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 2020, 1:30 pm   #81
Techman
Dekatron
 
Techman's Avatar
 
Join Date: Nov 2009
Location: Lincolnshire, UK.
Posts: 4,985
Default Re: Mullard 5-20 high DC Voltages

Looking in on this thread every so often and seeing all these four pages of comments and advice, I've up until now strongly resisted stating the very obvious (to me anyway), that why build or prototype this amp on a wooden chassis?

This is the sort of thing that I used to do as a kid down in the garden shed and I very soon learned that it was a very BAD idea!

This idea of building stuff on a chunk of wood was all very well back in the 1920s with low powered TRF radios and such like with performance not much better than a crystal set, but with an amplifier such as this, to me anyway, it just seems a complete waste of time and effort.

Even if this was got to perform correctly (or near correctly) on its wooden box, it would be all over the place if then built into a metal chassis due to all the changes in capacitance between wiring etc., that a metal chassis allowed for, and was built into the original design.

There's no excuse for not prototyping on a metal chassis. There must be loads of sheets of scrap aluminium from the likes of the closure plates from behind old removed gas fires, or the sheet metal from old washing machine cases etc. Just cut and bend to shape, pop rivet the corners and job done!

It's possible that I've missed the point of why this has to be built on a wooden box and the OP is obviously having fun doing it, so I can't fault him for that, but if this was a commercial venture it wouldn't pass any fire or otherwise safety regulations and I certainly wouldn't go out and leave it unattended. As for your house insurance if anything did happen is another matter entirely - I say forget the wooden box and knock up a metal chassis, it'll work much better IMHO.
Techman is offline  
Old 5th May 2020, 2:31 pm   #82
Kei-1986
Triode
 
Join Date: Apr 2020
Location: Newport, South Wales, UK.
Posts: 33
Default Re: Mullard 5-20 high DC Voltages

The use of wood was simply to "prototype" a layout and prove it worked. My plan is to get the final layout CNC machined. I would agree that I'm starting to think that not having a metal chassis is quite likely to be at least a part of my problems. I've got some 19" 2mm thick aluminium (old rack blanking plates) that I will try to machine a new top plate from. I don't have as much in the way of tools for dealing with metal hence preferring to use wood.

Did some open loop testing during my lunch break. Output from EF86 plate looks clean into the grids of 12AX7/5751. Out of the plates of the phase splitter it looks like a shark fin. No sign of any RF on the output with bandwidth limiting switched off.

10KHz SQ - EF86 out Vs 12AX7 out - Open Loop
[img]https://***********************/65535/49858992581_03511d6bab_4k.jpg[/img]

10KHz SQ - 12AX7 plates - Open Loop
[img]https://***********************/65535/49858456978_611073ef73_4k.jpg[/img]

DC conditions measured with HT at 385Vac at the rectifier

C15 - 452
C12 - 438
C5 - 409
C4 - 116

V1
A - 57
K - 0.54
G - 0

V2
A1 - 337
A2 - 336
G1 - 56.9
G2 - 51.9
K - 60.8

V3
A - 427
K - 30.1
SG - 425

V4
A - 427
K - 29.9
SG - 426
Kei-1986 is offline  
Old 5th May 2020, 2:45 pm   #83
Techman
Dekatron
 
Techman's Avatar
 
Join Date: Nov 2009
Location: Lincolnshire, UK.
Posts: 4,985
Default Re: Mullard 5-20 high DC Voltages

Quote:
Originally Posted by Kei-1986 View Post
The use of wood was simply to "prototype" a layout and prove it worked.
I thought that was probably the reason, but I couldn't remember what had actually been said about it without going back and reading through the original posts again.

The trouble is that I feel that you've spent an awful lot of time that's going to end up as largely wasted when you find that everything is different when you eventually build it into a metal chassis. As a kid I had easier access to knocking things up from wood down in the shed, rather than using metal. But, once I was older, I got tools to make things out of metal - so much better. I think you're going to find that even though making a metal chassis would have been far more difficult, in hindsight it would have been far easier in the long run.
Techman is offline  
Old 5th May 2020, 3:15 pm   #84
Boulevardier
Octode
 
Join Date: Sep 2017
Location: Bristol, UK.
Posts: 1,642
Default Re: Mullard 5-20 high DC Voltages

Quote:
Originally Posted by Techman View Post
Looking in on this thread every so often and seeing all these four pages of comments and advice, I've up until now strongly resisted stating the very obvious (to me anyway), that why build or prototype this amp on a wooden chassis?

