How much capacitance can I add (upgrade) to the amplifier's PSU?

[drussell]

I will check up those rectifiers. But they are no longer in production.. Don't know what will be most like the RS603

Are they actually RS603s?  I think they're still produced.  Regardless, the datasheet says:

Peak Forward Surge Current, 8.3 ms single half sine-wave superimposed on rated load (JEDEC method) = 200 A

There is a graph on the Diodes Inc. datasheet that shows a curve for "MAXIMUM NON-REPETITIVE FORWARD SURGE CURRENT" for 1 to 100 cycles of 60 Hz AC that still shows about 75A capability at 10 cycles, so I'm pretty sure he's gonna be fine, even on some kind of transient where the soft start failed to function on a restart, etc.

[Simon]

Indeed if the data on the diodes allows for such surges then it will be fine (while making no positive difference), no doubt the soft start resistors were calculated so that the surge current was not exceeded on start in the dead short situation of discharged capacitors and the charging time of the capacitors taken into account.

[G0MJW]

The simple answer is none - if it really is getting on for 20 years old I would replace the capacitors rather than add any more capacitance. Chances are more modern capacitors will have more capacitance for the same size. I doubt the value is important within 50%. Older capacitors can degrade - depending on the sort of life they have had. They may be and probably are just fine, but starting from a new replacement gives a good baseline. Its very important to get the voltage rating right. You also need to pay attention to ripple current ratings and ESR. ESR will have the greatest impact on the audio, ripple rating on the fire hazard.

Mike

[Simon]

The voltage should be 2x the RMS voltage, equipment from the 60's like the quad 303 had capacitors rated too low and were infamous for failure (and then idiots on ebay would advertise compatible replacements also rated too low). Depending on the construction of the amp/psu I'd use the space savings of the modern capacitor to increase the voltage if necessary and the temperature rating.

[drussell]

The simple answer is none - if it really is getting on for 20 years old I would replace the capacitors rather than add any more capacitance.

He is replacing them.  He just wanted to know if he can replace the 10,000 uF units with 15,000 uF units and have the amplifier function correctly.

The choice of filter capacitance in that kind of application is somewhat arbitrary anyway and the manufacturer might well have almost chosen 15,000 uF originally anyway.  I'm sure it will be just fine with the higher capacitance capacitors capably clarifying the current.   

The voltage should be 2x the RMS voltage, equipment from the 60's like the quad 303 had capacitors rated too low and were infamous for failure (and then idiots on ebay would advertise compatible replacements also rated too low). Depending on the construction of the amp/psu I'd use the space savings of the modern capacitor to increase the voltage if necessary and the temperature rating.

Indeed.  When replacing bulk filter capacitors in an application like this it is wise to select for the best temperature rating, life specification, robust physical construction and ripple rating while also ensuring adequate operating voltage margin in addition to the capacitance value when choosing a replacement.

These are all trade-offs with physical size and is the reason there are a zillion different electrolytic capacitor series / types.  I would think good quality parts with all those good specs, even at 15,000 uF should easily be available to fit the bill.

[Simon]

My take on most parts is that if you multiply all of the specs together or in some cases the inverses for some you will come up with a constant, keeping that constant you could alter another variable and see how it affects the others or you part grows in size, increases in cost ect. This is just an idea i have and not scientific, but if you look through a few datasheets of mosfets you start to get my drift, want stupidly high current handling capability ? sure..... but your input capacitance will go through the roof..... this has other knock-on effects that many beginners will overlook and think more current capacity is better when in fact it can be disastrous in a replacement situation.

If the capcitor technology has matured to the point where you can get a higher voltage rating (if required) and a higher temp rating AND a higher capacitance without having high ESR then sure put the capacitance up a bit. Obviously if we are being scientific increasing the capacitance by 50% means that the new value will still be -20% +50% so although we are talking 10'000 uF nominal, that was factored in for a maximum of 15'000uF, now if you fit a nominal 15'000uF you could have a 22'500uF at +50%, this is significant but may still work.

[kris2ff3r]

The replacement capacitor (15000µF) is of the same brand as the original (Nippon Chemi-con). It's also the only one I can find that can be mounted on the PCB without any further modifications.

