One thing you could try is that the power supply to the output stage at slightly higher voltage than the VAS stage. This causes the VAS to clip first. I found this out by accident. I tried the recomended VAS higher PSU and did not like the sound, the idea of that is maximum wattage. I simply reversed the thinking. Easy to do. A RC filter that throws away 3 V. If the VAS and input uses 15 mA for example that would be 200R. 2200 uF 63 V seems the ideal cap, no obvious time constant problem if so. Both - ve and + ve would need this. This also helps keep ripple out of the VAS. If this works well you might build a better PSU for the output stage alone. The latter needs measuring to be sure it is behaving. As the grounding is a bit suspect it probably will be difficult to do the latter.
On the grounding. You might reinforce the PCB ground tracks with 1.5 mm solid core. Just bend it point to point and solder at any component. This reduces the resitance between these points. Do it one track at a time to be sure nothing sounds worse. Again measure if possible with a scope. The ground between PSU caps should be prime, 2 pieces of 1.5 mm should fit well.
One thing people would not exspect is to reduce negative feedback can help voice quality. You will be very limited in how much you can try. If it were me I would double the gain. If the lower feedback arm is 1K add another 1K to make it 500 R. The volume control will not be as loud as you might think when changed. Bare in mind it will be at full volume earlier. The distortion will say when you must stop. What you might find is it will sound warmer. This works well with all feedback amps. When op amps gain of 1 or > 10 often is optimum. 2 to 9 less so. This is possibly because most designers must ensure gain of 1 is safe. Some op amps have compenstation pins. If so any gain can be made to work. Power amps also as most are big op-amps in layout. When a power amp the compensation is called Cdom ( most ). Very unwise to play with it if unsure. By doubling the gain the demands on the feedback loop are halved. In a technical sense distortion will double. Usually the distortion could rise by a factor of ten and still be OK on modern amps. The distortion that can be reduced was called TID. Some say it doesn't really exist. That is becuase they do not use music as a test. This gain doubling is 20 minutres to do and can work wonders. Some amplifers are much better as designed. That is rare. Yamaha might be one. I don't like their sound but would accept it is a type of perfection. They have no obvious faults. I only speak of the domestic designs. I dare say doubling the gain on a Yamaha would not change it for the better. Crossover distortion seems not to suffer by doing this. That would be the biggest doubt. I suspect other things prevent the correction which are actually helped by relaxing the loop. This is seen as crossover distortion away from the centre line ( 50 kHz + ). This is slew limiting and is not helped by feedback falling apart.
To be clear I love negative feedback and as much as possible, Problem is most amps are not really ideal for it's use. Too many stages and too slow. This seems to be caused by dogma. The Hypex amp is one I like very much. By curing it's class D problems it cures all the traditional ones. Simple reason is he threw much of the rule book away. The amp is very simple and it has to be to stand any chance of working. It has the dead band problem to solve. Ironically that solves similar traditional amp problems. I wish Bruno would make a class AB based on the same thinking. I suspect it would be better ?
You might raise the gain further. If so pot down the volume control to suit. To me that is a step too far for most as it is no longer is a very simple change. If someone can not find the feedback loop I also would say don't try. DC offset will double. < 50 mA when warmed up is OK. A Rotel RA 931 was 77 mA both sides. 100 mV wouldn't worry me. hiss will increase a bit. Some put their ears to the speaker and say not good enough. The question is would they hear it when real music is playing when 2 metres away? The signal to noise of a power amp might be - 100 dB. FM radio - 70 db on a good day. Seldom would anyone notice FM when so. Even the best microphones struggle to do better.
On the grounding. You might reinforce the PCB ground tracks with 1.5 mm solid core. Just bend it point to point and solder at any component. This reduces the resitance between these points. Do it one track at a time to be sure nothing sounds worse. Again measure if possible with a scope. The ground between PSU caps should be prime, 2 pieces of 1.5 mm should fit well.
One thing people would not exspect is to reduce negative feedback can help voice quality. You will be very limited in how much you can try. If it were me I would double the gain. If the lower feedback arm is 1K add another 1K to make it 500 R. The volume control will not be as loud as you might think when changed. Bare in mind it will be at full volume earlier. The distortion will say when you must stop. What you might find is it will sound warmer. This works well with all feedback amps. When op amps gain of 1 or > 10 often is optimum. 2 to 9 less so. This is possibly because most designers must ensure gain of 1 is safe. Some op amps have compenstation pins. If so any gain can be made to work. Power amps also as most are big op-amps in layout. When a power amp the compensation is called Cdom ( most ). Very unwise to play with it if unsure. By doubling the gain the demands on the feedback loop are halved. In a technical sense distortion will double. Usually the distortion could rise by a factor of ten and still be OK on modern amps. The distortion that can be reduced was called TID. Some say it doesn't really exist. That is becuase they do not use music as a test. This gain doubling is 20 minutres to do and can work wonders. Some amplifers are much better as designed. That is rare. Yamaha might be one. I don't like their sound but would accept it is a type of perfection. They have no obvious faults. I only speak of the domestic designs. I dare say doubling the gain on a Yamaha would not change it for the better. Crossover distortion seems not to suffer by doing this. That would be the biggest doubt. I suspect other things prevent the correction which are actually helped by relaxing the loop. This is seen as crossover distortion away from the centre line ( 50 kHz + ). This is slew limiting and is not helped by feedback falling apart.
