Hi all, resurrecting this thread sorry!
I am in the process of getting together a list of parts to mod my 640A V2 and looking for a bit of advise.
So far I have the following worked out:
Thanks in advance!
Pete
I am in the process of getting together a list of parts to mod my 640A V2 and looking for a bit of advise.
So far I have the following worked out:
- Upgrade the power supply electrolytics from 2200uF to 3300uF. I have seem some use Panasonic M and others use Nichicon FW. I can find the Nichicon FW caps through X-ON but not sure where to get the Panasonic M series from. Any reason why I should go one over the other?
- Should I upgrade all other electrolytics in the amp? Looking at a combination of Nichicon FG/KZ/ES and Elna Silmic II caps. Again all available from X-ON.
- Bypass balance pot as I never use it, and maybe the tone control...
- Some replace only one opamp (U7 I think) others replace all of them. What is the general consensus? Will be using the AD8599ARZ.
Thanks in advance!
Pete
Also another thing I have noticed is that in the power supply section, the factory is 2200uF 63V but most people seem to be using 3300uF 50V (to keep the case size the same). Is this going to allow enough voltage headroom? I am not sure what voltages are present here and can't measure it as I have pulled everything apart!!!
X-ON are Mouser agents and resellers for those that aren't aware of parts sellers and buying options in Oz or no clue who they are. They don't promote the full catalogue though.
The service manual is the usual pace to find parts details but there is no dimensional specification for the main smoothing caps. They sit close together though, so there is no room for error. I hope you didn't scrap the original parts yet because there is no way back unless someone opens up their amplifier to measure and remind you: Power Amp Schematic (left Channel) - Cambridge Audio 640A Service Manual [Page 19]
Cambridge Audio 640A Service Manual (Page 24 of 41)
The supply voltage rails are +/- 45V in the original version. The specified caps were 50V and output was 65W/8R rated. V2 is 75W/8R rated which suggests the power rails are higher at possibly 51V nominal, so 63V electrolytics would likely be necessary.
Think: You could surely measure the transformer's AC unloaded output and calculate the DC voltage after the rectifier anyway, even if all is in pieces.
The service manual is the usual pace to find parts details but there is no dimensional specification for the main smoothing caps. They sit close together though, so there is no room for error. I hope you didn't scrap the original parts yet because there is no way back unless someone opens up their amplifier to measure and remind you: Power Amp Schematic (left Channel) - Cambridge Audio 640A Service Manual [Page 19]
Cambridge Audio 640A Service Manual (Page 24 of 41)
The supply voltage rails are +/- 45V in the original version. The specified caps were 50V and output was 65W/8R rated. V2 is 75W/8R rated which suggests the power rails are higher at possibly 51V nominal, so 63V electrolytics would likely be necessary.
Think: You could surely measure the transformer's AC unloaded output and calculate the DC voltage after the rectifier anyway, even if all is in pieces.
Last edited:
Hi Ian,
I haven't pulled the caps out yet, just had the boards out to work out exactly what capacitors I need. I have all the factory values now, just need to double check diameters and pitch with the potential replacements to make sure everything fits.
There are a few 63V caps in the power supply, not just the eight 2200uF ones, so I probably should put it all back together and take some measurements to be safe. Will be very annoying if it does need 63V as that means I will have to stick with 2200uF instead of being able to increase the size.
I haven't pulled the caps out yet, just had the boards out to work out exactly what capacitors I need. I have all the factory values now, just need to double check diameters and pitch with the potential replacements to make sure everything fits.
There are a few 63V caps in the power supply, not just the eight 2200uF ones, so I probably should put it all back together and take some measurements to be safe. Will be very annoying if it does need 63V as that means I will have to stick with 2200uF instead of being able to increase the size.
Its Cambridge Audio, right? Not a single part of any significant cost will be higher rated than it must be, so the use of 50V rated parts won't be an option where 63V parts are specified and used. Use like for like and no less.
It's no loss really, if you can't squeeze more smoothing capacitance in. Nominal 10,000uF is absolutely fine for many amplifiers up to 100W/8R, Some folk are too busy parroting all-too generalised mantras like "add more capacitance" - "everyone does it, it makes night and day improvements if you use brand X like me" - "you can never have too much capacitance" etc. - but you can have too much with class AB. I've heard the differences subjectively with many iterations of basically the same few amplifier designs that I produced over some years. There are smarter ways to get the most out of smoothing capacitance rather that just kludge the amplifier with more.
