ok, so i think im going to keep the UCD400's. impressions to come later tonight...
im looking at some caps for the final power supply (im just using some cheap ones i had in the parts bin). it seems that ~20,000uf per channel, (10,000uf per rail) for a 4 ohm load is the optimum.
i dont want to spend a fortune on caps right now, they are pretty easily upgraded later as funds become available. im looking at two options...
1.) 2x 10,000uf 80v nichicon snap-ins
2.) 4 (or 6) 3,900 80v nichicon LK series snap ins.
ive always heard smaller value caps are better than a few large ones. is this the case? could i even do MANY 1,000uf caps too?
and are nichicons nice enough, or should i stick with a more "audiophile" grade cap?
i could also get 1,000uf 100v nichicons or panasonics for pretty cheap, and do 10 per rail, 20 per supply per channel. it would be cheaper than the larger value caps. it would just be more soldering work, which is fine.
im looking at some caps for the final power supply (im just using some cheap ones i had in the parts bin). it seems that ~20,000uf per channel, (10,000uf per rail) for a 4 ohm load is the optimum.
i dont want to spend a fortune on caps right now, they are pretty easily upgraded later as funds become available. im looking at two options...
1.) 2x 10,000uf 80v nichicon snap-ins
2.) 4 (or 6) 3,900 80v nichicon LK series snap ins.
ive always heard smaller value caps are better than a few large ones. is this the case? could i even do MANY 1,000uf caps too?
and are nichicons nice enough, or should i stick with a more "audiophile" grade cap?
i could also get 1,000uf 100v nichicons or panasonics for pretty cheap, and do 10 per rail, 20 per supply per channel. it would be cheaper than the larger value caps. it would just be more soldering work, which is fine.
Nichicons work fine. Panasonic FC may be a decent alternate if available in the sizes you are looking at. At least in low voltage supplies for my DAC and transport, these caps perform alike. Only Blackgates are better, but in the 10,000uF 80V size, you're looking at $170 a pop with BGs. I think a few smaller Nichicons or Panasonic, bypassed with quality film caps (e.g. a pair or 10uF and 0.47uF Auricaps per rail), should do the job.
Here's a list of caps I have collected over the last few days - not sure if this is anywhere near complete. If anyone can add some more I'd appreciate it (as well as comments to the parts listed here)
Here's a list of caps I have collected over the last few days - not sure if this is anywhere near complete. If anyone can add some more I'd appreciate it (as well as comments to the parts listed here)
Code:
ELNA Cerafine 64361 10,000uF 63V [url]http://www.partsconnexion.com/catalog/CapacitorsElectrolytic.html [/url]$32.50
Cornell-Dubliner 12,000uF 80V [url]http://www.partsconnexion.com/catalog/CapacitorsElectrolytic.html [/url]$11.95
Cornell-Dubliner 12,000uF 150V [url]https://www.alliedelec.com/Search/SearchResults.asp?SearchQuery=12%2C000uF&SearchType=STANDARD&submit=search+again [/url]$55.87
Cornell-Dubliner 10,000uF 63V [url]https://www.alliedelec.com/Search/SearchResults.asp?SearchQuery=10%2C000uF [/url]$5.84
Cornell-Dubliner 10,000uF 75V [url]https://www.alliedelec.com/Search/SearchResults.asp?SearchQuery=10%2C000uF [/url]$19.90
PACCOM 15,000uF 63V [url]http://www.partsconnexion.com/catalog/CapacitorsElectrolytic.html [/url]$11.95
Blackgate FK 10,000uF 63V [url]http://www.kyoto-electro.com/fk.html [/url]$140.20
Blackgate FK 10,000uF 80V [url]http://www.kyoto-electro.com/fk.html [/url]$168.68
Blackgate FK 10,000uF 100V [url]http://www.kyoto-electro.com/fk.html [/url]$189.80
Nichicon KG super 10,000uF 63V [url]http://www.percyaudio.com [/url]$21.95
Nichicon KG super 10,000uF 80V [url]http://www.percyaudio.com [/url]$36.95
Nichicon GSIV 10,000uF 100V [url]http://www.percyaudio.com [/url]$44.95
Nichicon KG super 15,000uF 63V [url]http://www.percyaudio.com [/url]$37.95
Nichicon KG super 15,000uF 80V [url]http://www.percyaudio.com [/url]$43.95
Nichicon KG super 15,000uF 100V [url]http://www.percyaudio.com [/url]$49.95
BHC ALS20/30 series 10,000uF 63V [url]http://www.thlaudio.com/indexE.htm [/url]$0.00
BHC ALS20/30 series 15,000 100V [url]http://www.thlaudio.com/indexE.htm [/url]$0.00
BHC slit foil 10,000uf 63V [url]http://www.schuro.de/preisl-slit-foils.htm [/url]$18.00