To be clear I love negative feedback and as much as possible, Problem is most amps are not really ideal for it's use. Too many stages and too slow. This seems to be caused by dogma. The Hypex amp is one I like very much. By curing it's class D problems it cures all the traditional ones. Simple reason is he threw much of the rule book away. The amp is very simple and it has to be to stand any chance of working. It has the dead band problem to solve. Ironically that solves similar traditional amp problems. I wish Bruno would make a class AB based on the same thinking. I suspect it would be better ?
You might raise the gain further. If so pot down the volume control to suit. To me that is a step too far for most as it is no longer is a very simple change. If someone can not find the feedback loop I also would say don't try. DC offset will double. < 50 mA when warmed up is OK. A Rotel RA 931 was 77 mA both sides. 100 mV wouldn't worry me. hiss will increase a bit. Some put their ears to the speaker and say not good enough. The question is would they hear it when real music is playing when 2 metres away? The signal to noise of a power amp might be - 100 dB. FM radio - 70 db on a good day. Seldom would anyone notice FM when so. Even the best microphones struggle to do better.
Hi sorry for the late reply but thanks a lot for the very kind and valuable advice.
i start from the idea just to improve the PS alone ... without messing to much with the circuit.
Usually the PS is where cost saving are done
I have no idea of the quality of the mains transformer used, of the PS caps.
I am always looking at the Mission Cyrus TWO as a reference.
Mission proposed a beefed-up PS that enhance the performance a lot indeed.
I would buy a new transformer, a 500VA thing. Very big caps that i would like very much to place in the amp at the place of the transformer.
Like 2x33.000uF caps ... a big energy reservoir indeed.
I still think that this is potentially a very good amp ... i get also some depth of soundstage
for me the best evidence of quality.
Thanks a lot for the very interesting advice. I think i could do this one quite easily.
wow ... this is so much beyond my abilities ... Thanks a lot anyway.
I would not mess with the actual circuit and try to keep it as original as possible.
I start assuming that the distortion is already low. Much lower than any possible tube amps for instance.
I would try only to decrease PS impedance and increase its capacitance.
Unfortunately i have absolutely no idea of the quality of the transformer used
To end all my rambling i think that the most sane idea would be to select a kit instead.
Every time i tried to modify something from stock i had bad experience ending most of the time with the destruction of the unit.
Modding is not an option here i am afraid.
Still the PS caps ... that should not do any harm ... i hope.
Thanks a lot again. Kind regards, gino
i start from the idea just to improve the PS alone ... without messing to much with the circuit.
Usually the PS is where cost saving are done
I have no idea of the quality of the mains transformer used, of the PS caps.
I am always looking at the Mission Cyrus TWO as a reference.
Mission proposed a beefed-up PS that enhance the performance a lot indeed.
I would buy a new transformer, a 500VA thing. Very big caps that i would like very much to place in the amp at the place of the transformer.
Like 2x33.000uF caps ... a big energy reservoir indeed.
I still think that this is potentially a very good amp ... i get also some depth of soundstage
for me the best evidence of quality.Thanks a lot for the very interesting advice. I think i could do this one quite easily.
wow ... this is so much beyond my abilities ... Thanks a lot anyway.
I would not mess with the actual circuit and try to keep it as original as possible.
I start assuming that the distortion is already low. Much lower than any possible tube amps for instance.
I would try only to decrease PS impedance and increase its capacitance.
Unfortunately i have absolutely no idea of the quality of the transformer used
To end all my rambling i think that the most sane idea would be to select a kit instead.
Every time i tried to modify something from stock i had bad experience ending most of the time with the destruction of the unit.
Modding is not an option here i am afraid.
Still the PS caps ... that should not do any harm ... i hope.
Thanks a lot again. Kind regards, gino
I think the 500 VA and 1.5 mm solid core should do all you need. Tiger in Peterborough UK did me one recently in the style of Naim Audio. You don't need that, just to say they were happy to say yes. I had some extra windings for a Raspberry Pi and phono stage. Price, quality and time to do were excellent. As 500 VA usually has a better regulation factor allow for that in when ordering. Don't assume larger capcacitor values mean better sound. When 500 VA you should be able to go for larger capacitors. Mission used DNM capacitors. They have slits in the foil to improve HF ability. A 10 000 uF DNM might considerably outperform " upgrades ".
