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 18th May 2023, 6:59 pm   #1
GlennJobson
Triode
 
Join Date: May 2023
Location: Cambridge, Cambridgeshire, UK.
Posts: 13
Default Quad 303 Voltage Query

I have been looking over a Quad 303 with some channel distortion.

The unit was last serviced by Quad in 1982, the owner said reasonably recently.

I have replaced the three electrolytics in each driver board and two 2200uF caps too, all had visible signs of leaking, the amplifier was quite a mess inside with a lot of fluff and dust.

After some adjustment of the voltages I have 67V on both driver boards and great sound on the right channel.

Looking over the left driver board I have the 67V but instead of the 33V on the schematic I can only see around 3V.

Measuring TR1L with respect to the schematic I see 67V where is shows connection to pins 2 (C) and 3 (B) but 3V on pin 4 (E). This of course makes all other comparable reading across the board that are not 67V way off.

Comparing a well performing channel (R) to the poor performing channel (L) I am now rather stuck. This may be a bit vague but what should I be looking at next to try and locate the 'problem'.

Many thanks if anyone offers any advice...
GlennJobson is offline  
Old 18th May 2023, 10:19 pm   #2
John_BS
Octode
 
John_BS's Avatar
 
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
Default Re: Quad 303 Voltage Query

Welcome to the forum.

The problem with DC-coupled amplifiers incorporating negative feedback is that once a problem occurs, all the DC voltages will be incorrect, even in stages which are perfectly sound.

Laborious as it sounds, if you can measure the BCE voltages of all the transistors and report back, we may be able to give you a steer.

John
John_BS is offline  
Old 19th May 2023, 6:24 am   #3
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Quad 303 Voltage Query

The Quad 303 has a DC blocking capacitor at its input, and a (large) DC blocking capacitor at its output, so overall the thing is AC coupled and this prevents any DC errors being applied to the speaker.

Between these two DC blocking capacitors, the main part of the amplifier is DC coupled. It has to be.

Transistors are not by themselves linear devices. To get their distortion down to acceptable levels feedback is needed - either global feedback around the whole thing, or degeneration on a local, stage-by-stage basis. To get the distortion down low enough, a LOT of feedback is needed, much more than will be found in traditional good quality valve amplifiers. So the main part of the amplifier needs to be very high gain in order to support all this feedback. Transistors have less gain per stage than typical valve amplifiers, so that means plenty of stages.

While the main section of a transistor amplifier could be done as a series of AC-coupled stages, that would mean several DC blocking capacitors and a separate bias arrangement for each individual stage. The whole thing is already getting a much larger component count than trad valve amps and then on top of this are the memories of the troubles with bias instability in early transistor amplifiers - thermal runaway.

The solution to these problems is to have only two bias arrangements working. One is to control the quiescent current down the output transistors, the other is to position the output voltage into the speaker DC block at about half the supply rail voltage. This leaves the output stage equal space to swing in each direction.

So the overall feedback around the amplifier is used to control the gain applied to signals, to reduce the distortion, to reduce the output impedance (= increase damping factor) and on top of all that it has to control the DC before the output blocker, keeping it near the midpoint. This doesn't make it terribly complicated, but it does make it terribly important.

The output blocker saves the speaker drom destruction if the DC conditions in the amplifier are slewed away from the midpoint. If something goes wrong, then you don't know about it until it's slewed so far off centre that the signal gets clipped early on one direction of swing or the other as the amplifier hits its endstops.

Because the amplifier is set up for so much forwards gain, without the DC aspect of the feedback system working to control the DC gain of the forward stages of the amplifier, it would be impossible to bias the input to reliably get the output biased at its midpoint. In other words, if the feedback loop is opened, the output will bang itself hard against one extreme or the other, trying to set the output as close as it can get to either 0v or 67v.

This is a bit of a devil for fixing a failed amplifier.

Because it's a loop, if things are wrong, the cause could be anywhere in the loop and it will throw everything else around the loop off of where it should be. Deducing the cause is very difficult. If you try opening the loop by disconnecting the feedback path, even an undamaged amplifier will ram itself against the end stops from quite normal bias inaccuracy at the front-end.

The DC coupling amongst the transistors allows one faulty transistor to damage its neighbours.

