UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Audio (record players, hi-fi etc)

Notices

Vintage Audio (record players, hi-fi etc) Amplifiers, speakers, gramophones and other audio equipment.

Closed Thread
 
Thread Tools
Old 5th May 2019, 8:35 am   #81
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

You can make up your own measurement software for the 8903 and most other HP/Agilent/Keysight hardware using a copy of HP/Agilent/Keysight VEE. Unfortunately they want money for it, it would have been a great sales tool as a freebie. It's a graphical interface test manager with lots of maths, plotting, automation capability and can amalgamate control of most things.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 5th May 2019, 10:38 am   #82
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

I have the GPIB cables and an old PC running the NI interface software and back in the late Cretaceous when I acquired all this stuff the plan was to write some software of my own to do the routine measurements. Unfortunately it's one of those situations where, over the years, the resulting package might have saved me days (weeks ?) of manual scanning. But at any given time I can't set aside the days/weeks needed to get on with the job. So I'm still pushing buttons. What will probably force the issue is when the keypad starts to fail .

As far as the kit itself is concerned, I was running in the 80-100kHz region with 1V RMS out when I started to see occasional higher frequency noise bursts appearing on the output. This went on for 10-15 minutes or so then suddenly the output dropped three orders of magnitude and now I can't change its amplitude. The signal still looks clean on the spectrum analyser and I can change the frequency. So the oscillator's probably fine and something will have gone west in the attenuator. Ho hum.

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Old 5th May 2019, 2:28 pm   #83
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

higher freq bursts sounds like the amplifier rather than the attenuator

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 5th May 2019, 5:16 pm   #84
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

The 8903A problem must have been something temporary - either a loose external cable or a poor contact (edge connector ?) internally. I took it out of the rack, opened it up, applied power and it was fine. So I reassembled it, put it back and finished the measurements.

With the transformers out of the chassis and a 16.5ohm resistor soldered onto the 16ohm secondary turrets I applied an AC signal to each half primary in turn - between the anode turret and the HT turret. I put the signals from the anode and screen grid turrets onto a scope and measured the time delay between their zero crossings. From that I worked out the phase shift between them. I then scanned the frequency from 80kHz to 100kHz (the maximum the 8903A can reach). The results are plotted in the attached pic. The green data are from the good transformer and the red from the bad one. The squares represent one half primary and the circles the other.

I can't see any real sign of resonances at these frequencies but perhaps, as kalee as pointed out, they would be affected in the amps both by stray capacitances and also by the effects of the valves. What is clear is that the two half primaries in the bad transformer seem much more similar to one another than the two in the good one. I wonder if the common mode oscillation I've been seeing with the bad transformer arises because of this closer 'match' between the two halves of the output stage ?

Cheers,

GJ
Attached Thumbnails
Click image for larger version

Name:	TL50 Plus OPT phase shifts.jpg
Views:	78
Size:	34.1 KB
ID:	182518  
__________________
http://www.ampregen.com

Last edited by GrimJosef; 5th May 2019 at 5:25 pm.
GrimJosef is offline  
Old 5th May 2019, 6:06 pm   #85
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Shouldn't there be 180 degrees difference between the phase of one primary (wrt secondary) and that of the other, but the traces seem to converge towards lower frequencies?

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 5th May 2019, 6:18 pm   #86
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Sorry I should have been more clear. I'm looking at the phase shift between the applied signal across a-HT and the measured signal across g2-HT as kalee suggested towards the bottom of post #75. The idea was to see whether the effects of leakage inductances and stray capacitances inside the transformer might be reversing the phase of the signal to g2 at high frequencies and thereby causing all the trouble. What I'm seeing is that with the transformer on its own we get more like a 90 degree shift rather than a 180 degree one, but the shifts vary a lot from transformer to transformer and from half-primary to half-primary so they must depend quite sensitively on some (poorly controlled) parameter. If the rest of the circuit (wiring, valves) contributes 'more' of whatever parameter it is then maybe oscillation will result ?

