Go Back   Home > Forums > >

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Oscillation in tube amps
Oscillation in tube amps
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools
Old 9th May 2020, 11:01 PM   #211
Potentiallyincorrect is offline Potentiallyincorrect  United States
diyAudio Member
Join Date: May 2020
Wow, I really love that beautiful hand drafted schematic with notes. I use LTspice for schematics now, but hand drafting one like that really focuses the brain on the circuit are hand. Many of us have gotten lazy with the great tools available, but much is also lost.
  Reply With Quote
Old 20th May 2020, 06:54 AM   #212
SemperFi is offline SemperFi  Wake Island
diyAudio Member
Join Date: Jan 2002
Location: nowhere
Default Chasing the osc

Digital oscilloscopes are handle little tools. I have a PC type (Picoscope 5444D). But I just experienced the under-sampled alias, and think this may be of interest. Most of us know about this digital limitation, but it is so easy to forget in the comfort of easy measurements. So this post is not about curing the oscillation, just to show how digital scopes can give bogus results.
So I fired up a new build, and of course instead of a fine and thin line the trace is wide and varies sporadically. Inserting a 1kHz test signal on the left ch gives weird bubbles on the right ch.
(This is a 6AS7 amp with a folded cascode input/gain stage. The measurements were with the output section not installed, only the input stage is active. The cascode transistors are 2SK1381, pretty fast devices, so no surprise regarding oscillations).
So I switched to FFT mode to pin point the oscillation frequency. The sample rate is 2x the max frequency selected for the FFT view. So if I select FFT to 100kHz, the sampling is 200ksps.
The FFT up to 100kHz shows a wide band (almost noise-ish) burst oscillation. Peak is 47kHz.
Increasing to 1MHz (992kHz really) FFT the peaks go up 323kHz and above, relatively wide band still. The sampling is now 1.984MSPS.
Going up to 8.929MHz, sample rate 17.86MSPS, the highest peak is at 1.75MHz.
Up to 125MHz, sampling 250MSPS, the peak is 105MHz. There is some wide band peaking here and there.
The scope's max range is 250MHz, sampling 500MSPS. The peak is still at 105MHz, but it is not dead solid.

Obviously what I am seeing is under-sampling. The time domain view set to highest sample rate shows a relatively stable 104.8MHz but I suspect this is under-sampled and the actual oscillation frequency could be higher.

Are there any simple formulae that lets me calculate what the real frequency is based on the sample rate, and supposed frequency peaks shown at the given rate?

So now it's troubleshooting time...
Attached Images
File Type: png testsignal_oscil.png (96.3 KB, 305 views)
File Type: png 200ksps.png (62.4 KB, 302 views)
File Type: png 1_984Msps.png (61.9 KB, 301 views)
File Type: png 17_86Msps.png (62.0 KB, 299 views)
File Type: png 250Msps.png (59.1 KB, 303 views)
File Type: png 500Msps.png (58.8 KB, 28 views)
File Type: png time_250Msps.png (65.1 KB, 29 views)
File Type: png time_500Msps.png (61.6 KB, 29 views)
  Reply With Quote
Old 23rd June 2020, 12:30 AM   #213
Mark Tillotson is offline Mark Tillotson
diyAudio Member
Join Date: May 2018
Location: Cambridge UK
For instance the 105MHz signal (which is probably correct as the 'scopes front-end amp will roll off from 250MHz quite severely) = 6 * 17.86MHz - 1.75MHz

The undersampling spurs for a tone at frequency f will be at n*Fs +/- f, where n is an integer and Fs is the sampling rate, as sampling is like RF mixing. You would have to guess n though, so the best strategy is start at the fastest timebase setting and work down.

Expensive 'scopes do proper rate-conversion and don't show such under-sampling artifacts. This also gives you free sample-averaging to obtain higher ENOB. It takes quite a lot more hardware to do this though as digital filters are needed running at the full sample rate and various sub-multiples of it too. Other tricks you see are tracking the max and min samples so that you can see there's signal present even when its faster than the current timebase would display. Sometimes that's annoying though as it can just be HF noise.

With lots of memory depth this processing can be done off-line, as no realtime
processing needs to be done, other than filling memory - but then the refresh rate can be disappointing.
  Reply With Quote
Old 16th July 2020, 08:27 PM   #214
wagonicfolding is offline wagonicfolding
diyAudio Member
Join Date: Jul 2020
This is almost certainly due to:

insufficient filtering in your +250 power supply or

connecting the output to a too-low impedance. The load on 6SN7 cathode follower like this should not be lower than about 1,000 or 2,000 ohms. Check your power amplifier's input impedance.

a high resistance ground connection to the negative side of the power supply and the amplifier's input stage ground point. A preamp like this usually works best if its chassis ground point is as close as possible to the ground point on the input jack.
  Reply With Quote
Old 18th February 2021, 01:53 AM   #215
longspeak is offline longspeak  United States
diyAudio Member
longspeak's Avatar
Join Date: Sep 2015
Location: Colorado
Oscillation in tube amps
I have a stereo tube amp that I'd been working on plugged into a dual AFCI/GFCI outlet and it tripped a couple times indicating an arc fault. I assumed it was a nusance tripping but put a scope on the psu grounding and found out it was oscillating in ~10Mhz bursts on a 120Hz cadence . It needed a gate stopper on a mosfet in a capacitance multiplier circuit in it. It had a very low level 120Hz buzzing sound in the speakers but it wasn't too objectionable. I'm not sure I would have caught the problem as quickly without the AFCI/GFCI tripping. I think a AFCI/GFCI outlet is a good idea with tube amps for safety and also for troubleshooting now.

Last edited by longspeak; 18th February 2021 at 02:15 AM. Reason: spelling
  Reply With Quote
Old Today, 04:01 PM   #216
adamus is offline adamus  United Kingdom
diyAudio Member
Join Date: Apr 2008
here is my square wave test. Any views on this would be helpful.
636p-dr parafeed, The 1khz squarewave looks the same from the plates as it does from the loaded transformer (same overshoot and minor ringing).

Gridstoppers on the pins (1k). Should i go chasing... or is this ok? The thickness of the horizontal is probe noise. I can flatten that out if i use short ground leads.

Click the image to open in full size.

Click the image to open in full size.

Last edited by adamus; Today at 04:13 PM.
  Reply to this post


Oscillation in tube ampsHide this!Advertise here!
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
For Sale: Tube Amps, Class-A Amps, BG Ribbons, Chassis and Other Goodies! opc Swap Meet 23 2nd November 2009 06:04 PM
Help Please with tube amps!!!!! bobcat1979 Instruments and Amps 1 28th June 2009 12:29 PM
Tube amps & MP3's jim Tubes / Valves 18 20th February 2008 05:15 AM
Tube Oscillation or Microphonic Question Trout Tubes / Valves 6 29th December 2005 05:18 AM

New To Site? Need Help?

All times are GMT. The time now is 07:49 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Copyright ©1999-2021 diyAudio