Quad 303, possible improvement

[spendorman]

I have a fairly good selection of amplifiers, but my favourites are Radford STA25 III and Quad II power amplifiers.

None of my transistor amps seem to give quite the same satisfaction. I am fond of the Quad 303, and have four of them. They are all set up pretty well to spec.

I thought that I might try an experiment, increase the output stage bias current to about double.

I have done this, 20 mV across the pair of resistors R124/125, instead of the usual 10 mV.

Heatsinks do not heat up excessively and I think that I am a little step nearer to the sound of the Radford STA25 / Quad II.

[Rare Bird]

Try using MJ15003/MJ15004 in fully complementary output as Net do..

[spendorman]

Try using MJ15003/MJ15004 in fully complementary output as Net do..

Thanks, thought about that, but decided may try that at some other time. Also thought about just trying MJ15003 and keeping the circuit as original.

[Rare Bird]

O Mr Spendorman your PM box is full ;)

[spendorman]

O Mr Spendorman your PM box is full ;)

Fixed now, thanks

[DSJR]

Andr'e is right and the option once output devices are changed to remove or significantly relax the current limiting (4A to at least 8A) would help with more modern low impedance speakers. Possibly the fully-complimentary conversion which can be done with current transistors.

Please check out the Dada website as well :)

Of course, if Quad thought that the original tweak was a good thing, don't you think they'd have done ot at the time?

Others' work on the equally antiquated Naim amp design shows that there is an optimum bias current level and increasing it makes no beneficial difference - see the Acoustica site for links regarding this..

[spendorman]

Andr'e is right and the option once output devices are changed to remove or significantly relax the current limiting (4A to at least 8A) would help with more modern low impedance speakers. Possibly the fully-complimentary conversion which can be done with current transistors.

Please check out the Dada website as well :)

Of course, if Quad thought that the original tweak was a good thing, don't you think they'd have done ot at the time?

Others' work on the equally antiquated Naim amp design shows that there is an optimum bias current level and increasing it makes no beneficial difference - see the Acoustica site for links regarding this..

303 as I guess you know uses NPN devices for the final output. One cannot make this fully complimentary using the "current" transistors, one needs an NPN and PNP pair.

I am a great one for leaving things standard, but this bias mod has definitely improved the amp.

It might just be a one off with this amp that is improved by increasing the bias, I have three more 303's to try the same mod on.

[DSJR]

That's fine, but i'd like to know why this was never done by quad, that's all. Maybe there's a perfectly logical and genuine reason why not :) But don't ever think that they'd not have done it because they hadn't thought of it at the time ;)

[Reid Malenfant]

There wasn't any complementary devices about when the 303 was developed & going over to a fully complemetary output stage would involve modifying circuit boards. Why fix what aint broke :eyebrows:

Obviously a competant modder could easily figure it out, but you'd need an oscilloscope & all sorts of loads to test it on for stability issues once modified.

:doh: Looks like i was answering the wrong question :eyebrows:

[spendorman]

That's fine, but i'd like to know why this was never done by quad, that's all. Maybe there's a perfectly logical and genuine reason why not :) But don't ever think that they'd not have done it because they hadn't thought of it at the time ;)

Well, I don't know, but in my opinion Quad 303's (in standard form) are not that good in retaining their output stage bias setting. Over the years I have serviced quite a few.

Increasing the bias a bit over standard, I suppose makes it less likely that over a period of time, the amp will likely have at least the specified bias current.

[spendorman]

There wasn't any complementary devices about when the 303 was developed & going over to a fully complemetary output stage would involve modifying circuit boards. Why fix what aint broke :eyebrows:

Obviously a competant modder could easily figure it out, but you'd need an oscilloscope & all sorts of loads to test it on for stability issues once modified.

I believe Net Audio and others have figured it out, I know a guy who has the Net Audio mod. He has had some problems though.