- SMH series
- Same lead spacing, same diameter, bit longer (of course) but not too much.
- Max operating temp of 85ºC
- Max 0.017 ohm ESR @ +20ºC,120Hz (80V, 100V Surge)

[Simon]

How does it compare to the old one on temperature, how good is the PSU cooling? Is the PSU 40V RMS?

Sent from my Moto G (4) using Tapatalk

[kris2ff3r]

How does it compare to the old one on temperature, how good is the PSU cooling? Is the PSU 40V RMS?

Sent from my Moto G (4) using Tapatalk

The old one is also specified to 85ºC.
There is no active cooling in the powersupply, the only heatsinks are in the amplifier section.

In standard configuration I haven't noticed any heat in the powersupply (without load). Hard to tell while in use...
I want to check with a Fluke IR cam, but that would require me to own a cam  ..
Not sure about the voltage, I'm out of town at the moment, but I BELIEVE the VCC is 63VDC (no load/no signal input)

[Simon]

A thermocouple poked into the cap area will do fine. Costs less than £10, it cameras have their pitfalls too.

Sent from my Moto G (4) using Tapatalk

[kris2ff3r]

A thermocouple poked into the cap area will do fine. Costs less than £10, it cameras have their pitfalls too.

Sent from my Moto G (4) using Tapatalk

Do you know about a good thermocouple?
The capacitance will also be 50% larger in the VDD supply. Same wiring, but using two sets of MUR120 in FWB, and two 2200µF (Now 3300µF).

I can't monitor the heating in the powersupply with the 10000µF's, cause they are already removed.
One of the caps were bulky, that's why this whole thing came up ''maybe use 15000µF, when they're already removed''

Before I came here with the question I actually believed that you could add a shitload of capacitance if you really wanted. It could never be too much to damage anything. Now I see different, but I remember another hifi manufacturer (also from Norway).
They made a A/B 70W (in 8 ) SS Amp.
The amplifier had a reservoir of 41600µF. They do sell the unit also as an UP edition. The UP edition is the exact same amplifier, but with a reservoir of 87600µF, and this is still a 70W amp. Still produced today with the standard reservoir, but they offer the upgrade.

(Not my statement) It gives noticeable sonic improvements. Makes the unit more stable with low impedance speakers. And in bridge mode it gives you more bass, and ''tonality''.
- What do you all think about these statements?

[Simon]

The statements are for audiofools. The signal goes in and the amplifier should faithfully amplify it. provided you meet the minimum required to not have mains hum it makes no difference.

Sent from my Moto G (4) using Tapatalk

[IanB]

The statements are for audiofools. The signal goes in and the amplifier should faithfully amplify it. provided you meet the minimum required to not have mains hum it makes no difference.

Not necessarily. A larger buffer can affect dynamics and transient performance. See the comment below from drussell:

Adding a bit more capacitance isn't going to make the amplifier more powerful.  It is still going to clip at the point set by the circuitry.  It will still output the same maximum 100 Watts.

It might potentially improve the transient response on large peaks a little bit, though, especially if you're driving a very low impedance load...  You might get a bit more solid bass "punch", for example.  Probably not much, though.  The manufacturer chose 20,000 uF for a reason, however, if you think it seems a bit anemic you can always try it and see if you like the result. 

It is virtually impossible for it will blow anything up but it will likely not make a noticeable difference and might actually make any power supply diode noise worse, as floobydust pointed out, but I don't think you'd hear that either with a change of 20,000 -> 30,000 uF.

[Simon]

I would have thought the amplifier would have a design slew rate etc and therefore the original capacitance be set to cope with any eventuality. Presumably the lack of sufficient capacitance will mean that for a large output the capacitor would be over discharged and the amplifier would start to clip and distort.

I can't really see how in a well designed amplifier the power supply capacitance would have any effect on sound "quality" providing the minimum is met. The frequencies we can hear are not exactly that high, if anything I'd be adding some faster caps if that sort of thing was a concern. My assumption is that if the capacitance is too small it won't hold the voltage up very well and so the output will clip and the mains hum will come through, once you are over that what more is there to gain ?

If we are talking about capacitors in series with a speaker on a single supply then yes i can see a bigger cap being better but most power amps have a split supply and don't need output capacitors.