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Old 5th May 2019, 6:20 pm   #87
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Quote:
Originally Posted by GrimJosef View Post
I went out this afternoon to run kalee's frequency scan across the a-HT winding monitoring the g2-HT section...What I did learn was that, essentially unloaded (1M/14pF scope probe), the g2-HT waveform went from lagging the a-HT one by 65 degrees at 80kHz to lagging it by 88 degrees at 96kHz. This was on one half of the primary, with the other half being distinctly different and substantially less out of phase.
That's a LOT of phase shift, for something which is supposed to be in phase!

At high frequencies the transformer's primary inductance is going to have a massive reactance so we can forget it. The impedance presented to the output valve is, I suspect, largely capacitive due to the self-capacitance of the winding. Although the load is supposed to be in parallel with this, there's the primary-secondary leakage inductance to consider, which increasingly acts to isolate the load as frequency rises. So it is quite plausible that the overall load on the valve is capacitive.

That being so, Va will be 90deg out of phase with Ia. Add your 88deg of phase shift between Va and Vg2 and you have 178 degrees of overall phase shift, which is 2 degrees from oscillation. It's not hard to pick up 2 degrees...

A good transformer will keep the voltages at the anode and screen tap in phase well up to the frequency where the self-capacitance has reduced gain to below 1, so it will never oscillate.

Assuming that I've understood your results correctly, that explains the oscillation. It would have been good to be able to examine the transformer winding configuration to see what is bad, but that'd wreck the evidence!

It would be worth checking the transformer at 1kHz whereupon the 88 degrees should be zero-ish (it would be really hard to make a transformer that's wrong in the middle of the audio range).

You could also try the amplifier with the screen taps disconnected and the screen-grids taken to HT+ (tetrode operation) where I bet it'll be fine. It's probably going to run the valves at non-optimal conditions, but it will prove a point.
kalee20 is offline  
Old 5th May 2019, 6:22 pm   #88
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Ref above post.. Others have been typing same time as me, I'll go and read!!
kalee20 is offline  
Old 5th May 2019, 7:05 pm   #89
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Ah, sorry. As phase shift betweed an anode terminal and the g2 tap, that phase shift is scary. I agree with Kalee20.

Considering that the HT-g2-a winding is essentially an autotransformer, which wins a bit in the leakage inductance stakes, and that the g2 tap is well away from the HT end, both good and bad transformers look rather poor.

I think you've discovered a serious design problem running through all TL50+ and that those which operate stably are on thin ice.

I'm not so much wondering why that one hoots so much as why the others don't.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 5th May 2019, 8:09 pm   #90
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Given that the results suggest the transformer isn't great I thought I'd better go out and run some sanity/reassurance checks.

The voltages are measured by clipping scope probes onto the transformer turrets. If I clip both probes onto the same turret then the scope traces overlap to within the visible resolution on the screen (I tested at 80kHz so this was better than 20ns or so, which is less than a degree of phase shift).

Then I reversed the probes, in case there was some systematic delay in the test equipment. There wasn't. At 80 kHz I originally measured a 3.02us delay on the good transformer's 'left hand' (arbitrarily named by me) half-primary. When I went back out and reversed the probes I measured -3.04us - essentially the same number to within the above resolution.

Finally I stepped the frequency down to the audio range, checking the delay (i.e. phase shift) as I went. Converting to degrees it was

64 degrees at 70kHz
47 degrees at 60kHz
35 degrees at 50kHz
27 degrees at 40kHz
21 degrees at 30kHz
14 degrees at 20kHz
11 degrees at 15kHz
8 degrees at 10kHz
4 degrees at 5kHz

and I couldn't measure it at 1kHz - it was too small. Interestingly the delay from 60kHz downwards was constant at around 2us +/- 0.2us. I don't know if that represents some constant of the transformer ?