This is the sort of thing that I used to do as a kid down in the garden shed and I very soon learned that it was a very BAD idea!

This idea of building stuff on a chunk of wood was all very well back in the 1920s with low powered TRF radios and such like with performance not much better than a crystal set, but with an amplifier such as this, to me anyway, it just seems a complete waste of time and effort.

Even if this was got to perform correctly (or near correctly) on its wooden box, it would be all over the place if then built into a metal chassis due to all the changes in capacitance between wiring etc., that a metal chassis allowed for, and was built into the original design.

There's no excuse for not prototyping on a metal chassis. There must be loads of sheets of scrap aluminium from the likes of the closure plates from behind old removed gas fires, or the sheet metal from old washing machine cases etc. Just cut and bend to shape, pop rivet the corners and job done!

It's possible that I've missed the point of why this has to be built on a wooden box and the OP is obviously having fun doing it, so I can't fault him for that, but if this was a commercial venture it wouldn't pass any fire or otherwise safety regulations and I certainly wouldn't go out and leave it unattended. As for your house insurance if anything did happen is another matter entirely - I say forget the wooden box and knock up a metal chassis, it'll work much better IMHO.
I’m with Techman on this. Most of the suggestions so far have been about circuit minutiae like adjusting resistor/capacitor values, etc, and I wonder if all these aren’t just missing what could just turn out to be the “elephant in the room” - the wooden ‘chassis’. It’s impossible to say exactly what difference a metal chassis might make without trying it, but trying it would at least remove one huge (IMHO!) question mark.
I also think that it would be better to construct a metal chassis from one, appropriately bent, piece of aluminium. Bolting five pieces together at the edges/corners runs the risk of creating new earth currents through the inevitable resistances across the joints. Again, impossible to predict in advance whether or not this would be a problem in practice, but better to avoid a possible source of problems from the start. Again, the Mullard article gives details of their chassis build as a starting point at least.

Mike
Boulevardier is online now  
Old 5th May 2020, 3:27 pm   #85
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Mullard 5-20 high DC Voltages

The style of construction needed by something is not only a function of the frequency range you intend it to work over, but also a function of the frequency range the active devices could oscillate up to!

A good design must not only do all the things it must do, it must also not do all the things it must not do. They are just as important and may even be harder work.

People get a little too close to things: When thinking of filters passband shape gets the concentration and the stopband gets ignored, but the stopband is what it's all about, if you didn't want a stopband, a piece of simple wire would make a great filter. Passes everything, and flat!

Someone thinking of buying a spectrum analyser for his shack said "Something going to 100MHz will be fine for me, I don't build anything over 30MHz" It sounds logical, but misses the need to spot spurious oscillation. He needed to go to the oscillation limit of the devices he used to be sure nothing nasty was going on.

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 2020, 4:18 pm   #86
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Mullard 5-20 high DC Voltages

The output from the phase splitter looks very odd to me. One of the outputs is also looking considerably worse than the other. Are you using x10 probes? What is the probe input capacitance?
PJL is offline  
Old 5th May 2020, 4:52 pm   #87
Kei-1986
Triode
 
Join Date: Apr 2020
Location: Newport, South Wales, UK.
Posts: 33
Default Re: Mullard 5-20 high DC Voltages

Probes are x10 Tektronix P6137 which communicate with the scope so the scale numbers tally up. Input capacitance is 10.8pF. All of the probes have been set correctly on the scope calibrator. The output from the phase splitter really does look that bad on a 10KHz square wave when run as open loop. I can give it another go with the 12AX7 in place of the 5751 that was in there but I don't think there is going to be any change.

The chassis I was planning to build was never a complete one piece metal enclosure as the sides were going to be wooden with metal panels for the cutouts, leaving only the top and bottom as fully metal. The two pieces of aluminium I have are 3mm thick 19" 4U rack blanking covers. I can drill holes in it but I have no way to bend it. I might know someone with some greenlee punches so might have a chance to do better than using a hole saw.
Kei-1986 is offline  
Old 5th May 2020, 5:56 pm   #88
Kei-1986
Triode
 
Join Date: Apr 2020
Location: Newport, South Wales, UK.
Posts: 33
Default Re: Mullard 5-20 high DC Voltages

Someone on DIYaudio suggested performing the grid to plate check with no output valves fitted. Voltages were extremely high as expected. Output from the PI with no output stage looks good.