This leaves a problem. You can't confidently test a transistor and just replace it because if it wasn't the cause your new one will fall victim the moment you switch on. If it was the cause, others may be damaged and they can still kill your new one. It is possible to get yourself trapped in an endless circle of keeping replacing parts.

Two ways can get you out of this trap. Good deductive skills may lead you to understanding which stages are not working rather than simply being biased off by a different stage. This is difficult and not always possible. The other approach is to keep power off and check all semiconductors either in-circuit or temporarily removed and then to replace all bad ones AND ANY SUSPECT ONES all at once before power is ever re-applied. Sometimes it's simpler to just replace the whole lot in a oner. Small signal transistors are pretty cheap.

Here we come to a problem specific to the Quad 303. The TO-3 cased output transistors are early versions of the very common 2N3055. They are used in some sub-circuits usually referred to as the "Quad triples" which is an odd name in itself. These are little feedback loops in themselves and their stability relies on the 2N3055 being rather slow and really dominating the speed of the little loops. Newer 2N3055 devices are a lot faster and are too fast to make the quad triples reliably stable. Fitting new devices often results in bursts of oscillation at some points in a signal waveform - oscillation at several megahertz. This can make finding bad enough output transistors for use as replacements in these amplifiers a bit of a problem.

So with a 303, check the big output transistors and if they're OK, get on with checking the small signal devices. You can also check the DC voltage into the base of the first transistor - this is where the centre point bias starts.

Sorry, it's a lot to dump on you, but I've tried to explain the whys and wherefores because otherwise some of the things with transistor amplifiers can look rather silly.

Transistor amps can be fixed, but they require a different approach. There isn't any fast way that doesn't require luck, so you have to grind your way round the long way.

The 303 was really the first transistor amplifier that was generally accepted as being good.

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 19th May 2023, 6:53 am   #4
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Quad 303 Voltage Query

That is of course full of David's wisdom. But damping factor is a red herring.

It was a term originally coined by Fritz Langford-Smith, the editor and author of the truly excellent Radiotron Designer's Handbook.

He then refuted the term in a letter to Wireless World back in the late 40's (I don't have the specific date) when he recognised that the dominant impedance was the loudspeaker's voice coil.

But it lives on regardless through the decades - with audio spec wars regarding Titanicically large and larger damping factors.

That this is indeed a red herring, consider current driving the speaker. This has a long history of success, since the force on the cone is proportional to the current flowing in the voice coil, not the voltage applied. That drives the speaker from a high driving impedance - from the collectors of the output pairs rather than the emitters. There are no damping related problems associated.

Craig
__________________
Doomed for a certain term to walk the night
Craig Sawyers is offline  
Old 19th May 2023, 7:21 am   #5
G4_Pete
Hexode
 
Join Date: Feb 2014
Location: Southampton, Hampshire, UK.
Posts: 419
Default Re: Quad 303 Voltage Query

Also worth mentioning I have a letter here from Quad service dated Sept 1998 stating that the 2N3055,s were specifically tested and selected and that the replacement is the MJ15003. I also have a pencil note on the schematic showing a 0.1 uF fitted from the the top of the R104 (67V rail ) to the input earth pin 10 on the schematic " To cure Instability on peaks" .But cant remember if that was after fitting the MJ devices or how successful that was!! This was a 303 that came through, not owned by me.
G4_Pete is offline  
Old 19th May 2023, 8:01 am   #6
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Quad 303 Voltage Query

If an amplifier is designed to give a reasonably low output impedance, it at least gives the speaker crossover designer a known point to start from rather than throwing yet another uncertain variable at the poor soul. Fighting for ultra-low output impedance is a waste of time.

Damping factor is a terrible name. An output Z of zero or infinite Ohms can do no damping of speaker resonances whatsoever. Zero and infinity resistive impedances can dissipate no real power at all, therefore there can be no damping, no extraction of energy from resonances, no reduction of their Q.

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 19th May 2023, 8:21 am   #7
John_BS
Octode
 
John_BS's Avatar
 
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
Default Re: Quad 303 Voltage Query

This is interesting but heading off-topic and we should start a new thread under Components & Circuits

John
John_BS is offline  
Old 19th May 2023, 9:01 am   #8
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Quad 303 Voltage Query

Quote:
Originally Posted by Radio Wrangler View Post
If an amplifier is designed to give a reasonably low output impedance, it at least gives the speaker crossover designer a known point to start from rather than throwing yet another uncertain variable at the poor soul. Fighting for ultra-low output impedance is a waste of time.