BHC slit foil 22,000uf 63V [url]http://www.schuro.de/preisl-slit-foils.htm [/url]$25.00
BHC T-Network 10,000uf 63V [url]http://www.schuro.de/preisl-tnet.htm [/url]$40.00
BC Components series 051 10,000uf 100v [url]http://www.schuro.de/preisl-bc051.htm [/url]$14.00
Epcos 10,000uF 100V [url]http://www.mercateo.com/p/live~s.20*102-501150[/url](2d)BP/AL_EL_KONDENSATOR_SCHRAUB_LL_10000UF100V.html $30.00
Panasonic ECE-T2AP103FA 10,000uF 100V [url]http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=300100&Row=424070&Site=US [/url]$14.08
Panasonic EEG-A2A103FHE. 10,000uF 100V [url]http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=300100&Row=424212&Site=US [/url]$23.51
Panasonic ECE-P2AP103HA 10,000uF 100V [url]http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=300100&Row=430254&Site=US [/url]$22.24
Panasonic ECE-T1KA103EA 10,000uF 80V [url]http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=300100&Row=425102&Site=US [/url]$12.54
Panasonic ECE-T2AA682FA 6,800uF 100V [url]http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=304148&Row=425383&Site=US [/url]$11.96
Panasonic ECE-T2AA822EA 8,200uF 100V [url]http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=304975&Row=425103&Site=US [/url]$13.37great, thanks.
it looks like for the price, the nichicons i found are a good way to go. in either scenario i gave, i would be spending a max of $30 for BOTH supplies (2x 2-channel supplies) for the caps. its a bargain. im looking to keep these amps cheap. i might build the "cost-no-object" ones next if i like the UCD's that much.
i have a buch of 0.47 WIMA caps in my parts bin. would those work well for bypass caps? or should i stick with a higher quality film cap?
it looks like for the price, the nichicons i found are a good way to go. in either scenario i gave, i would be spending a max of $30 for BOTH supplies (2x 2-channel supplies) for the caps. its a bargain. im looking to keep these amps cheap. i might build the "cost-no-object" ones next if i like the UCD's that much.
i have a buch of 0.47 WIMA caps in my parts bin. would those work well for bypass caps? or should i stick with a higher quality film cap?
cowanrg said:
I have Wima caps as bypass in my current amps. Depends on the type though (MKp4 I think is what's in there). I'd also put a much larger bypass cap on top of the 0.47uF, something between 10 and 100uF. There are some pretty cheap Poly film caps at ApexJr for up to 23uF I think - big suckers, but they do sound pretty good in my preamp PSU, and they don't cost too much.
Peter
cowanrg said:
Hi,
The nichicons seem alright. Since you've already decided on them I'll say that TNT audio made some comment as to the question of "one big cap or several in small ones paralleled" in their power supply design write ups. Worth reading.
I'd imagine the smaller caps would still end up taking alot of real estate in the case, and unless you make up a PCB for it you'll have a nice little antenna array.
That said I think you made the right choice.
You may want to read existing threads about bypassing them. Long story short (I'll save you some time) you can either snub your supply in order to compensate for the parasitic inductance, or leave it alone. The kind of high frequency bypass caps you're discussing is already there where it is most effective, on the module itself, and you really don't want to mess with it.
Regards,
Chris
pburke said:
yeah, mine are mks4's. i have 6 or so of them already, and the local surplus shop by me sells them for $0.25 each, so they wont be bad to use.
what do you mean by putting a larger one on top of it? so bypass the bypass cap?
i also saw a few people putting small poly caps on their bridge rectifiers. is this something is should do? i can get these caps VERY cheap as i stated locally, so the extra cost is nothing, so if there is a benefit, i will go for it.
classd4sure said:
well, i havent totally made up my mind, it just seems to be a good "value". there doesnt seem to be a significantly better option unless you spend 5x more.
taking up space isnt a concern. if anything, i want to fill some of the extra space that i have inside the case! i will probably be using a store-bought board that i would just mount caps to. my electronics store has some boards that would be setup very well for cap banks. or i would just get blank perfboard and solder them to it and make my own traces.