Don't forget that there are complimentary power transistors in the 303, the drivers are, and they are quite big. However, the PNP ones are more fragile than the NPN ones.

Anyway, I am tempted to increase the output stage bias on my other three 303's and a friend has requested that I do his one.

[Reid Malenfant]

I don't see why you can't take it to 100mA bias if you wanted to, that would give some class A power at normal listening levels with only 6W of dissipation per channel ;) Higher isn't out of the question either as 12W is naff all even on a heatsink the size that the 303 has :)

[spendorman]

I don't see why you can't take it to 100mA bias if you wanted to, that would give some class A power at normal listening levels with only 6W of dissipation ;) Higher isn't out of the question either as 6W is naff all even on a heatsink the size that the 303 has :)

Absolutely, but I thought try a little first!

Perhaps someone else would experiment and report findings.

[spendorman]

I don't see why you can't take it to 100mA bias if you wanted to, that would give some class A power at normal listening levels with only 6W of dissipation ;) Higher isn't out of the question either as 6W is naff all even on a heatsink the size that the 303 has :)

Bear in mind that the heatsink is also used for the series power regulator. On a standard 303 this can heat up the heatsink more that anything else (amp at idle or low volume).

This is another interesting point. Depending on which taps Quad used on the main transformer secondary (and they did vary this) more or less is voltage is dropped across the regulator.

I discovered this because at idle one 303 ran warmer than another, even though both had their output bias set to spec. This difference of temp was due to the different tappings used.

[Reid Malenfant]

If i find the time i might have a crack at it myself as i have a 303 & 33.

Just out of interest what is the standard quiescent current? A guess would be about 20 - 25mA :scratch:

[spendorman]

If i find the time i might have a crack at it myself as i have a 303 & 33.

Just out of interest what is the standard quiescent current? A guess would be about 20 - 25mA :scratch:

Have to look back, but I think emitter resistors are 0.3 Ohm, I measure across both to double the voltage and voltage is what I mentioned earlier in the thread. My memory is bad!

[spendorman]

I have a fairly good selection of amplifiers, but my favourites are Radford STA25 III and Quad II power amplifiers.

None of my transistor amps seem to give quite the same satisfaction. I am fond of the Quad 303, and have four of them. They are all set up pretty well to spec.

I thought that I might try an experiment, increase the output stage bias current to about double.

I have done this, 20 mV across the pair of resistors R124/125, instead of the usual 10 mV.

Heatsinks do not heat up excessively and I think that I am a little step nearer to the sound of the Radford STA25 / Quad II.

Here it is:


So it looks like about 16.7 mA as standard

33.4 mA as modded.

Just will check diag to see if I remembered the emitter resistor value correctly.

[Reid Malenfant]

Heck that's a little low, i guess they could get away with it though due to the output triple. 0.016A :scratch:

[spendorman]

Heck that's a little low, i guess they could get away with it though due to the output triple. 0.016A :scratch:

Just checked, emitter resistors (one is actually a collector resistor) are 0.3 Ohms.

[DSJR]

have you seen this? From the Revisionspot site...

"The next myth: high bias current

In the Quad service manual it is stated that the DC bias current of the power transistors must be between 10mA and 5mA. A popular “mod” is increasing the current to 35mA and in some cases even higher. Well, in real life the distortion is only 0.05% at a bias current of 0.5mA. At 5mA it is 0.01% and will not get lower at higher bias currents. Also the spectrum analyzer will not give a better (cleaner) picture at high levels of bias current. So Quad was (is) right! For practical reasons, stability of the circuit, 10mA is a good value for the bias current."

[Reid Malenfant]

Hi Dave, all well & good ;) However with some power avaialble in class A distortion measurements simply don't tell the whole story :) Once again, it's not about how much distortion is present, but the type of distortion become much more significant :cool:

In class A there can be no crossover distortion, so higher harmonics are out of the window & lower harmonics predominate which tend to be pretty innocuous.