This is the grid of the PI vs the plate
[img]https://***********************/65535/49859179138_4f3b0490fd_k.jpg[/img][/url]

I repeated the earlier test just to make sure the 12AX7 behaved the same as the 5751.
[img]https://***********************/65535/49859200283_57f5a6c671_k.jpg[/img]
Kei-1986 is offline  
Old 5th May 2020, 6:00 pm   #89
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Mullard 5-20 high DC Voltages

Very strange.

It almost looks like there is a 300pF or so load on the ECC83 anodes. I wonder if this has anything to do with your EL34 relay sockets. I suggest you disconnect the mauve wires that go from EL34 pin 6 to the caps and, to avoid instability, connect them via 10K (or something around that) resistor to earth and scope the LTP anodes again.

PS: We crossed posts - maybe there is an issue with using pin 6. You could try lifting the components off the pin.
PJL is offline  
Old 5th May 2020, 9:24 pm   #90
Sideband
Dekatron
 
Sideband's Avatar
 
Join Date: Oct 2003
Location: Croydon, Surrey, UK.
Posts: 7,548
Default Re: Mullard 5-20 high DC Voltages

Quote:
Originally Posted by Techman View Post
Looking in on this thread every so often and seeing all these four pages of comments and advice, I've up until now strongly resisted stating the very obvious (to me anyway), that why build or prototype this amp on a wooden chassis?

This is the sort of thing that I used to do as a kid down in the garden shed and I very soon learned that it was a very BAD idea!

This idea of building stuff on a chunk of wood was all very well back in the 1920s with low powered TRF radios and such like with performance not much better than a crystal set, but with an amplifier such as this, to me anyway, it just seems a complete waste of time and effort.

Even if this was got to perform correctly (or near correctly) on its wooden box, it would be all over the place if then built into a metal chassis due to all the changes in capacitance between wiring etc., that a metal chassis allowed for, and was built into the original design.

There's no excuse for not prototyping on a metal chassis. There must be loads of sheets of scrap aluminium from the likes of the closure plates from behind old removed gas fires, or the sheet metal from old washing machine cases etc. Just cut and bend to shape, pop rivet the corners and job done!

It's possible that I've missed the point of why this has to be built on a wooden box and the OP is obviously having fun doing it, so I can't fault him for that, but if this was a commercial venture it wouldn't pass any fire or otherwise safety regulations and I certainly wouldn't go out and leave it unattended. As for your house insurance if anything did happen is another matter entirely - I say forget the wooden box and knock up a metal chassis, it'll work much better IMHO.
I've been dipping in and out of this thread and come to the same conclusion. The 5-20 is a tried and trusted design from a company that knew what they were doing......it was after all designed at the Mullard Research Labs to promote their 'new' valves available at the time.

I never built the 5-20 but I did build the 5-10 (which is pretty much the same design, just lower power)....in fact two of them as monoblocks and later the stereo pre-amp. I stuck rigidly to the design layout on all occasions....fortunately I did have someone who could bend the chassis for me (my dad). The two 5-10's were built as 'mirror image' using Parmeko transformers and ultra-linear. No problems at all from the start and I had the whole lot for many years from the late 60's until the mid 70's when I gave them to a friend who had them for a further 10 years or so. I don't recall any breakdowns or problems...they just worked!
__________________
There are lots of brilliant keyboard players and then there is Rick Wakeman.....
Sideband is offline  
Old 5th May 2020, 11:18 pm   #91
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Mullard 5-20 high DC Voltages

We are getting close. The EL34 grid capacitance is far too small to cause this. The fact it does not occur with the output valves removed suggests it is not capacitive coupling to earth. The fact that both LTP outputs behave similarly implies it is unlikely to be induced. It might be RF instability so you could try scoping the EL34 grid but if it were I would have expected it to be everywhere.

My favourite remains the EL34 valve socket but capacitive coupling from pin 4 (g2) to pin 5 (g1) which would create this effect. Not sure how you can prove it other than by fitting a replacement socket but you should order some ceramic sockets for the EL34's in any case as these will cope with the high voltage and temperature.

Last edited by PJL; 5th May 2020 at 11:30 pm.
PJL is offline  
Old 6th May 2020, 6:51 am   #92
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Mullard 5-20 high DC Voltages

Don't agree completely 100% with the wood chassis comments, I prototype using wood for the OP stage, as do quite a few successful, very knowledgable builders over on DIYaudio, however once the general shape of the amp is sorted I use thin ali, 0.7mm bent into a U shape to then further develop the amp, but agree moving on from wood you have to then further de-bug things.