Damping factor is a terrible name. An output Z of zero or infinite Ohms can do no damping of speaker resonances whatsoever. Zero and infinity resistive impedances can dissipate no real power at all, therefore there can be no damping, no extraction of energy from resonances, no reduction of their Q.

David
Take your point regarding crossovers.

Yup regarding energy dissipation. There is a wide misunderstanding that a low output impedance in some way controls back emf - which it does not. Since the efficiency of a moving coil driver is around 1% (because the motional impedance of the cone is far removed from the impedance of the air load) the real world damping is largely done by the voice coil - it dissipates 99% of the power coming out of your amp, and heats up as a result.

The only way you can increase the coupling efficiency to the air is using an impedance transformer - AKA a horn. The problem is that to get real bass involves truly massive horns. Or a transmission line such as Nelson Pass's https://www.firstwatt.com/pdf/art_elpipeo.pdf

Craig
__________________
Doomed for a certain term to walk the night
Craig Sawyers is offline  
Old 20th May 2023, 8:06 pm   #9
GlennJobson
Triode
 
Join Date: May 2023
Location: Cambridge, Cambridgeshire, UK.
Posts: 13
Default Re: Quad 303 Voltage Query

Thanks for all the replies, looks like I have more work to do, I will report back...
GlennJobson is offline  
Old 21st May 2023, 10:17 am   #10
GlennJobson
Triode
 
Join Date: May 2023
Location: Cambridge, Cambridgeshire, UK.
Posts: 13
Default Re: Quad 303 Voltage Query

OK, Sunday morning and a multimeter... I have looked over the left driver board, these are the reading once the amp has been on for 10 minutes. Not sure how these should be presented, I have seen all sorts on here, perhaps this is another confusing variation... sorry if that is the case.

TR100 B 4.4 C 5.0 E 2.0
TR101 B 0.6 C 0.0 E 1.4
TR102 B 1.4 C 2.0 E 0.7
TR103 B 3.3 C 2.9 E 67.0
TR104 B 2.0 C 1.2 E 2.7
TR105 B 3.4 C 67.0 E 2.8
TR106 B 1.2 C 2.8 E 0.5
TR107 B 2.7 C 2.0 E 3.3

TR1L B 3.0 C 67.0 E 2.8
TR2L B 0.5 C 3.0 E 0.0

Many thanks.....

Glenn
GlennJobson is offline  
Old 21st May 2023, 12:08 pm   #11
G4_Pete
Hexode
 
Join Date: Feb 2014
Location: Southampton, Hampshire, UK.
Posts: 419
Default Re: Quad 303 Voltage Query

Hi Glenn ,You might find this useful I drew it out years ago when I worked on these amps. I suspect someone else has done a better one by now but I could not find one on a quick scan this morning hence posting mine.

If it is not readable I will try to enhance it a bit, also treat it as a guide as it was accurate to the best of my abilities at the time!!

Also check the tantalum capacitor C100 has not gone short circuit , as David said sometimes you get lucky.

Pete
Attached Thumbnails
Click image for larger version

Name:	QUAD 303.jpg
Views:	63
Size:	88.6 KB
ID:	278537  

Last edited by G4_Pete; 21st May 2023 at 12:13 pm. Reason: added accuracy comment.
G4_Pete is offline  
Old 21st May 2023, 12:34 pm   #12
GlennJobson
Triode
 
Join Date: May 2023
Location: Cambridge, Cambridgeshire, UK.
Posts: 13
Default Re: Quad 303 Voltage Query

Quote:
Originally Posted by G4_Pete View Post
Hi Glenn ,You might find this useful I drew it out years ago when I worked on these amps. I suspect someone else has done a better one by now but I could not find one on a quick scan this morning hence posting mine.

If it is not readable I will try to enhance it a bit, also treat it as a guide as it was accurate to the best of my abilities at the time!!

Also check the tantalum capacitor C100 has not gone short circuit , as David said sometimes you get lucky.

