Hi,
Recently I built this amplifier and I have a problem with noise that I can't seem to get rid of.
This is the power supply,
I'm using 300 uF for C101 and C102, 5H, 250 mA choke, 470uf for C103 and C104. Power trans is Hammond 1000VCT.
Amplifier schematic,
I'm having a problem with this noise spike.
This is taken on the speaker output, 2mS per division about 100 mV in amplitude.
I'm using a DC supply for the front end filaments and it is free of noise.
Using AC on the power tube filaments.
No spike on the 150 V supplies but it is present on the plates of V1B and V2B.
I can hear a slight buzz from the PT as the supply comes up to full power.
I have tried another pair of diodes but there was no change.
What should I look at next?
Thanks,
Recently I built this amplifier and I have a problem with noise that I can't seem to get rid of.
This is the power supply,
I'm using 300 uF for C101 and C102, 5H, 250 mA choke, 470uf for C103 and C104. Power trans is Hammond 1000VCT.
Amplifier schematic,
I'm having a problem with this noise spike.
This is taken on the speaker output, 2mS per division about 100 mV in amplitude.
I'm using a DC supply for the front end filaments and it is free of noise.
Using AC on the power tube filaments.
No spike on the 150 V supplies but it is present on the plates of V1B and V2B.
I can hear a slight buzz from the PT as the supply comes up to full power.
I have tried another pair of diodes but there was no change.
What should I look at next?
Thanks,
That looks like may be line frequency related. Check it with respect to the ac out of one transformer winding to see if it is locked to the line freq. If so, I'd look at the Lamda power supply next to see if it is kicking noise on the line. Check the line to see if there is another source causing noise on it.
Sorry I was not too clear, not using the Lambda's.
AC on the 572 filaments with balance pot and transformer/full wave rectifier/cap type linear supply on the 6BM8's.
I turn the lights off in the shop for the scope shot.
No other devices are powered up at the time.
So I should measure the frequency of the incoming line, and each of the secondaries to the CT to ensure the same frequency?
Could frequency be changed on one half of the secondary?
I have an RF filter on the AC input, I may try bypassing that.
I'll report back later.
AC on the 572 filaments with balance pot and transformer/full wave rectifier/cap type linear supply on the 6BM8's.
I turn the lights off in the shop for the scope shot.
No other devices are powered up at the time.
So I should measure the frequency of the incoming line, and each of the secondaries to the CT to ensure the same frequency?
Could frequency be changed on one half of the secondary?
I have an RF filter on the AC input, I may try bypassing that.
I'll report back later.
The spikes are a tad over 8ms apart, that's 120HZ. That means it's on a full wave rectified DC line. Odd for a tube rectified PS. The 150 Volt supplies are half wave and you verified there's no noise there, it would be 60 Hz rate anyway. I'm building guitar amps that use 2 x 5U4s and big caps also, no spikes that I've seen. I've had problems using solid state diode noise getting into the heater windings and then into the cathodes of the input tube and had the exact same output. A separate heater transformer for the input tubes cured that. My point is to scope your input tube heater voltage and see if there is a spike. In fact try turning the input heaters off to see if the spike goes away.
Craig
Craig
Ok, line freq is 60.00 hz, both secondaries are solid at 60.00 hz.
533 VAC from each half of the secondaries.
The spike is present at the R25/C14 junction so it looks like it is on the main supply.
The filament supply does not show the spike.
The +/- 150 VDC supplies do not show the spike.
Bah, Think I'll just turn it up for a while and rock out..
533 VAC from each half of the secondaries.
The spike is present at the R25/C14 junction so it looks like it is on the main supply.
The filament supply does not show the spike.
The +/- 150 VDC supplies do not show the spike.
Bah, Think I'll just turn it up for a while and rock out..
Where is the high voltage center tap of your supply connected to? Generally I find that restricting the capacitor charging current loop to the first filter cap is a good idea, i.e. connect the center tap to the first filter cap and from there go to the star ground. Might or might not help in this case.
Ok, line freq is 60.00 hz, both secondaries are solid at 60.00 hz.
533 VAC from each half of the secondaries.
The spike is present at the R25/C14 junction so it looks like it is on the main supply.
The filament supply does not show the spike.
The +/- 150 VDC supplies do not show the spike.
Bah, Think I'll just turn it up for a while and rock out..
Actually what you said is not necessarily true, is it also present on the center tap of your OPT? or the other side of R25?
If you disconnect the feedback from the output of the amplifier does the spike at C14 go away? I'm wondering if you have magnetic coupling between the choke and the OPT or an issue with the way you have configured your internal grounds.
With the feedback disconnected check all stages for presence of those spikes. You may be picking it up somewhere else.
Can you post pictures of the internal layout and wiring? Something might be more obvious them..
Lets see,
The CT is grounded to the chassis and negative buss at the PT, one of the next components along the buss is the first cap.
I did not probe the high side of R25 or the CT as I am leery of blowing up my scope.
Can I put a cap of sufficent voltage rating in series with my scope probe to ensure DC blockage when measuring the B+?
I can't work on the amp until tomorow night but
I'll snap a few pics tonight and post them tomorrow morning.