[DSJR]

I accept that. Although I haven't measured the bias on my 303, the sound is like velvet with no aggro at all. Any grit or grain may be due to something else?????????

As long as no damage is caused though :)

[Reid Malenfant]

I was always very happy listening to my dads 303/33 combo, it blew away every music centre that anyone else owned back in the day :lol:

No idea about this one either tbh, but a little calculation reveals about 0.5W of class A or a tad more might be possible per channel with moderate heatsink temperatures. It might not sound like a lot, but if you don't listen to things loud it may well be worth the fettling.

I wouldn't expect distortion to increase too much once it enters class B either due to the output triple & it's huge current gain.

[Rare Bird]

Dave have you tried the '303' with say a 10K Passive pre?

[spendorman]

have you seen this? From the Revisionspot site...

"The next myth: high bias current

In the Quad service manual it is stated that the DC bias current of the power transistors must be between 10mA and 5mA. A popular “mod” is increasing the current to 35mA and in some cases even higher. Well, in real life the distortion is only 0.05% at a bias current of 0.5mA. At 5mA it is 0.01% and will not get lower at higher bias currents. Also the spectrum analyzer will not give a better (cleaner) picture at high levels of bias current. So Quad was (is) right! For practical reasons, stability of the circuit, 10mA is a good value for the bias current."

Very interesting, I still reckon that my amp with the increased bias sounds better. It could be that mine has some problems and that he increased bias overcomes. I doubt it though, it is in pretty good condition.

Tomorrow, my friend was supposed to be bringing me his 303 (which I set up as standard about a year ago) for the increased bias setting, but that is going to be delayed a while. He will have a listen to mine though.

I wonder why the the increased bias mod was popular.

There is no doubt in my mind that a 303 with bias that's low, can sound rough. It's mainly in the treble.

I purchased my first 303 in the early 70's. It was second hand and in good condition. I did not like the sound, gritty in the treble when compared to my Quad II's.

The 303 sat on the shelf for some years until I decided to set it up to spec. It was not massively out. But the improvement in sound quality was considerable.

This amp is one of the very early 303's (not the one I have done the mod on) and is still working fine, all original components. None of the electrolytics show any sign of leakage.

The other three 303's are much newer, one is a late model with the IEC power connector, strangely, all the newer 303's needed their big power supply and output capacitors replacing.

Well, perhaps it's not so strange that the newer amps need cap replacement, more like it's strange that the oldest one is still fine on it's original capacitors.

[spendorman]

Hi Dave, all well & good ;) However with some power avaialble in class A distortion measurements simply don't tell the whole story :) Once again, it's not about how much distortion is present, but the type of distortion become much more significant :cool:

In class A there can be no crossover distortion, so higher harmonics are out of the window & lower harmonics predominate which tend to be pretty innocuous.

Spot on, in my view

[spendorman]

Dave have you tried the '303' with say a 10K Passive pre?

I have, worked fine.

[spendorman]

Hi Dave, all well & good ;) However with some power avaialble in class A distortion measurements simply don't tell the whole story :) Once again, it's not about how much distortion is present, but the type of distortion become much more significant :cool:

In class A there can be no crossover distortion, so higher harmonics are out of the window & lower harmonics predominate which tend to be pretty innocuous.

Many years ago, I was given a Rogers Cadet III with a blown output transformer. I could not find a Rogers one, but G W Smith sold one that they said would be an equivalent. I cant remember exactly, but it was not straightforward to fit it.

Think I had to change the feedback and something else. However on a cheap test speaker it worked. I had a scope, various signal generators and did some basic tests, all seemed OK.

So, I connected the amp to the Hi Fi speakers, and was shocked that the repaired channel sounded so bad. The measurement (although basic) did not tell anyway near as much as my ears could.

Years later, that amp is still like that. Sowter can do a transformer, but it's costly.