i have a bit of vertical space that wont be used. my idea was to put tranformer in front, than have extra boards (soft start, 12v trigger, etc.) on TOP on a shelf. then, UCD's would be mounted at back of case on the sides, and caps would be on another top shelf near the back.
ill have to find the size of the 1,000uf caps, but i like the idea of 20 of those per supply. ill have to read that TNT article though. answers to any of my questions would be helpful.
classd4sure said:
I tried a couple of combinations of bypass capacitors on the power supply, and could hear very little to no change, with if anything, a slight degradation in the sound. I used a couple of 10,000uf Nichicons on each rail which sounded excellent, but the sound improved a bit when I used some monster United Chemicons I had laying around, 2 82,000uf per rail.
GregD said:
I tried a couple of combinations of bypass capacitors on the power supply, and could hear very little to no change, with if anything, a slight degradation in the sound. I used a couple of 10,000uf Nichicons on each rail which sounded excellent, but the sound improved a bit when I used some monster United Chemicons I had laying around, 2 82,000uf per rail.
hum, interesting. there is a LOT of talk against using big cap banks.
what are your speakers? 4/8 ohm? do you run them full range, etc...
cowanrg said:
I have a pair of Wilson Watt/Puppy 6s, nominally 4 ohm. I'm using the UCD400s in a mono amp configuration, so each UCD400 has a total of 164,000uf of capacitance per rail. I also tried using a single 82,000uf capacitor per bank, and the difference was minimal, if any. Using 10,000uf Nichicons, I tried 10,000uf, 20,000uf, and 40,000uf per rail. 20,000uf sounded much better than 10,000uf, but there was no noticeable difference with 40,000uf. I've also read the comments about big cap banks, but it definitely sounded better with the big caps in my application, maybe they're just better quality. 😕
I also tried a couple of different transformers for the amps, a couple of Avel Lindberg 625VA, and custom wound 1500VA trannies running through dual bridge rectifiers before the capacitors. The 1500VA transformers sounded a little better than the Lindberg 625VA. The type of change I noticed with the bigger transformers was similar to the type of change I experienced with the bigger capacitors. In both cases, the noise floor seemed to be lowered, just a little bit of roughness or hashiness in the sound was reduced, and the presentation seemed to have more presence and be more effortless
But the overall change in sound for the options I tried is really quite small. If the best the UCDs have sounded in my system is 100, then knock off 1 or 2 for the smaller capacitors, and 1 or 2 for the smaller transformers. When I tried some smaller filter capacitors, any change I may have noted would be less than 1. By comparison, my prior amplifier, a Threshold S/500 II which I used to like quite a bit, I would rate as an 85 to 90 on the same scale.
GregD said:
good info. so basically dont go and dump twice as much money on each part, because you should only get a fraction of improvement...
i think i might go with 20,000uf per rail, 40,000uf per supply pre channel. its a good price/performance ratio. i can always tack on more if i need it. thankfully, my transformers are 1.7kva. i might go stereo with each case. in that situation, ill dump the single transformers and go with a single 800va for each amp. its a little less VA, but a separate transformer and cap bank for each amp, i think thats a better way to go.
good to hear from someone who has tried going bigger and found it to be benefitial. i was having a hard time understanding why more capacitance was worse...
cowanrg said:
Because it affects power factor adversly. Bruno recommended no more than 20 000uF, I think that was per rail per channel, but am not sure. If that's the case you're still in the optimal range.
However, this is true with all amps. AKSA from Aspen Amplifiers has some good notes on this stuff too. I believe his optimal was also 10 000uF, having said something like more can slow the attack down. That's not a direct quote though, you're free to check his website out yourself 🙂 Interesting reading there.
Regards,
Chris
Bruno made calculations resulting in 8250uF/rail/channel for 8 ohm use, double that for 4 ohm, it's in the monsterthread somewhere.
Even those calculations were based on a higher amp output than the musical average.
Hypex seems to follow that trend: Their HG PS is 10.000uF per rail per channel. I'm using a single 500VA for 2 channels, which might be a compromise, but still overdimensioned compared to most amps, even decent ones.
I think about 800VA per channel would be able to handle constant sinus power!
Anyway, I went this route, thinking I can always split up the amp by getting another 500VA/Highgrade PS and go dual mono with power to spare.
We should never forget that these amps are 93% effective, and get a lot more out of a supply than class AB ever will.
Even those calculations were based on a higher amp output than the musical average.
