The main low end restriction is the 2.000 uF output capacitor. With a 4 Ohms load, response will be 6dB lower at 20 Hz. The input time constant, even with the original 0,68uF, will have the -6dB at around 10Hz. Quad is known for never been a big fan of wide bandwith products... ;)
Jorge Santos : true, but I wanted to see what the affect of the input cap was, since I was going to change it anyway. It is slight but visible. A bigger output cap would definitely make a huge difference.
Response will be -3dB at 20Hz in voltage terms. The output coupling capacitors were the biggest ones that would fit. You can easily double them or more nowadays,
Just repaired a 303 with a fault with a reduced output on right channel compared to left channel. Problem was the right channel decoupling cap C100, which was a tantalum type 0.68uF. Signal was reduced immediately after this input cap. The ESR was 150 ohm on the faulty cap. I also changed the electrolytics ( same as you ), as the capacitance value had increased 30% above original spec. I note that your amp is an early type, as mine has the Rev 10 driver boards, ie: > serial number 11,500. I'm hoping that I can use a similar trick to set the bias current to 5-10mA, to save having to disconnect the wiring to insert an ammeter. There isn't a lot of slack on the board wiring and the wires are susceptible to breakage at the entry hole on the board. During work, both sides had the end wire break when titling the board up and down. The wire insulation is almost bonded onto the internal wire, so had to be scraped off with a Stanley blade. Going to check online for a bias procedure for my later driver board. Thanks for the informative video.
Excellent video, thank you. I did the same thing on a Leak Stereo 30+ i.e. changed all the caps on one channel only. I found as you did that the main benefit was a flatter frequency response, and also faster rise time, but I can't honestly say I could hear a difference. Small point: I suspect you should measure voltages with an analogue meter (20k ohms/ volt) if you are to compare with the design figures; I use an old Avo 8 for this (great excuse to use a lovely piece of kit!).
Terry Nicklin : Hi Terry. I agree that the differences in sound often require better ears than mine to discern. However, most of the owners of this gear have in fact better ears, so I’m always happy to oblige. I also have an Avo 8 that I’ve restored, so should in fact use it more often.
1. All the necessary Bulgin connectors are still available and no problem at all. 2. Measuring the voltage between pins 6 and 1 is slightly invalid. 3. The difference between 3.8VRMS and 3.6VRMS is neither significant nor audible. 4. The LF response isn't affected in the slightest by the midpoint or bias adjustments, or by C101 or C106, but it is affected by C100 and C104 and the output capacitor, which should be doubled. 5. All Peter Walker's amplifiers follow a consistent theme of -3dB points at 15Hz and 55KHz, and the 303 is no exception, contrary to what is claimed in other comments here.
The manual recommends 6-10mA across pins 4-6. Given that the emitter resistors sum to 0R6, that means 3.6-6mV across those pins, not 6-10mV. You have it slightly overbiased. I usually end up at 5mV even with the aid of a distortion analyser. Above that it tends to run away, but it varies unit to unit. You will get a access to a bit more bass by doubling C104 to 22uF. This defines the pole of the DC servo, and it works in conjunction with the input capacitor and output caps to set the LF limit. Changing only one of those three by itself can't do much. There are persistent myths that Pater Walker designed these for the LS3/5A with nothing below 70Hz, etc., but they are just myths,
I am owner of a 303 and a 33 by Quad for more than 20 years. Unfortunately here in Rio de Janeiro is very dificult to find good technical people for repairs and therefore we usually just have 2 options for these kind of equipment by here: or to Pray or go to Sao Paulo city try find some technical people, what is not also an easy option also. Could you plese tell me what fuse especificatrions the 303 use at its back? I would like thank you for the nice video and so sorry my limited English.
O inglês não é problema entre países irmaos 😊 De acordo com o manual de serviço, o fusível é de 2A, lento. www.meridian-audio.info/public/quad-303-service-data%5B1961%5D.pdf
The main reason for the poor LF was that the 33/303 was designed to work with the electrostatic speaker. 90% of people what purchased Quad would also buy the complete system, as indeed I did. The original quad valve amplifier only produced about 15 watts, but when the 303 came out with 40 + watts the problems began. The transformer in the electrostatic would saturate and present a very low impedance which the amplifier did not like. The answer was to limit the LF response in both the power amp but mostly in the preamp (as you will find) This removed the issue but made it sound very lacking on standard loudspeakers, plus current limiting made the 303 a very un dynamic amplifier. At least these modes stopped people damaging the electrostatic and blowing up the 303. The bias needs to be on the high side to stop crossover effects showing up on a distortion analyser at 10Khz and over, slight but visible...audible...well that's debatable!!