[Rare Bird]

I have, worked fine. I have aswell & it does work fine your right, for me it gives a more open sound it's as if the '303' has been set free to exhibit what a great performance it's capable of..don't get me wrong i totally adore the looks of the '33' but not it's sound regardless of what mods you perform on it.

[spendorman]

I have aswell & it does work fine your right, for me it gives a more open sound it's as if the '303' has been set free to exhibit what a great performance it's capable of..don't get me wrong i totally adore the looks of the '33' but not it's sound regardless of what mods you perform on it.

It was when I started running the 303 straight off my fairly good soundcard that I found how good it was.

Using the NAD 3130 as a pre for it does not seem to noticeably degrade the sound. The NAD as a pe amp seems to better than both my 33 and 44, my 44 is an early one though.

[DSJR]

My 303 is one of the last, according to hifi dave who sold hundreds to a large chain of record shops and has, of course, been sensitively "updated" to high heavan by the previous owner. left to idle with no speaker load, it gets gently warm on the heatsinks, so may be set high as it is for all I know.

An old-time work colleague is the son of a once famous pop/country star who had a passion for yodelling :eek: - looked like his Dad too.. When Dad emigrated back to Oz, he left behind his Quad 33/303 and Tannoy Lancasters (I think they were). The speakers had been exposed to high humidity and the voice coil gaps had corrosion in them, needing a complete and expensive rebuild. These were sold twenty years ago for £500 but the amp was retained. It sounded truly awful and exactly as you described above - gritty, noisy and generally unpleasant. Our engineer re-capped a lot of it and re-set up the amp, but although loads better after, it never really sang like I've heard other 33/303's do.

What got me back into 33/303's was the trading in of a lovely set, totally original and sweet as a nut with air and space. Only a small scale sound and frequency-field showed the age of the design as standard. Great that both items can be sonically opened out and allowed to fly :)

Quad themselves shot us in the feet by recommending QED style passive volume controls for their older power amps - :steam: I got very cross with one of their engineers once as I lost a sale of a 34 because of this.

[Macca]

Quad themselves shot us in the feet by recommending QED style passive volume controls for their older power amps - :steam: I got very cross with one of their engineers once as I lost a sale of a 34 because of this.

It did work well though Dave. A pal of mine had a QEDpassive to 303 set up back in the early 90s and it sounded lovely - open and sweet with good bass

[DSJR]

I'm very sure it did :) Mine's up here in the "office" awaiting some small vintage speakers (I really should get the N5's out, but I have an itch to scratch first)..

[spendorman]

My 303 is one of the last, according to hifi dave who sold hundreds to a large chain of record shops and has, of course, been sensitively "updated" to high heavan by the previous owner. left to idle with no speaker load, it gets gently warm on the heatsinks, so may be set high as it is for all I know.

An old-time work colleague is the son of a once famous pop/country star who had a passion for yodelling :eek: - looked like his Dad too.. When Dad emigrated back to Oz, he left behind his Quad 33/303 and Tannoy Lancasters (I think they were). The speakers had been exposed to high humidity and the voice coil gaps had corrosion in them, needing a complete and expensive rebuild. These were sold twenty years ago for £500 but the amp was retained. It sounded truly awful and exactly as you described above - gritty, noisy and generally unpleasant. Our engineer re-capped a lot of it and re-set up the amp, but although loads better after, it never really sang like I've heard other 33/303's do.

What got me back into 33/303's was the trading in of a lovely set, totally original and sweet as a nut with air and space. Only a small scale sound and frequency-field showed the age of the design as standard. Great that both items can be sonically opened out and allowed to fly :)

Quad themselves shot us in the feet by recommending QED style passive volume controls for their older power amps - :steam: I got very cross with one of their engineers once as I lost a sale of a 34 because of this.


Just because the heatsink warm up a fair bit at idle, does not mean that that bias is set high, as mentioned earlier, the series regulator is mounted on the heatsink. The heat will depend on which secondary tappings are used and how much the regulator is dropping.