so I have control of it.

But of more concern is even with the load resistor and not the speaker I develop a headache after only a few min of the amp being on. I can't hear anything but I think I can sense something, so it has to be something in the 18-22kHz range.
 
Administrator
Joined 2004
Paid Member
Ill redo the noise measurement at 1w but this time with the signal generator so I have control of it.

But of more concern is even with the load resistor and not the speaker I develop a headache after only a few min of the amp being on. I can't hear anything but I think I can sense something, so it has to be something in the 18-22kHz range.

I would look for ultrasonic oscillations in the output with the amplifier driven, a scope should be very revealing if such is present, and it is quite possible and maybe has something to do with singing OPT as well. I have had some similar experiences as well, mostly I'll admit with solid state gear.

This must be terribly frustrating, but at the same time it is an excellent (if expensive) learning opportunity.

I would recommend trying the pink noise source later (arta or audiotester demo software installed on a PC) and see what this gets you if nothing else proves illuminating.
 
Ok at 1.5w ( I cannot turn my generator below 50mVpp) at below 5khz i have to have my ear only the OPT to hear it, around 8-9Khz i can clearly hear it from a foot or two away.

This amp has a ton of gain and way more power than i think ill need. What is involved in converting the kt88/OPT into triode ?
 
Administrator
Joined 2004
Paid Member
Ok at 1.5w ( I cannot turn my generator below 50mVpp) at below 5khz i have to have my ear only the OPT to hear it, around 8-9Khz i can clearly hear it from a foot or two away.

This amp has a ton of gain and way more power than i think ill need. What is involved in converting the kt88/OPT into triode ?

OK this is the first bit of good news, sounds to me like you were driving the amp very hard the first time around.

Do check for oscillation (hair) with a 1kHz sinewave input applied into a resistive load. There might still be some lurking stability issue.

To convert to triode disconnect the KT88 screens (pin 4) from their respective transformer taps and insulate the taps. Next solder a 330 ohm 1W resistor or similar (you may go as high as 1K) between the pin 3 and pin 4 of each tube.. I think you can also remove the zobel network between the screen and plate taps on the OPT, but save for later. Check bias - very minor adjustment may be required.
 
Last edited:
Previous posts you mentioned instability, I wonder if the output stage primary is HF oscillating with a signal, and the core iron singing and demodulating it into noise. Have you "scoped" the secondary on load.....what does it look like ? MAke sure no LP filter is set on the scope.

richy
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.