_What_ 'poor LF'? The evidence is entirely against this claim. The 303 is -3dB at 15Hz, and the 33 at about 20Hz, mainly I believe as an anti-rumble measure, and they actually _lowered_ the LF pole in the 33 during production. And an LF pole in a preamp isn't much of a strategy to protect a loudspeaker.
michael - you sound like u know your shit - if only i understood your language. i have a couple of 303's paired with ditton 44's and i bypass any preamps altogether- and to tame the sound i've got a dacmagic100 feeding it. i'm getting audio definition i've never thought possible digitally - and its risen again after my last 303 arrived the other day from Ebay and by all accounts its been in storage for 20 years and belonged to the vendors good friend prior to that which explains, atleast to me, how perfection got owned .. - and i find flaws in most 'stuff' - where do i go from here ? - - also, electrostatic speakers - ?
This is urban myth. It is certainly untrue that 90% of 33/303 customers also bought ESLs. The sales figures tell you that. 120,000 33s, 94,000 303s, 56,000 ESLs, and many of those were driven by Radford, McIntosh, etc. All Quad power amplifiers are -3dB at about 14Hz.
Did you mean to adjust the 67V supply with a dim bulb in circuit? That's what the video implied. I don't think the original -1dB at 30Hz is bad, do you?
I have one of these amps in for repair at the moment. I would be interested to see how you input the signal to the amp and how you connected the scope to each channel. Did you use a dummy load? Great video though. I was thing of trying some audiophile caps to see if I could get any measured improvements. Thanks Graham
Radiocruncher : the inputs on mine are via the two rca sockets, but this was a mod. The original has a proprietary socket and you can follow the wiring into the two amp boards via a small screened cable. It’s fairly straightforward. I made a ‘box’ that received the speaker out connectors and then directs the signal either to a set of speakers or to a dummy load (2x 8ohm resistors that I use as 8 ohms or short to make it 4 ohms. The same applies to the other channel, and both are completely separate as far as ground is concerned. I tap the signal prior to the speaker/dummy load Re-direct from each side and connect to two bnc connectors to be able to connect each side to the scope. Beware of the scope part: if you connect a scope you effectively short the negative to earth, which may not be acceptable in some cases. Also, even if your scope is on an isolation transformer, if you connect BOTH scope connectors, your are again shorting the two speaker grounds to each other. In this case I believe everything is referenced to the same ground anyway so it won’t be a problem, but I wired it this way to leave my options open when working with amps that have crazy grounding schemes. This is a fairly simple amp and there is los of info available on the web including setup and biasing instructions. Good luck.
Thank you very much for the comprehensive reply. I will see if I can rig something up to test them. I assume you used an AF generator or waveform generator for the inputs. Thanks Graham
I have just started the refresh of my 1979 303 sr 85849, the four large capacitors have been replaced. One had bulged and others had evidence of leaking. I noticed that you used the current limiter with a 40-watt bulb. Was the current limiter still in the circuit when you measured the voltages and bias currents?
@@electronicsoldandnew That's what I thought, because the voltage drop across the limiter would cause a lower voltage input to the mains transformer and therefore a lower output voltage. Lovely tutorial very very clear and concise.. Are there many Quads in Madeira???. Plenty where I live in England !
I just picked one of these 303/33 combos up to go with the Quad ESL 57s I’ve had in storage for about 20 years. Tell me, is it dangerous to remove the cover of these amps and have a peek inside? What precautions do you have to take? Thanks.
![Megan Thee Stallion - Neva Play (feat. RM) [Official Video]](http://i.ytimg.com/vi/TpYTyAaTRts/mqdefault.jpg)
Thank you for a comprehensive description of the completed work... Very good indeed. !
Thanks
The main low end restriction is the 2.000 uF output capacitor. With a 4 Ohms load, response will be 6dB lower at 20 Hz. The input time constant, even with the original 0,68uF, will have the -6dB at around 10Hz. Quad is known for never been a big fan of wide bandwith products... ;)
Jorge Santos : true, but I wanted to see what the affect of the input cap was, since I was going to change it anyway. It is slight but visible. A bigger output cap would definitely make a huge difference.