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Old 5th May 2019, 8:16 pm   #91
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

As the 'good' transformer is showing the worst phase shift, I am not convinced that this is the cause. This level of phase shift is unexpected so we may want to look at the measurement method.

Last edited by PJL; 5th May 2019 at 8:23 pm.
PJL is offline  
Old 5th May 2019, 8:48 pm   #92
Valvepower
Octode
 
Valvepower's Avatar
 
Join Date: Aug 2008
Location: Rayleigh near Southend-On-Sea, Essex, UK.
Posts: 1,852
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Hello,

As a long shot to see how it behaves with a different (lower gm..?) valve you could try the KT66. This would mean running the amplifier via a variac to drop the voltage by 10% or thereabouts to get a lower HT of 450V keeping anode dissipation down – the filaments will just be on their minimum, but they should be ok – alright, the biasing will not be totally correct but should suffice for this experiment. Its real daft idea, but worth a try to see if it affects the oscillation?

Terry
Valvepower is offline  
Old 5th May 2019, 10:44 pm   #93
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Quote:
Originally Posted by PJL View Post
... This level of phase shift is unexpected so we may want to look at the measurement method.
It's certainly worth scrutinising this. I've never done this measurement before and I haven't thought the interpretation through, so I wouldn't necessarily spot a howler simply by looking at the results. That was partly why I wanted to see the phase shift (in degrees at least) pretty much disappear as we moved down into the audio region. Which it did.

By way of a bit more reassurance I've just hauled my Radford STA15 out to the workshop and used exactly the same test rig to do some quick checks on one of its output transformers. Radford were famous for their transformer expertise. In this case I didn't take the transformer out of the amp. So the wiring loom was still connected as was the secondary feedback network. I selected the 16ohm output tapping and soldered my 16.5ohm resistor across the speaker terminal tags. I then connected my signal source to the anode tag on the valve socket and to a wire from the HT pin on the pcb. The scope probes were 'grounded' to that wire and were hooked to the anode and g2 tags on the valve socket.

The delays I measured were

19 degrees at 80kHz
5.5 degrees at 40kHz
2.7 degrees at 20kHz

So the test rig is at least capable of measuring small phase shifts at relatively high frequencies. Of course what is controlling them is the really interesting question.

EDIT: For what it's worth the delays at 40kHz and 20kHz were, once again, the same. But instead of being 2us as in the Leak transformer, they were 0.38us.

Cheers,

GJ
__________________
http://www.ampregen.com

Last edited by GrimJosef; 5th May 2019 at 10:49 pm.
GrimJosef is offline  
Old 5th May 2019, 11:40 pm   #94
julie_m
Dekatron
 
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Quote:
Originally Posted by Radio Wrangler View Post
I think you've discovered a serious design problem running through all TL50+ and that those which operate stably are on thin ice.

I'm not so much wondering why that one hoots so much as why the others don't.
Could its tendency to oscillate have been accidentally tamed by the passives of the day? Modern resistors have more stray inductance than older carbon ones; whereas modern capacitors have less stray inductance than older ones with larger areas of thicker dielectrics.

I'm sure I'm not the only person to have fitted new 2N3055s in an old amplifier, only to discover that it's now "a bit lively" .....
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments.
julie_m is offline  
Old 6th May 2019, 12:16 am   #95
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

This is one of the most interesting threads I've seen. It's opened up an aspect of valve amplifier design I hadn't thought about before.

The difference between the Leak transformer and the Radford one is quite large.