Pete
Thanks Pete, that is brilliant, I have done something similar but I am left handed so once written I can't read it!
GlennJobson is offline  
Old 21st May 2023, 12:49 pm   #13
G4_Pete
Hexode
 
Join Date: Feb 2014
Location: Southampton, Hampshire, UK.
Posts: 419
Default Re: Quad 303 Voltage Query

Hi I have done another slightly clearer one
Pete
Attached Thumbnails
Click image for larger version

Name:	quad 303 scan 3.jpg
Views:	53
Size:	88.5 KB
ID:	278539  
G4_Pete is offline  
Old 21st May 2023, 12:57 pm   #14
fetteler
Octode
 
fetteler's Avatar
 
Join Date: Jul 2011
Location: Staffordshire Moorlands, UK.
Posts: 1,464
Default Re: Quad 303 Voltage Query

TR 103 appears to have 63.7V between the base and emitter, Something's clearly very wrong there and I would suspect the B-E has gone open circuit otherwise with that many volts across it it would be on fire!

The B-E voltage of Tr100 at 2.4V is also high...

Steve.
__________________
Those who lack imagination cannot imagine what is lacking...
fetteler is offline  
Old 21st May 2023, 12:58 pm   #15
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Quad 303 Voltage Query

The layout depends on what generation of board it is. The later boards were issue 10!

The most significant difference between earlier and later boards was the addition of a Vbe muliplier, repositioning the quiescent current pot and some component value changes and physical positions.

I changed the input tant on mine as a matter of course, replaced both by plastic film (probably polyester).

Craig
__________________
Doomed for a certain term to walk the night
Craig Sawyers is offline  
Old 21st May 2023, 1:50 pm   #16
John_BS
Octode
 
John_BS's Avatar
 
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
Default Re: Quad 303 Voltage Query

Glenn: can you double-check the Tr terminals ?
For instance, the collector of 101 must be the same potential as base of 102. Ditto, collector 103 = base 105 etc
John_BS is offline  
Old 21st May 2023, 1:53 pm   #17
G4_Pete
Hexode
 
Join Date: Feb 2014
Location: Southampton, Hampshire, UK.
Posts: 419
Default Re: Quad 303 Voltage Query

Hi the layout was from a board marked M12038-10, as it has a TR107 that would appear to be the later one from ser number 11500 onwards.

Pete
G4_Pete is offline  
Old 21st May 2023, 2:06 pm   #18
John_BS
Octode
 
John_BS's Avatar
 
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
Default Re: Quad 303 Voltage Query

1. Check C101 / is in the right way round ?
2. try replacing Tr100 with a general purpose pnp e.g 2N2906 or similar.
John_BS is offline  
Old 21st May 2023, 7:09 pm   #19
GlennJobson
Triode
 
Join Date: May 2023
Location: Cambridge, Cambridgeshire, UK.
Posts: 13
Default Re: Quad 303 Voltage Query

OK, a couple of corrections from me.... Thanks for the help. I had misidentified a couple of transistors...

TR103 B 3.7 C 67.0 E 2.9

John BS - For instance, the collector of 101 must be the same potential as base of 102. Ditto, collector 103 = base 105 etc

That is the case, another correction

Collector 103 = Base 105 = 67.0V
Collector 101 = Base 102 = 1.4V

>The B-E voltage of Tr100 at 2.4V is also high...

TR100 B 4.5V C 1.9V E 5.1V

The diagram from Pete was super valuable.
GlennJobson is offline  
Old 21st May 2023, 8:16 pm   #20
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Quad 303 Voltage Query

The bias conditions for TR100 are set like this:

The voltage across R107 has to be about the base emitter voltage of TR101. So the current through R107 has to be about 0.6V, giving the collector current of 0.6/2200 = 0.27mA, giving a collector voltage of 2V. Which is what you have (1.9V).

The emitter voltage is largely set by the potential divider R102 and R104. That, for 67V on the plus rail gives an emitter voltage of about 10V (the schematic says about 9V).

Now you get 5.1V - which is wrong.

What is interesting is the emitter current is right (from the 2V), and the base emitter voltage is right (5.1-4.5 = 0.6V - remember this is a pnp transistor).

The only way for this to be wrong is for either C100 (0.68u) or C101 (either 300u or 470u) to be leaky.

Craig
__________________
Doomed for a certain term to walk the night
Craig Sawyers is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 8:19 am.


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.