[IanB]

The power supply reservoir capacitors provide the power supply voltage that the amplifier "sees". If you draw power from the reservoir faster than it can be replenished then the supply voltage drops, which will reduce the maximum instantaneous output power the amplifier can develop (crudely, output power = output voltage squared over output impedance). If the supply voltage drops, the output voltage drops and the output power sags. This can cause output transients to lose their accuracy. How bad this is in practice depends on what the amplifier designers thought was acceptable. Every amplifier design has constraints on what is expected of it.

[IanB]

And of course, this may only be noticed at full volume, by which time performance will probably be at its worst anyway.

[Simon]

And of course, this may only be noticed at full volume, by which time performance will probably be at its worst anyway.

Quite, I don't know what the slew rate of the average audio power amp is but the caps will be replenished at 100Hz, unless you demand a 0 to Vmax rise instantly you should not see a problem, at the end of the day if your are expecting to pulse the speakers like an idiot teenager because you think that is bass then you deserve to loose performance

[rdl]

The replacement capacitor (15000µF) is of the same brand as the original (Nippon Chemi-con). It's also the only one I can find that can be mounted on the PCB without any further modifications.

- SMH series
- Same lead spacing, same diameter, bit longer (of course) but not too much.
- Max operating temp of 85ºC
- Max 0.017 ohm ESR @ +20ºC,120Hz (80V, 100V Surge)

It isn't necessary but, all else being equal, I always replace 85C capacitors with 105C rated.

[Zero999]

I would have thought the amplifier would have a design slew rate etc and therefore the original capacitance be set to cope with any eventuality. Presumably the lack of sufficient capacitance will mean that for a large output the capacitor would be over discharged and the amplifier would start to clip and distort.

I can't really see how in a well designed amplifier the power supply capacitance would have any effect on sound "quality" providing the minimum is met. The frequencies we can hear are not exactly that high, if anything I'd be adding some faster caps if that sort of thing was a concern. My assumption is that if the capacitance is too small it won't hold the voltage up very well and so the output will clip and the mains hum will come through, once you are over that what more is there to gain ?

If we are talking about capacitors in series with a speaker on a single supply then yes i can see a bigger cap being better but most power amps have a split supply and don't need output capacitors.

The reservoir capacitors do more to boost low frequency response, than high frequencies. The larger capacitance can be useful, when the volume is high and it's playing bass heavy music, such as reggae, drum & bass, rap and funk, which I know you love. 

[Simon]

The reservoir capacitors do more to boost low frequency response, than high frequencies. The larger capacitance can be useful, when the volume is high and it's playing bass heavy music, such as reggae, drum & bass, rap and funk, which I know you love. 

Exactly, unnecessary music so unnecessary requirement 

[Simon]

It isn't necessary but, all else being equal, I always replace 85C capacitors with 105C rated.

Which is why i was asking what the temperature is in the capacitor compartment.

[kris2ff3r]

I will check up those rectifiers. But they are no longer in production.. Don't know what will be most like the RS603

Are they actually RS603s?

See attached

[Gyro]

The reservoir capacitors do more to boost low frequency response, than high frequencies. The larger capacitance can be useful, when the volume is high and it's playing bass heavy music, such as reggae, drum & bass, rap and funk, which I know you love. 

Exactly, unnecessary music so unnecessary requirement 

Actually I think you'll find that it's large orchestral works that typically have the lowest fundamentals and demand the most sustained bass handling. A lot of the stuff mentioned above actually relies on higher bass frequencies to give the 'in the chest' feeling when played at high volume through PA speakers.

[Simon]

oh all right But I say again, the amplifier will have a slew rate, after you have provided enough capacity to reach the output the amplifier is capable of more won't help. If a manufacturer is advertising an "upgrade" is makes you wonder why they didn't just make it right in the first place. Clearly the manufacturer is aiming at a certain market, capacitors are not that expensive in the grand scheme of things.

[Gyro]

I'm not going to disagree with you there. There's really no point in going past a certain value - you might just as well go for regulated supply rails, as used by John Linsley Hood in his ultimate Mosfet amp designs.


[Diversion]

Of course you want real bass you need to go to the Proms on a night when they fire up the 64' stop (8.2Hz) on the Royal Albert Hall organ. Or even better, if you're man enough, there's a Cathedral in Germany with a 128' pipe (4.1Hz) - probably enough to do interesting thing to your bowels! 



[/Diversion]