Occasionally, more capacitance delivers crisper bass from wimpy transformer power supplies - often though, it just clags the life out of the sound for reasons I'm still not very clear on. Unless you use the amplifier at clipping levels, I would not suggest upping the smoothing capacitance and really, it can't be more than 10 years old and a recap isn't due yet unless the caps are 1,000 hr no-name trash, already showing bulges if not leaking.
It's no loss really, if you can't squeeze more smoothing capacitance in. Nominal 10,000uF is absolutely fine for many amplifiers up to 100W/8R, Some folk are too busy parroting all-too generalised mantras like "add more capacitance" - "everyone does it, it makes night and day improvements if you use brand X like me" - "you can never have too much capacitance" etc. - but you can have too much with class AB. I've heard the differences subjectively with many iterations of basically the same few amplifier designs that I produced over some years. There are smarter ways to get the most out of smoothing capacitance rather that just kludge the amplifier with more.
Occasionally, more capacitance delivers crisper bass from wimpy transformer power supplies - often though, it just clags the life out of the sound for reasons I'm still not very clear on. Unless you use the amplifier at clipping levels, I would not suggest upping the smoothing capacitance and really, it can't be more than 10 years old and a recap isn't due yet unless the caps are 1,000 hr no-name trash, already showing bulges if not leaking
.
The 540A V2 is:
60W into 8 ohms
90W into 4 ohms
The 640A V2 is:
75W into 8 ohms
120W into 4 ohms
I think you are right though regarding leaving the power caps the same value. I had a look at the schematics and I think there is the possibility of causing other issues as the unregulated DC voltage level would increase.
So will just upgrade them to a better quality cap, but same value and voltage. Nichicon FW have the correct cap. Could also get the FK, but given that this is simply filtering the power and not in the audio line it isn't really worth it.
Here is the schematic of the power supply caps and rails:
60W into 8 ohms
90W into 4 ohms
The 640A V2 is:
75W into 8 ohms
120W into 4 ohms
I think you are right though regarding leaving the power caps the same value. I had a look at the schematics and I think there is the possibility of causing other issues as the unregulated DC voltage level would increase.
So will just upgrade them to a better quality cap, but same value and voltage. Nichicon FW have the correct cap. Could also get the FK, but given that this is simply filtering the power and not in the audio line it isn't really worth it.
Here is the schematic of the power supply caps and rails:
OK more queries!! This time on capacitor type. For the power supply/amplifier PCB, I can't use Elna Silmic II anymore as they don't have the values in 63V only 50V.
So looking at the Nichicon range. I can find them in FW and KW (bi-polar I have all Nichicon Muse ES). Anyone have any preference or other recommendations?
Cheers
Pete
So looking at the Nichicon range. I can find them in FW and KW (bi-polar I have all Nichicon Muse ES). Anyone have any preference or other recommendations?
Cheers
Pete
The power amplifier designs and supplies are essentially the same in 540/2 and 640/1 so rating it 60W instead of 65 is likely a paperwork figure, perhaps permitting a smaller transformer to be used and give people a reason to pay more for the 640A V2, which you could say is no different to 540A V2 in practical listening terms and features. It's not all strict scientific fact in marketing so I wouldn't stake my life or even a recap job on what can be squeezed out of SAP15s with 45V rails. The SAP15s are now obsolete but STD01 carry on where they left off.
http://www.semicon.sanken-ele.co.jp/sk_content/std01n_ds_en.pdf
In theory, 45V rails should permit 100W/8R with an EF output stage but I suspect the AC from the small transformer secondary windings is quite rubbery and dips under load. This is common in UK audio design and works well to protect the output devices in the event of overload but at the cost of dynamic range. The modest difference between power into 8R and 4R loads all but spells out this design approach.
Don't be concerned about an increased DC rail voltage with 50% more capacitance, as it won't be that much and will quickly vanish under output load, where voltage will count. The question centres around the benefit, if any, of increasing capacitance above even as little as 4,700uF. Unless the amplifier is higher power and thrashed at its peak power levels, the extra can't help linearity unless it just compensates for badly deteriorated caps - unlikely at present.
Missed your post but if you have Nichicon KW, not the best but fine because of size mainly.
http://www.semicon.sanken-ele.co.jp/sk_content/std01n_ds_en.pdf
In theory, 45V rails should permit 100W/8R with an EF output stage but I suspect the AC from the small transformer secondary windings is quite rubbery and dips under load. This is common in UK audio design and works well to protect the output devices in the event of overload but at the cost of dynamic range. The modest difference between power into 8R and 4R loads all but spells out this design approach.
