Quad 303 Power Amplifier - Power Supply Repair

[Jay_Diddy_B]

Hi Group,

I am going to share a repair that I performed on a Quad 303 Power Amplifier. The 303 was introduced in 1967 and was sold until 1985. The unit that I repaired was built around 1977.

I bought the unit locally for a reasonable price. It looks like no modifications have been done the unit.

I open the unit and found that there was evidence of overheating on the power supply board:









I found that are a few place offering replacement boards for this amplifier. But, I decided that I would build a replacement board using as many of the original components as possible.
I found the schematic on the internet:




And I redrew the section that I needed:



Then I generated the artwork for the board. I took measurements from the original board. Here is a picture showing the tool path generated by the LPKF software:



Here is a picture of the new board:




The assembled board, using many of the original components:



The board installed in the amplifier:




I adjusted the pot to set the output voltage to 67V.

I gave the amplifier a quick test it seems o.k. The next step is to replace the four large electrolytic capacitors. Two are used in the power supply and two are in series with the speaker outputs.

Regards,

Jay_Diddy_B

[rbm]

Cheers Jay Diddy!! I have a Quad 405, 33 and 303 in my stash.  I love the simplicity of Quad's audio gear and appreciate that others are contributing to their longevity.

[oldway]

Something strange: evidence of overheating seems not to be an issue of the board, nor of his components, but from something other....Perhaps something under the board.....a transformer ?....Picture is not clear enough to see it.....It should be investigated as the new board will suffer the same problem...

[tronde]

Something strange: evidence of overheating seems not to be an issue of the board, nor of his components, but from something other....Perhaps something under the board.....a transformer ?....Picture is not clear enough to see it.....It should be investigated as the new board will suffer the same problem...

The evidence is visible in the picture of the new PCB. One resistor with long legs.

[SeanB]

R204, looks like it should have been a 10k 2W resistor there, not a 0.5W unit.

[Jay_Diddy_B]

Hi,

Sean is right, the hot resistor was R204. The value of R204 may have changed over the product life of the unit.

The original value was 6.8k. The voltage across the resistor is 67V - 12V = 55V.

So the original dissipation was 55 x 55 / 6800 = 0.45W

Some of the schematics show R204 as 10K

The power dissipation is 55 x 55 / 10000 = 0.3W

I stood the resistor off the board as precaution. I will probably replace the resistor with one with a larger body size. I think the overheating was caused by running the amplifier resting on carpet. The carpet blocks the lower cooling holes. Also the original board was SRBP.

It is also possible that the trimpot failed, it was overheated. If the wiper went open, there would be a further increase in dissipation and temperature rise.

Here are some pictures of the dead board:



You can see that the resistor has a blue band, probably 6.8k




Regards,

Jay_Diddy_B

[Jay_Diddy_B]

Hi Group,

I did a little more work on my Quad 303 today. I replaced the electrolytic capacitors. The original capacitors 2200uF 100V were manufactured in 1977. There are a total of four. Two are connected in parallel in the power supply and one in series with each loudspeaker.






The capacitors in the power supply circuit were bulging.

I decided to buy new capacitors from Digikey rather risk getting counterfeits (or worse) from eBay. I chose these parts:



The specifications are:

4700µF 100V Aluminum Electrolytic Capacitors Radial, Can - Snap-In 53 mOhm @ 120Hz 2000 Hrs @ 105°C

These are good quality industrial capacitors that cost $6.25 USD each.

These capacitor were metric in a 35mm diameter can. They are slightly smaller than 1.5 inch diameter original capacitors.
I wrapped some strips of paper around the new capacitors so they would fit the original clamps.
I mounted the capacitors in the same orientation as the originals with the vents at the top.



Regards,

Jay_Diddy_B

[nfmax]

Well done!

I used to work at QUAD, a long, long, time ago. There were many changes to the 303 over it's life, including changing the orientation of the main electrolytics. Early units hand them mounted tags (& vent) down. This was subsequently changed to improve their reliability. I have a paper copy of the service notes, but I think it may be slightly older than the version you found online.

They are still very good audio amplifiers. PJ was a very talented engineer!