I did put another cap just past the diodes as suggested but no improvement was found.
I have used these diodes before at higher voltage and current and have not had this problem before so I'm thinking induction or a ground issue.
Thank you guys very much...
The CT is grounded to the chassis and negative buss at the PT, one of the next components along the buss is the first cap.
I did not probe the high side of R25 or the CT as I am leery of blowing up my scope.
Can I put a cap of sufficent voltage rating in series with my scope probe to ensure DC blockage when measuring the B+?
I can't work on the amp until tomorow night but
I'll snap a few pics tonight and post them tomorrow morning.
I did put another cap just past the diodes as suggested but no improvement was found.
I have used these diodes before at higher voltage and current and have not had this problem before so I'm thinking induction or a ground issue.
Thank you guys very much...
It needs something like 100nf to quench the spikes.
An electrolytic is too inductive to do it.
That should work ok but you need one in each leg.
Each diode needs a capacitor across its output to ground.
Family issues have kept me away from the shop until yesterday when I had a few minutes to probe the circuit.
I made a small coil and connected it to the scope and was able to see the spikes are from the diodes conducting.
As the diodes have top cap I placed a 0.1 uf at the junction where the wires join at the main fuse.
This seems to quiet the amp down a bit, for example I could hear switching noise from my computer PS before, none after the cap was added.
Also the spike is now cleaner, less ringing and a fast recovery
I realize that the HV wire from the top cap is an aerial, I could use a shielded wire but the spikes are radiated right through the glass envelope.
I then probed the circuit and find that the power supplies do not show the spike.
Looks like kevin kr was correct, the feedback wire is carrying the spike back into the front end. I think the noise is coupled into the wire as I ran it too close to the rectifiers.
First I'll open the loop and hopefully things will quiet down. Then relocation to a quieter location in the chassis and I should be set.
Here's a pic of the underside. Originally I had it buttoned up nicely but troubleshooting and adding caps has made things a mess.
I plan to eliminate the noise, clean things up and fab a botton cover.
I made a small coil and connected it to the scope and was able to see the spikes are from the diodes conducting.
As the diodes have top cap I placed a 0.1 uf at the junction where the wires join at the main fuse.
This seems to quiet the amp down a bit, for example I could hear switching noise from my computer PS before, none after the cap was added.
Also the spike is now cleaner, less ringing and a fast recovery
I realize that the HV wire from the top cap is an aerial, I could use a shielded wire but the spikes are radiated right through the glass envelope.
I then probed the circuit and find that the power supplies do not show the spike.
Looks like kevin kr was correct, the feedback wire is carrying the spike back into the front end. I think the noise is coupled into the wire as I ran it too close to the rectifiers.
First I'll open the loop and hopefully things will quiet down. Then relocation to a quieter location in the chassis and I should be set.
Here's a pic of the underside. Originally I had it buttoned up nicely but troubleshooting and adding caps has made things a mess.
I plan to eliminate the noise, clean things up and fab a botton cover.
Very crowded inside that chassis.. In looking I also noticed that there are a lot of excessively long wiring loops - filaments, drive to the final grids, etc. - you really need to shorten up these lines, and dress them close to the metal chassis plate. One grid line in particular is looping all over the place..
I'd also recommend using tie strips or similar rather than the many floating connections I see whether insulated or not.
There are so many transformers scattered around the chassis hopefully none of them are inducing into the OPT.
I'd also recommend using tie strips or similar rather than the many floating connections I see whether insulated or not.
There are so many transformers scattered around the chassis hopefully none of them are inducing into the OPT.
I hsve never had the need to bias the heaters in my previous SE build.
Last night I was probing and pushing wires around and noteed that the feedback pot seemed especially sensitive to noise like a magnet for noise
actually.
Today I rewired the amp and the noise problem is gone, vanished completely.
I still have a few things I want to fix but today I am a happy camper indeed.
Lessons learned (again)...
Pics forthcoming
Last night I was probing and pushing wires around and noteed that the feedback pot seemed especially sensitive to noise like a magnet for noise
actually.
Today I rewired the amp and the noise problem is gone, vanished completely.
I still have a few things I want to fix but today I am a happy camper indeed.
Lessons learned (again)...
Pics forthcoming
Sounds like all good news.
I do completely agree with Mike about the rectifier filaments - it is a good practice as it minimizes cathode to filament insulation stress. In the event that you don't trust the filament windings to handle the full B+ voltage you can often use a resistive divider and reference the filaments to that with the proviso that the voltage after division should be within the insulation capability of both the filament/cathode and the transformer winding insulation. NOTE: This obviously only applies to rectifiers such as indirect heated types without a common filament/cathode connection..
The exception to this rule is in lower voltage applications where you wish to use the same 6.3V filament supply to heat a bunch of tubes in which case the filament supply is usually referenced to some % of filtered B+ that is safe for all tubes used. Needless to say this is not a very good idea in applications where very high voltages are present on the cathode of the rectifier. (Say anything over 350V or so) 6V4, 6CA4, 6AU4, 6BY5 are all types where this may be done safely at reasonable voltages.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.