Response will be -3dB at 20Hz in voltage terms. The output coupling capacitors were the biggest ones that would fit. You can easily double them or more nowadays,
Thank you for a good video
Pleasure
Just repaired a 303 with a fault with a reduced output on right channel compared to left channel. Problem was the right channel decoupling cap C100, which was a tantalum type 0.68uF. Signal was reduced immediately after this input cap. The ESR was 150 ohm on the faulty cap. I also changed the electrolytics ( same as you ), as the capacitance value had increased 30% above original spec. I note that your amp is an early type, as mine has the Rev 10 driver boards, ie: > serial number 11,500. I'm hoping that I can use a similar trick to set the bias current to 5-10mA, to save having to disconnect the wiring to insert an ammeter. There isn't a lot of slack on the board wiring and the wires are susceptible to breakage at the entry hole on the board. During work, both sides had the end wire break when titling the board up and down. The wire insulation is almost bonded onto the internal wire, so had to be scraped off with a Stanley blade. Going to check online for a bias procedure for my later driver board. Thanks for the informative video.
Good luck with yours. The wire breakage is a real problem, and can catch you out if you’re not aware if the problem.
Ok, it seems that the only difference is that the newer board uses better bias circuitry but the process for measuring is the same. Thanks again
👍
C100 is a coupling cap, not a decoupling cap. Once those wires start to break, another will break, and another. Keep a close watch.
Excellent video, thank you. I did the same thing on a Leak Stereo 30+ i.e. changed all the caps on one channel only. I found as you did that the main benefit was a flatter frequency response, and also faster rise time, but I can't honestly say I could hear a difference. Small point: I suspect you should measure voltages with an analogue meter (20k ohms/ volt) if you are to compare with the design figures; I use an old Avo 8 for this (great excuse to use a lovely piece of kit!).
Terry Nicklin : Hi Terry. I agree that the differences in sound often require better ears than mine to discern. However, most of the owners of this gear have in fact better ears, so I’m always happy to oblige. I also have an Avo 8 that I’ve restored, so should in fact use it more often.
The difference in sound after recaping should be night and day! If you dont hear it you have a problen!
Thanks for the video, I've got a bid going on eBay for one of these.
Good luck 👍
1. All the necessary Bulgin connectors are still available and no problem at all. 2. Measuring the voltage between pins 6 and 1 is slightly invalid. 3. The difference between 3.8VRMS and 3.6VRMS is neither significant nor audible. 4. The LF response isn't affected in the slightest by the midpoint or bias adjustments, or by C101 or C106, but it is affected by C100 and C104 and the output capacitor, which should be doubled. 5. All Peter Walker's amplifiers follow a consistent theme of -3dB points at 15Hz and 55KHz, and the 303 is no exception, contrary to what is claimed in other comments here.
EJP : thanks for your input.
The manual recommends 6-10mA across pins 4-6. Given that the emitter resistors sum to 0R6, that means 3.6-6mV across those pins, not 6-10mV. You have it slightly overbiased. I usually end up at 5mV even with the aid of a distortion analyser. Above that it tends to run away, but it varies unit to unit. You will get a access to a bit more bass by doubling C104 to 22uF. This defines the pole of the DC servo, and it works in conjunction with the input capacitor and output caps to set the LF limit. Changing only one of those three by itself can't do much. There are persistent myths that Pater Walker designed these for the LS3/5A with nothing below 70Hz, etc., but they are just myths,
my fav stereo next to b&o
Dan Ludington: yes, it is quite special.
i agree
I am owner of a 303 and a 33 by Quad for more than 20 years. Unfortunately here in Rio de Janeiro is very dificult to find good technical people for repairs and therefore we usually just have 2 options for these kind of equipment by here: or to Pray or go to Sao Paulo city try find some technical people, what is not also an easy option also. Could you plese tell me what fuse especificatrions the 303 use at its back? I would like thank you for the nice video and so sorry my limited English.
O inglês não é problema entre países irmaos 😊
De acordo com o manual de serviço, o fusível é de 2A, lento.
www.meridian-audio.info/public/quad-303-service-data%5B1961%5D.pdf
@@electronicsoldandnew , muito obrigado pela sua resposta!! Espero encontrar o fusível por aqui no Rio. Grande abraço e obrigado!!
prazer
The main reason for the poor LF was that the 33/303 was designed to work with the electrostatic speaker. 90% of people what purchased Quad would also buy the complete system, as indeed I did. The original quad valve amplifier only produced about 15 watts, but when the 303 came out with 40 + watts the problems began. The transformer in the electrostatic would saturate and present a very low impedance which the amplifier did not like. The answer was to limit the LF response in both the power amp but mostly in the preamp (as you will find)
This removed the issue but made it sound very lacking on standard loudspeakers, plus current limiting made the 303 a very un dynamic amplifier. At least these modes stopped people damaging the electrostatic and blowing up the 303.