There's stray capacitance somewhere, yesterday whilst messing about with my amp I'm building, I accidentally overloaded the FB capacitor using a decade box, the OP rang like the OP's, the LTP OP looks like an OP when too much capacitance is used as the FB C, no wonder the OP has been looking so dreadfull.

Scope the LTP at pins 5 on the EL34 base and at ECC83 anodes if the scope will take it with EL34's pulled.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 6th May 2020, 8:51 am   #93
vidjoman
Dekatron
 
Join Date: Jun 2013
Location: East Sussex, UK.
Posts: 3,315
Default Re: Mullard 5-20 high DC Voltages

You don't have to bend metal to make a chassis, Use 5 flat bits, 4 sides and a top, held together with small metal angle brackets or aluminium angle. Simple to do and you do then have proper screening etc. Done it many times when I didn't have metal bending facilities.
vidjoman is online now  
Old 6th May 2020, 9:59 am   #94
Kei-1986
Triode
 
Join Date: Apr 2020
Location: Newport, South Wales, UK.
Posts: 33
Default Re: Mullard 5-20 high DC Voltages

Quote:
Originally Posted by Diabolical Artificer View Post
Don't agree completely 100% with the wood chassis comments, I prototype using wood for the OP stage, as do quite a few successful, very knowledgable builders over on DIYaudio, however once the general shape of the amp is sorted I use thin ali, 0.7mm bent into a U shape to then further develop the amp, but agree moving on from wood you have to then further de-bug things.

There's stray capacitance somewhere, yesterday whilst messing about with my amp I'm building, I accidentally overloaded the FB capacitor using a decade box, the OP rang like the OP's, the LTP OP looks like an OP when too much capacitance is used as the FB C, no wonder the OP has been looking so dreadfull.

Scope the LTP at pins 5 on the EL34 base and at ECC83 anodes if the scope will take it with EL34's pulled.

Andy.
Currently, all of my testing is open loop with no feedback. The distortion out of the phase inverter when the output valves are fitted has me a little baffled but I do wonder about the operating points on the EF86 vs the ECC83 as I've not adjusted it since changing the front end.

Quote:
Originally Posted by vidjoman View Post
You don't have to bend metal to make a chassis, Use 5 flat bits, 4 sides and a top, held together with small metal angle brackets or aluminium angle. Simple to do and you do then have proper screening etc. Done it many times when I didn't have metal bending facilities.
Some 2" U section aluminium is the plan I'm looking at for this.
Kei-1986 is offline  
Old 6th May 2020, 10:04 am   #95
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Mullard 5-20 high DC Voltages

"but I do wonder about the operating points on the EF86 vs the ECC83" yes, your right, & after having just read your DIYaudio thread I picked up on what someone said. Look at the DC figures, EF86 anode 57v , ECC83 grids 56.9v, so there's your problem straight off, one half of the ECC83 is biased too hot, way too hot, it's going into grid current and then blocking, acting like a diode. The other half is in cutoff at 51v.

So, you need to change the EF86 DC conditions so that the anodes of the EF86 V rises by about a volt. Pop a pot in series with the cathode resistor of the EF86, something like a 100r, connected as a reostat, EG wiper connected to one outer tab, adjust so anode V goes up, they should equalise.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 6th May 2020, 10:44 am   #96
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Mullard 5-20 high DC Voltages

It is not going to be the LTP.
The grid voltage for V2 g2 measures low because of the 10M input resistance of the DMM.
The current taken by the LTP anodes can be calculated from ohms law:
i=(Vc5 - Va) which gives us (409-337)/180000=400uA and (409-336)/180000=406uA. That is a very very near perfect match.

The problem is highly likely to be the inter pin capacitance in the EL34 valve socket. A small capacitance between pins 4 (g2) & 5 (g1) will be 'magnified' by the gain of the valve in a similar manner to the miller effect.

Do you have another socket you can try? Do you have a spare socket you could extract pin 5 from? You could then drill out the hole to give clearance and use the extracted pin to make the contact to the grid but you would also need to lift the connections off pin 6 and leave the joint floating.

PS: There is a non-destructive way to prove this is the problem by eliminating the signal from g2. Disconnect the g2 stopper from the ultra-linear tap connection and wire it to HT (centre tap of OT). You only need to do one EL34 to prove this. This has no effect on DC operating conditions.