Don't be concerned about an increased DC rail voltage with 50% more capacitance, as it won't be that much and will quickly vanish under output load, where voltage will count. The question centres around the benefit, if any, of increasing capacitance above even as little as 4,700uF. Unless the amplifier is higher power and thrashed at its peak power levels, the extra can't help linearity unless it just compensates for badly deteriorated caps - unlikely at present.
Missed your post but if you have Nichicon KW, not the best but fine because of size mainly.
Last edited:
I haven't bought any of the caps yet, just working my way through the boards to find what I can replace them with. I wanted Elna Silmic II but they don't come in the 63V required Might be able to use them on the smaller voltage caps. Hence looking at the KW for the 63V ones unless you have a better suggestion??
Using Nichicon PW for the 8 PSU caps. Nichicon ES for the bi-polar.
Edit: Found the Nichicon FG and KZ series. They appear to be better so checking them out.
Might be able to use them on the smaller voltage caps. Hence looking at the KW for the 63V ones unless you have a better suggestion??Using Nichicon PW for the 8 PSU caps. Nichicon ES for the bi-polar.
Edit: Found the Nichicon FG and KZ series. They appear to be better so checking them out.
Last edited:
if you check scheme carefully you will notice there is already 2x2200uf per each rail per channel as caps are in parallel. you will not find same voltage rating cap with more capacitance as it broke some base physics then. what you can do is to use some free space near to power supply coil and add new caps there. this is how I've done my mod by adding 3300uf to existing one. you need 3 wires and relevant bridging within new 4 caps section (per channel).
unfortunately tapatalk is broken again and i can't add pictures here.
tapnięte talkiem
the differece between 540 and 640 is mainly in transformer, the letter one has additional secondary for power stage, both rated 2*32.5V/2A wherein 540 has single secondary rated 2*30V/3A.
i've not finished my mod as awaiting for replacement of rubbish xunda's all around but quick check confirms it sounds with 20kuF at least the same as before mod
for me more important is to replace relays which cause channels to off and to add dac for optical input handling from media streamer. and ofc it's a lot fun
i've not finished my mod as awaiting for replacement of rubbish xunda's all around but quick check confirms it sounds with 20kuF at least the same as before mod
for me more important is to replace relays which cause channels to off and to add dac for optical input handling from media streamer. and ofc it's a lot fun
I noticed that the decoupling caps throughout the whole board are cheap 47nF 63V polyester MKT.
Now I can upgrade these to some WIMA polypropylene MKP which seems to be a common thing to do, but what is the consensus on on using tantalum caps? They typically have very good high frequency rejection which I would have though to be an advantage in a decoupling capacitor? Would replace them with 100nF 35V tant.
I have seen some threads mentioning that tants are not recommended in the signal path, but these aren't, they are just used for rail decoupling on the opamps.
Now I can upgrade these to some WIMA polypropylene MKP which seems to be a common thing to do, but what is the consensus on on using tantalum caps? They typically have very good high frequency rejection which I would have though to be an advantage in a decoupling capacitor? Would replace them with 100nF 35V tant.
I have seen some threads mentioning that tants are not recommended in the signal path, but these aren't, they are just used for rail decoupling on the opamps.
What's wrong with the original MKT film caps or Multilayer ceramics - the preferred types for decoupling? Size is important with unshielded components and polypropylene film types lose by size, lead length and noise proneness what may be gained in microscopically lower distortion. Tantalums may work fine, just like wet electrolytics, if suitable voltage ranges are available and there is no possibility of exceeding it. They are polarized, not long-life components and only have a size advantage to offer audio.
Don't swap everything in sight based on a few specifications since there are more factors in component choices than the obvious ones that beginners make such a blinkered, ritualistic fuss of. I suspect there is little to improve in such a recent model and most that is done will be retrograde as far as performance is concerned. How it sounds has little to do with fidelity though - that's more the "flavour" of distortion.
Don't swap everything in sight based on a few specifications since there are more factors in component choices than the obvious ones that beginners make such a blinkered, ritualistic fuss of. I suspect there is little to improve in such a recent model and most that is done will be retrograde as far as performance is concerned. How it sounds has little to do with fidelity though - that's more the "flavour" of distortion.