The mystery now is what constructional differences bring about the performance differences.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 6th May 2019, 12:29 am   #96
Trevor
Octode
 
Join Date: Aug 2006
Location: Stoke-on-Trent, Staffordshire, UK.
Posts: 1,082
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Could this be something to do with the actual lamination's are they the same for both transformers
I feel the Radford transformer is a completely different beast being much smaller with possibly thinner laminations
Trev
Trevor is online now  
Old 6th May 2019, 10:04 am   #97
Valvepower
Octode
 
Valvepower's Avatar
 
Join Date: Aug 2008
Location: Rayleigh near Southend-On-Sea, Essex, UK.
Posts: 1,852
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Hello,

From the outset and GJ’s first post I’ve been racking my brains to remember where I’d heard about another Leak amplifier hooting around 90K and last night in bed I remembered, and it was in one of Dave Porter’s “tricks of the trade” articles on the BBCeng.info site. This described a MWT BD228A 50 kW transmitter installed at Brookman’s Park in the early 60’s..

Link to David's article on the BBCeng.info site:

https://www.bbceng.info/Technical%20...l_Issue_18.pdf

If you read the article MWT used the a TL12+ as a preamplifier This amplifier had a penchant for hooting at 80KHz when the EF86 gave up the ghost! Seeing this amplifier would have been made in the early 60’s there is a strong likelihood the output transformer had a very similar constriction to that of the one in the TL50. David Porter says in the article they replaced the EF86, but they also changed all the valves just in case – this has me thinking if the output transformer wasn’t blameless in this amplifier?

Radford STA15 with Bailey driver and Radford transformers was an exemplary example of a valve amplifier technology.

Finally, it would be interesting to compare the early and later Leak output transformers with different primary and secondary sectionalisation – for example the 3925 and 8615 as used in the TL25 plus. I have both here but it maybe a day or two before I can get them onto the bench, as I’m currently doing house preparation for the sparks and heating engineer to rewire and fit modern heating in my late parents’ house.

Regards
Terry
Valvepower is offline  
Old 6th May 2019, 11:31 am   #98
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

Interesting.

The TL12 output transformer might be a bit better as balanced outputs and higher impedances ought to allow better sectioning, thoug there is a third output used to keep the feedback similar to standard.

I see MWT equipped their TT21 push pull stage with 1k stoppers in both g1 and g2 of each valve... and TT21s are KT88s with top caps and more cathode pins. Even the next stage of parallelled-up twin TT21s got 1k on each g1 and g2. Quite generous stopping!

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 6th May 2019, 12:01 pm   #99
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

I am no transformer expert, but it seems to me that phase shift within what is essentially an autotransformer means that there is insufficient coupling between sections of the same winding. Using too thick laminations, or poorly insulated laminations, could do this at higher frequencies. A poor interleaving scheme could also do this.
G8HQP Dave is offline  
Old 6th May 2019, 12:02 pm   #100
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cause of ultrasonic oscillation in a push-pull output stage ?

For completeness I repeated the above measurements on a couple of other transformers this morning - both ones that I've had in the loft with a view to building pairs of Mullard 5-10s and 5-20s (one day, one day ...):

Parmeko P2642 (43% option for the 5-10)

2.3 degrees at 80kHz
Shift too small to measure at 40kHz
2.5 degrees at 20kHz

VVT VTP203200 (for the 5-20 - now replaced in the catalogue by VTP12923-1200)

-28 degrees at 80kHz
-4 degrees at 40kHz
11 degrees at 20kHz

The Parmeko seems to perform particularly well in this regard. The VVT is a perfectly decent transformer, but is not expensive. Its phase variation is larger than the Radford and the Parmeko but still not nearly as large as the Leak at very high frequency. Interestingly the sign of the HF shift is different in this transformer.

Quote:
Originally Posted by Trevor View Post
... I feel the Radford transformer is a completely different beast being much smaller ...
Trev
I checked the Radford laminations. The stack is the same thickness as the Leak one - 1.5". The lams themselves are smaller though - the stack is 3" x 2.5" externally (compared with 4.5" x 3.75" on the Leak). That's a factor of more than 2 difference in iron volume. But the Leak amp is rated for more than a factor of 3 difference in output power, so the Radford iron isn't working harder.

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 1:04 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.