Last edited by PJL; 6th May 2020 at 10:55 am.
PJL is offline  
Old 6th May 2020, 11:42 am   #97
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Mullard 5-20 high DC Voltages

Sorry, have to disagree there PJL, look at these voltages -

V1
(A - 57)
K - 0.54
G - 0

V2
A1 - 337
A2 - 336
(G1 - 56.9)
(G2 - 51.9)
K - 60.8
The anode of V1 is direct coupled to V2's g1, so Vgk = 0.1v!. I've never had a 10m IP Z load a grid down like that. The miss-matched grids is a classic example of an incorrectly biased LTP.

There is still an issue with stray C & or leakage inductance. There are three issue's as I sere it with this amp, 1) LTP misbiased 2) one EL34 red plating 3) ringing on OP/instability.

Also disagree that the Mullard 5-20 is a perfect amp, yes it could work straight off, but others with notable credibility have found the circuit lacking. I built the 5-10 as my first amp, after solving various issues like layout it still distorted badly, I now know the sensitivity is way too low for modern music sources, same applies to the 5-20. The choice of an ECC83 as LTP is the main fault, it's running at way to low an anode current - 180k anode R's! The better and beeffier the driver of the OP the better.

Andy.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 6th May 2020, 12:35 pm   #98
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Mullard 5-20 high DC Voltages

You can do alot with some basic maths:

The grid bias for V2a, Vgk = G1 - K = 56.9 - 60.8 = -3.9V. Looking at the datasheet for Va=280V and ia=0.4mA suggests it should be about -3V so pretty close.

All these readings are referenced from earth so we are trying to subtract some large numbers to obtain a small one. The mains voltage and consequently the HT is not stable so you should always expect some inaccuracy working this way, the correct method is to measure Vgk directly on the valve.

V2a has a grid voltage of 56.9V. When the DMM is connected to V2b grid it creates a potential divider formed by R10 (1M) and the DMM (10M). Ohms law for potential dividers says Vout= Vin*DMM/(DMM+R10) = 56.9*10000000/(10000000+1000000) = 51.73V.

The Mullard 5-20 is a well established and known to work design and was the basis of the original build and the output transformers are a good brand with the correct design specification for the 5-20. It should therefore be possible to get this amp working with the 5-20 standard circuit. We are making good progress and have seen a fault condition that is not related to the transformer.

I agree with you that a wooden chassis should be OK due to the use of an earth bus and care taken on the layout and wiring. I am sure the 5-20 can be improved upon but we need to stick with it and resolve the faults rather than try to workaround them.

Peter
PJL is offline  
Old 6th May 2020, 1:06 pm   #99
Techman
Dekatron
 
Techman's Avatar
 
Join Date: Nov 2009
Location: Lincolnshire, UK.
Posts: 4,985
Default Re: Mullard 5-20 high DC Voltages

Quote:
Originally Posted by Kei-1986 View Post
The chassis I was planning to build was never a complete one piece metal enclosure as the sides were going to be wooden with metal panels for the cutouts, leaving only the top and bottom as fully metal
The chassis needs to have full metal sides. I don't know if you're old enough to remember the CB radio craze of the late 70s and early 80s, where AM radio transmissions used to break through on music systems and amplifiers etc., but it's having a big revival again with AM once again becoming popular - there was a crowd chatting away on it in my area last night, and in fact on most nights if I have a listen round. If you've got any radio operators, be it CB or Ham radio near to you, then they'll be all over that un-screened amplifier like a rash - and it won't be their fault, so a proper full metal chassis build is a must!
Techman is offline  
Old 6th May 2020, 5:32 pm   #100
Herald1360
Dekatron
 
Herald1360's Avatar
 
Join Date: Feb 2007
Location: Leominster, Herefordshire, UK.
Posts: 16,526
Default Re: Mullard 5-20 high DC Voltages

Up to a point, but I doubt most 5-20s would have had chassis with welded seams and a tight fitting bottom plate and you would still have 5 little aerials and all the ironmongery up top to shield as well in order for a full metal jacket of a chassis to be worth the effort.

Judicious use of ferrite beads/ small resistors and small capacitors to keep stray RF out of the sensitive active bits would be my first addition after any problems with the official layout and stray RF.

Adding decorative wood to a standard chassis rather than using the wood structurally would be the best approach I think.
__________________
....__________
....|____||__|__\_____
.=.| _---\__|__|_---_|.
.........O..Chris....O
Herald1360 is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 5:16 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.