[Jay_Diddy_B]

nfmax and the group,

It is good to meet to somebody who had some insight into why things were done a certain way. I was made sensitive to the vent location issue at a company that I worked at. They had used Sangamo computer grade electrolytic capacitors. They learned the hard way that the capacitor core was held in the can with pitch. They had mounted some capacitors vent down. If you ran the capacitor hot, the pitch would flow and block the vents. The vents no longer worked and the capacitors would fail without venting 

I have several pieces of Quad equipment:

1) 2x FM4 tuners

2) 34 Preamplifier

3) 606 mk II power amplifier

4) 66 preamp with IR remote control. I cannot get this to work in a reliable fashion. I am having problems with the display (digital board). The board is double sided but it uses eyelets (rivets) for the through connections. I take it apart reflow some of the connections, it works for a while then it goes bad. Short of cloning the whole board I don't know how to fix this.

I would like to find another 34 or a 44 preamp.

Who is PJ? Is P.J. Baxandall ( of tone control fame)?

Regards,

Jay_Diddy_B
 

[nfmax]

Who is PJ? Is P.J. Baxandall ( of tone control fame)?

PJ is of course P J Walker, the company founder http://www.aes.org/aeshc/docs/jaes.obit/JAES_V52_3_PG316.pdf. I joined the company in 1982 as a (the) junior design engineer, and stayed for three years or so. I got the chance to work with PJ and Mike Albinson on a range of things, including the 34 control unit, which was near completion, and various items of production equipment; troubleshooting production issues with the ESL-63; and designing a new output transistor protection circuit for the 405 power amplifier, turning it into the 405-2.

Peter Baxandall was an occasional visitor to Huntingdon, and we used to meet up at AES meetings in London as well.

It all seems such a long time ago!

[PeterSorvag]

Hi all,

I'm new here, my name is Peter.
I found this thread, and I just had to sign up :-)

It is an old thread, I know, but I hope you guys are still there.

I bought on old 303, and of course the voltage and the bias were totally off, but adjusting it, it plays very well.
But some components has to be replaced. Among them, the psu and output capacitors .
Nfmax, fantastic to hear that you worked with these guys :-)
I have a question: Restoring these amplifiers mostly includes increasing the values on the filter and output capacitors. I may be stupid, but why is that desirable ? The 303 I bough, the one of the filter capacitors is starting to bulge, and I guess they have been in there since ancient time, but the amp is still very silent ! If they loose value during the years, as we know they do, they must have been over dimensioned to begin with . No?
And the output capacitors with 2000 uF should be straight down to 20 Hz at 4 ohms, or do I miscalculate ?
If I put in 10000 uF per channel, as some recommend, don't I change the conditions for the drivers at the same time ?

When I have been talking to constructors of loudspeakers about upgrading their old constructions, wheather is is Mats at QLN or Stig Carlsson, they mean that every construction is a delicate balance between different behavior with different components. To put in modern speaker elements and new filters just because they are new, and measure better, and can take more effect, doesn't make it a better speaker.
Instead it can ruin years of research, how to overcome the limitations and problems that exists in every construction.

An inventor of Peter Walker's format, he must have been aware of that the C 100 was taking a little slice of the peaks under 30 Hz.

Of course I want my 303 sound as good as possible, I just don't want the magic to go away :-)

All the best
Peter

[PeterSorvag]

I recapped my Quad today, and I thought "what the heck", and I put in 4700 uF on the output, and changed C4 to 22 uF. I left C 100 as it was. Its recommended that you change that to have more bass from 0.68 to 1 uF, but I wanted to hear what the other changes made with the sound.

In my opinion , this changed the character of the sound, from a very dry and defined sound to a more "consumer hifi" sound.
There is more bass, but the bass is out of balance. The double bass sounds like it separates the ground notes and the sub resonances, and the overtones. Some of the energy is gone.
So first thing tomorrow is to change back C 4 to 12 uF and see what happens, and then the output caps back to 2000 uF.

[duak]

Peter, are you using phono as a source?  Perhaps the extended low frequency response is feeding back acoustically through the arm and turntable.  It all depends on the system and room acoustics but I've found that it seemingly increases bass while reducing bass definition.  It also tends to use up amplifier headroom by wasting power on something that doesn't belong.