The bias needs to be on the high side to stop crossover effects showing up on a distortion analyser at 10Khz and over, slight but visible...audible...well that's debatable!!
That makes a lot of sense. Hadn’t thought of the problem of the ESL’s overloading this little amp.
_What_ 'poor LF'? The evidence is entirely against this claim. The 303 is -3dB at 15Hz, and the 33 at about 20Hz, mainly I believe as an anti-rumble measure, and they actually _lowered_ the LF pole in the 33 during production. And an LF pole in a preamp isn't much of a strategy to protect a loudspeaker.
michael - you sound like u know your shit - if only i understood your language.
i have a couple of 303's paired with ditton 44's and i bypass any preamps altogether- and to tame the sound i've got a dacmagic100 feeding it. i'm getting audio definition i've never thought possible digitally - and its risen again after my last 303 arrived the other day from Ebay and by all accounts its been in storage for 20 years and belonged to the vendors good friend prior to that which explains, atleast to me, how perfection got owned .. - and i find flaws in most 'stuff' - where do i go from here ? - - also, electrostatic speakers - ?
This is urban myth. It is certainly untrue that 90% of 33/303 customers also bought ESLs. The sales figures tell you that. 120,000 33s, 94,000 303s, 56,000 ESLs, and many of those were driven by Radford, McIntosh, etc. All Quad power amplifiers are -3dB at about 14Hz.
Did you mean to adjust the 67V supply with a dim bulb in circuit? That's what the video implied.
I don't think the original -1dB at 30Hz is bad, do you?
No, the final voltage adjustment must be done at full line voltage. I think the low frequency response is very good.
I have one of these amps in for repair at the moment. I would be interested to see how you input the signal to the amp and how you connected the scope to each channel. Did you use a dummy load? Great video though. I was thing of trying some audiophile caps to see if I could get any measured improvements. Thanks Graham
Radiocruncher : the inputs on mine are via the two rca sockets, but this was a mod. The original has a proprietary socket and you can follow the wiring into the two amp boards via a small screened cable. It’s fairly straightforward.
I made a ‘box’ that received the speaker out connectors and then directs the signal either to a set of speakers or to a dummy load (2x 8ohm resistors that I use as 8 ohms or short to make it 4 ohms. The same applies to the other channel, and both are completely separate as far as ground is concerned.
I tap the signal prior to the speaker/dummy load Re-direct from each side and connect to two bnc connectors to be able to connect each side to the scope. Beware of the scope part: if you connect a scope you effectively short the negative to earth, which may not be acceptable in some cases. Also, even if your scope is on an isolation transformer, if you connect BOTH scope connectors, your are again shorting the two speaker grounds to each other. In this case I believe everything is referenced to the same ground anyway so it won’t be a problem, but I wired it this way to leave my options open when working with amps that have crazy grounding schemes.
This is a fairly simple amp and there is los of info available on the web including setup and biasing instructions.
Good luck.
Thank you very much for the comprehensive reply. I will see if I can rig something up to test them. I assume you used an AF generator or waveform generator for the inputs. Thanks Graham
Radiocruncher : yes, signal generator at any reasonable frequency like 1kHz, Ana start at about 20mV amplitude then go up.
Electronics Old and New by M Caldeira It's not s proprietary socket, it is a standard 4-pin DIN Socket, available anywhere.
I have just started the refresh of my 1979 303 sr 85849, the four large capacitors have been replaced. One had bulged and others had evidence of leaking. I noticed that you used the current limiter with a 40-watt bulb. Was the current limiter still in the circuit when you measured the voltages and bias currents?
No, when you set that up you must have the full supply on.
@@electronicsoldandnew That's what I thought, because the voltage drop across the limiter would cause a lower voltage input to the mains transformer and therefore a lower output voltage. Lovely tutorial very very clear and concise.. Are there many Quads in Madeira???. Plenty where I live in England !
There actually are quite a few. Much more that I ever expected for such a small place.
I just picked one of these 303/33 combos up to go with the Quad ESL 57s I’ve had in storage for about 20 years. Tell me, is it dangerous to remove the cover of these amps and have a peek inside? What precautions do you have to take? Thanks.
beware of shocks 😊
Anywhere there is mains wiring, you’re risking a zap ⚡️
It seems that there is quite a lot of wind blowing around your house, is it? You can hear it clearly.
Jacob Fikse : I live on the top floor of an apartment building, so with the great views we also have some wind. A fair trade, I think.
May I ask what is the best option for C100 component replacement?
Thanks :)
Luke Konieczny : I used 1uF film cap.
Do not buy the preamp. It is a real piece of sh..