One interesting thing about the Mission. It has an ideal volume ( about 60% ). Above or below that it changes tone colour for want of a better way to say it. I have looked at the circuit and have no idea why ( feedback and Cdom ? ). Rather a nice circuit. As luck would have it the sweet spot suits most people. Like colour balance on older TV this is noticed when using something better. For this reason I would as you say limit yourself to the easy stuff. If you sell the amp put the old stuff back. The transformer would be ideal for something new and DIY. On older amps the smaller capacitors seemed to produce better midband. These might be only 4700 uF. Upgrading to 10 000 uF often produced a bloated bass. Had the less obvious upgrade been done either with original or 10 000 uF plus big transformer then the bloated bass is replaced by a cleaner/deeper sound. Musically this is easy to explain. Just like a church organ deep bass gives the overall tonal colour. Also the midband seems deeper when right. So big capacitors can mean a change for the worse. The organist also has to time the bass!!! If not the midband is less deep and less sweet.
Douglas Self's book on audio power amp design is worth having. I don't aggree with all he says. I don't think any of it will send you down a badly wrong route.
One interesting thing about the Mission. It has an ideal volume ( about 60% ). Above or below that it changes tone colour for want of a better way to say it. I have looked at the circuit and have no idea why ( feedback and Cdom ? ). Rather a nice circuit. As luck would have it the sweet spot suits most people. Like colour balance on older TV this is noticed when using something better. For this reason I would as you say limit yourself to the easy stuff. If you sell the amp put the old stuff back. The transformer would be ideal for something new and DIY. On older amps the smaller capacitors seemed to produce better midband. These might be only 4700 uF. Upgrading to 10 000 uF often produced a bloated bass. Had the less obvious upgrade been done either with original or 10 000 uF plus big transformer then the bloated bass is replaced by a cleaner/deeper sound. Musically this is easy to explain. Just like a church organ deep bass gives the overall tonal colour. Also the midband seems deeper when right. So big capacitors can mean a change for the worse. The organist also has to time the bass!!! If not the midband is less deep and less sweet.
Douglas Self's book on audio power amp design is worth having. I don't aggree with all he says. I don't think any of it will send you down a badly wrong route.
Nige, did you take a good look at the Cambridge 340A internal pic and pcb layout ?.
It really is not conducive to modding, especially by a novice.
About the most that could be done is to change a handful of coupling caps to bipolar type, change NFB caps to BP and add small value shunt, and change chipamp power rail decoupler caps.
Gino, you could do the above, but really I think you should consider building a 'proper' amplifier.
modulus-86-build-thread.html would be a good starting point.
Tom Christensen has designed a very good pcb, and the performance is pretty much SOTA.
Economical and easy to build, this design should suit you admirably.
Dan.
It really is not conducive to modding, especially by a novice.
About the most that could be done is to change a handful of coupling caps to bipolar type, change NFB caps to BP and add small value shunt, and change chipamp power rail decoupler caps.
Gino, you could do the above, but really I think you should consider building a 'proper' amplifier.
modulus-86-build-thread.html would be a good starting point.
Tom Christensen has designed a very good pcb, and the performance is pretty much SOTA.
Economical and easy to build, this design should suit you admirably.
Dan.
Hi and thanks a lot for the very kind and valuable advice
I am only worried about the number of secondaries needed ... but i could go with even two transformers, one small for the op-amps supply and one big for the amp stages. If the space inside allowas of course.
What i wanted to get with more uF in the PS is a bigger feeling of ease in the sound delivery, that should result in lower noise and more spatial sound
I did some very brutal experiments with very big caps on an old amp and the result was quite impressive ... i think they were 2x60mF blue caps from Mallory ... high grade ones. I liked what they were doing to sound.
When i switched off the amp the sound went on for 10 seconds at least
I used the Mission PSX as an example of external power supply
I still like the caps placed as close as possible to the output devices
But all the reviews are positive about the effect of this psx on overall sound.
i am doubtful about this bloated bass ... i think that this has to do with amp damping factor.
Maybe there is an optimal value for uF in the PS.
But i would stay more on 20mF than 10mF x 2 anyway of good quality.
If you look at the datasheet you see that a 20mF cap has better specs than a 10mF capof same series and WV. Every parameter is better, price aside
thanks a lot again for the very useful directions.
Despite what everyone tell me this is not a bad amp. The sound when not pushed is quite transparent and dimensional.
But i understand that modding is much more difficult than buy a complete kit and start soldering it. In a commercial unit there are many constraints, space for instance.
Kindest regards, gino
I did look at the Cambridge. I think over laying the PCB grounds might help. Original path made lower in reistance. That way nothing should change for the worse. New transformer also. Other than that it needs a complete redesign I fear.
The bloated thing. It is what people sometimes call timing. It can sound exactly like damping factor. I wonder if damping factor might be we can hear the power suppy? Some amps with " poor " damping factor do not boom or sound bloated.
The bloated thing. It is what people sometimes call timing. It can sound exactly like damping factor. I wonder if damping factor might be we can hear the power suppy? Some amps with " poor " damping factor do not boom or sound bloated.
The inductance of "normal" speaker wire, is so low to begin with, relative to system inductance from other sources such as the voice coils, passive crossover inductors, etc., that even if it would have otherwise been relevant (it would not be) it is so far far below possibly being significant.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.