Hypex seems to follow that trend: Their HG PS is 10.000uF per rail per channel. I'm using a single 500VA for 2 channels, which might be a compromise, but still overdimensioned compared to most amps, even decent ones.
I think about 800VA per channel would be able to handle constant sinus power!
Anyway, I went this route, thinking I can always split up the amp by getting another 500VA/Highgrade PS and go dual mono with power to spare.
We should never forget that these amps are 93% effective, and get a lot more out of a supply than class AB ever will.
I would recommend this to reduce switching noise on the rectifiers. I have heard that it makes more of a difference than using the more exotic rectifier diodes, but I can't seem to find the reference to the test. Anyway, I have also heard that this is the best place to put bypass caps in a UcD power supply- there is already a snubber on the module.
Anyway- small film caps are cheap, bridge rectifiers are cheap and it takes a couple minutes to do- try it and, if you don't like it, put it back- you won't be out more than $5.
Best of luck-
Larry
ClassD, you're not correct, but neither was I...
Bruno's specs were 6250uF (and 12500uF in 4 ohms) - so 10.000uF is a nice value.
Bruno's specs were 6250uF (and 12500uF in 4 ohms) - so 10.000uF is a nice value.
Bruno recommends 47nf at the rectifiers. I wonder if it would be even better to put an RC instead to reduce further the switching noise. Say 1R(carbon composite) + 47nf MKP.
I surelly need a scope for Christmass 😀
My 10000 uF Sikorel's would handle 30A picks, if I understood the chart
Regards
M
ok, after all your guy's talk, i think im going to do this...
3x 3900uf nichicons per rail, 6 per supply. so that gives me 23,400uf per supply, right below the "ideal" 25,000uf.
my conversions are always terrible, 47nf is equal to 0.047uf right? im not sure if i have any of those laying around. i might as well put some caps across the rectifiers, no one seems to think it hurts anything, and its cheap. plus, i FINALLY got leads for my scope and its running. its going to take forever to learn how to use it (ive never used a scope before), but i hooked it up to the AC outlet and i saw a nice (not perfect though) AC sinewave. so, its working i guess.
i havent put it on my amp to test it yet though. ill put it on the supply and see how noisy it is, or if it is.
OH, im also getting a little bit of mechanical hum through one of my transformers. here is the thing... it ONLY happens with one of them, the other is silent with ear touching the transformer. and, it doesnt always happen. sometimes, its just as silent as the other, but other times, it buzzes kinda loud. its not "bad". i could put some rubber washers on top and bottom of it to stop if from transmitting into the case, but i would like to know why its doing that.
i read somewhere that when DC gets into a transformer it can do that. i measured like 0.5vdc coming into it. is that enough? it doesnt really seem like it... its not a big deal, it happens maybe a few minutes out of every day it seems. as soon as i start messing with it (turning stuff on and off to find out of anything is causing it) it just stops on its own.
3x 3900uf nichicons per rail, 6 per supply. so that gives me 23,400uf per supply, right below the "ideal" 25,000uf.
my conversions are always terrible, 47nf is equal to 0.047uf right? im not sure if i have any of those laying around. i might as well put some caps across the rectifiers, no one seems to think it hurts anything, and its cheap. plus, i FINALLY got leads for my scope and its running. its going to take forever to learn how to use it (ive never used a scope before), but i hooked it up to the AC outlet and i saw a nice (not perfect though) AC sinewave. so, its working i guess.
i havent put it on my amp to test it yet though. ill put it on the supply and see how noisy it is, or if it is.
OH, im also getting a little bit of mechanical hum through one of my transformers. here is the thing... it ONLY happens with one of them, the other is silent with ear touching the transformer. and, it doesnt always happen. sometimes, its just as silent as the other, but other times, it buzzes kinda loud. its not "bad". i could put some rubber washers on top and bottom of it to stop if from transmitting into the case, but i would like to know why its doing that.
i read somewhere that when DC gets into a transformer it can do that. i measured like 0.5vdc coming into it. is that enough? it doesnt really seem like it... its not a big deal, it happens maybe a few minutes out of every day it seems. as soon as i start messing with it (turning stuff on and off to find out of anything is causing it) it just stops on its own.
Yves Smolders said:
i will. i have today, tomorrow, thursday, friday, saturday, and sunday off. i will be playing with them the whole time. i still havent switched back to my rotel amps, but so far, im happy with them.
i should have an actual decision made by tomorrow 🙂
- Status
- Not open for further replies.