[PeterSorvag]

Hi,

I have the 303 in the chain that I use for mastering, so there is no amplification like phono or other preamps between the 303 and the DAC.
The room acoustics are measured, and I am used to it. So the difference is before I changed the caps and after.
I wasn't shure what you meant: did you mean that turntables can increase the bass and deteriorate control over the bass, or did you mean that you have made the same observation regarding the caps ? :-)

[duak]

Hello Peter,

I was wondering if your system used a turntable and vinyl records as a signal source.  One problem with such a system is that low frequencies can couple acoustically from the speakers to the turntable.  Extending the low frequency response can make this worse because the system can resonate at certain frequencies.  Since you are using a DAC, there shouldn't be any acoustic feedback.

One thing that comes to mind is that the larger output capacitors (C1L) may have a different and probably lower ESR (series resistance) than the originals.  If you look at the schematic you will see they are outside of the feedback loop and directly affect the output impedance and the damping factor DF.  My anecdotal experience is that lower output impedance (higher DF) tends to reduce speaker coloration and what some people may call warmth or even musicallity.  A quick experiment would be to place a 0.2 to 1 ohm resistor in series with each loudspeaker and then see if there is a difference in sound.

Looking at the specs on the 303, I see that it is capable of driving 4 to 16 ohm speakers, however, it can only deliver its rated power at 8 ohms.  4 ohm speakers are more difficult to drive and require more current than do higher impedance speakers.  Note that a speaker's impedance is nominal and that at some frequencies it can be even lower. To get back to my original thought, is there a possibiliy of any subaudible, low frequency signals getting in to your signal chain and perhaps causing the 303 to distort?

Cheers,

[PeterSorvag]

Hi Duak,

no the source is digital, and the change in sound is from changing the caps. I decided to try the 4700 uF, as recommended by some upgrade sellers, and I altered the C 101 to 22 uF instead of 12.

A change in sound was expected, because I changed old caps for new ones, but I did not expect a worse sound than I had with the old caps :-)

I will change back to the correct values and let you know :-)

All the best
Peter

[PeterSorvag]

Hi,

I changed back to the original values, except for the output capacitors, where I went down to 1800 uF in parallel with a 100 uF.

It's beginning to sound good. The lack of energy that I noticed with the "upgrade" recommendations is gone, and it has a full sound and a good response for room acoustics.

Only one thing is disappointing: I noticed a edgy character in sound in the treble, and when I checked with a magnet, both Panasonic FC series and Nichicon audio grade capacitors have magnetic legs.
That is something I would never expect in components for audio applications.

Elna Silmic is non-magnetic.

I havent checked any other makes for this phenomenon, but maybe you guys have recommendations for good non magnetic electrolythic capacitors ? :-)

All the best
Peter

[floobydust]

You won't find people here sympathetic to "magnetic distortion" due to components made of steel, having extra inductance that is non-linear. It seems to show up only on high current paths outside the main feedback loop, such as loudspeaker output connectors or protection relay. I doubt you could hear it at all, as the 303 is not ultra low distortion at  the 0.001% THD floor where it shows up. Many parts have changed from copper to steel leads over the years.

The original input capacitor C100 0.68uF is tantalum which is not ideal. Best is a polypropylene film capacitor there, you can hear that difference instantly. Like Wima MKP4 0.68uF or 1uF 100VDC with 15mm lead spacing, if that fits.

[PeterSorvag]

You won't find people here sympathetic to "magnetic distortion" due to components made of steel, having extra inductance that is non-linear. It seems to show up only on high current paths outside the main feedback loop, such as loudspeaker output connectors or protection relay. I doubt you could hear it at all, as the 303 is not ultra low distortion at  the 0.001% THD floor where it shows up. Many parts have changed from copper to steel leads over the years.

The original input capacitor C100 0.68uF is tantalum which is not ideal. Best is a polypropylene film capacitor there, you can hear that difference instantly. Like Wima MKP4 0.68uF or 1uF 100VDC with 15mm lead spacing, if that fits.

Thank you for this reply,
I will test the polypropylene for the 0.68 uF cap.

As I understood it, the 303 is built on the "positive " side, and hysteresis distortion appears when the signal is going through zero. Same thing with vacuum tubes.
I was just surprised that capacitors that are supposted to "audio grade" contains magnetic material....