ray bronk said:
Hi Ray,
I replaced the output caps in my UcD 180 modules that are used for midrange and tweeters in an active speaker system.
My impression after I changed to the WIMA caps is that the soundstage is larger and deeper and more acurate at the same time. With accurate I mean the stability of the position of each instrument. Also stability of sounds that are at the extreme left or right and in front of the speakers, as appear for example in "Amused to death" of Roger Waters, a CD I always use to quickly check imaging, especially imaging in front of the speakers and at the sides. I also think to hear increased resolution that make low level sounds that were previously more masked by the higher level sounds easier distinguishable. The sound seems to be somewhat "cleaner".
I have to admit, I did no AB comparison and before I did the mod I had not been listening regularly (very busy at work). However, since the mods, I spent more time listening, a good sign I would say. The WIMA cap mod is a very cheap one. I think they were about US$ 1.50 a piece at Mouser, so 3$ per UcD. Worth the try I would say and they are quite easy to install as they are small.
I also can not say that I miss dynamics with the UcDs. In fact I think that they are more dynamic than many other amps. However, as the sound stays clean even at very high level, they may subjectively sound less loud. When I just had them and used them on my passive speakers, just for testing, I quickly realized that I turned up the volume much higher than normal as the sound stayed so clean (no significant increase in distortion at high levels I guess).
Best regards
Gertjan
Hi,
Gerjan, I can agree totally it is very easy to turn it up and up and up... sometimes it's 4AM and I've got it way too loud and don't even realize it. Even just going from one song with its' particular recording level to the next can be a very eye opening experience.
Dynamics also come into play at the source, alot of compressed junk is out there these days, play something orchestral or that new tool album I mentioned and it's very mindblowing how dynamic it truly is, with the stock decoupling caps (at least for my module) the same is true for the 3D staging.
Poor wiring etc will diminish the the dynamics somewhat, the soundstage, and also the musicality, which can be measured as the level of enjoyement of the prior two. So listening to the amp more is actually really an excellent sign of improvement.
I found much the same as Gertjan changing my filter cap, I'm also fairly certain we both had the older style of filter cap though, and the newer one may be very much improved, or not, experimetc.
With the new modules you personally may prefer the sound of the AD op amp with the CRD tied to the negative rail, or you may prefer the sound of the 2134 with the CRDs tied to the positive rail. You truly do have to experiment.. I'd say practice your soldering and go from there. Old motherboards and peripheral cards are a dime a dozen if not cheaper, hack em up.
I do use good quality solder as well, I can't say it affects the sound, but it does give a much stronger joint.
No I don't think you'll be able to mod the STD supply with any 4 pole cap, you're dealing with 4 leads instead of 2.
Regards,
Chris
Gerjan, I can agree totally it is very easy to turn it up and up and up... sometimes it's 4AM and I've got it way too loud and don't even realize it. Even just going from one song with its' particular recording level to the next can be a very eye opening experience.
Dynamics also come into play at the source, alot of compressed junk is out there these days, play something orchestral or that new tool album I mentioned and it's very mindblowing how dynamic it truly is, with the stock decoupling caps (at least for my module) the same is true for the 3D staging.
Poor wiring etc will diminish the the dynamics somewhat, the soundstage, and also the musicality, which can be measured as the level of enjoyement of the prior two. So listening to the amp more is actually really an excellent sign of improvement.
I found much the same as Gertjan changing my filter cap, I'm also fairly certain we both had the older style of filter cap though, and the newer one may be very much improved, or not, experimetc.
With the new modules you personally may prefer the sound of the AD op amp with the CRD tied to the negative rail, or you may prefer the sound of the 2134 with the CRDs tied to the positive rail. You truly do have to experiment.. I'd say practice your soldering and go from there. Old motherboards and peripheral cards are a dime a dozen if not cheaper, hack em up.
I do use good quality solder as well, I can't say it affects the sound, but it does give a much stronger joint.
No I don't think you'll be able to mod the STD supply with any 4 pole cap, you're dealing with 4 leads instead of 2.
Regards,
Chris
ray bronk said:
Hi Ray,
In the UcD180s I have the NE5532 chip, actually I have one of the first UcD180 modules. The coupling caps on those modules have been replaced with BG-N types, relatively small ones as I use them for tweeter and midrange. I have UcD400 modules in use for the woofers, those have the opa2134. I have other UcD400 modules with the AD8620 but have not installed those yet. Plan to use those instead of the UcD180 modules for mid and tweeter amps. I know, from a power standpoint of view it is complete overkill, however, I expect that the UcD400 maybe sonically better than the UcD180, although the UcD180 sounds great.
I also have bought the SST511 current source to be used as a load for the AD8620. Plan to install those and the WIMA caps on the UcD400 modules. For the 470uF caps, I have Panasonic FC 680uF 100V at home and plan to get my hands on Rubycon ZL caps as well and try those out. For power supplies I'm thinking of using SMPS (actually now using an SMPS for the woofer amps). I want to try the coldamp SMPS once available (should be soon).
Best regards
Gertjan
Best regards
Gertjan
deandob said:
classd4sure said:
Hey all,
I can try to take on the group buy. Any US distributers you guys can steer me to.
I've got a good cover for my day job into why we'd need such a cap, and we have a shipping department I can use to, so it might be less painful for me.
All I ask is that you help me find a source.
If need be, I have channels I could tap in china perhaps... if that's where they are made.
Mike
Hi Mike,
Very cool of you to offer us that.
If you check here
http://www.rubycon.co.jp/en/contact/
You'll have the list of all of their distributors. I contacted one in canada and they weren't at all interested unless it was large volume, so I think we ought to start with seeing what kind of interest we can generate, and maybe then have an easier time in finding a distributor to work with.
TY.
Very cool of you to offer us that.
If you check here
http://www.rubycon.co.jp/en/contact/
You'll have the list of all of their distributors. I contacted one in canada and they weren't at all interested unless it was large volume, so I think we ought to start with seeing what kind of interest we can generate, and maybe then have an easier time in finding a distributor to work with.
TY.
classd4sure said:
Started to email them then thought, heck, not sure what I should get?
What ZL's do we want if we can get anything?
What would be "okay" to get? voltage ranges, cap ranges, diameter...
I'm quite busy right now with a big work project. You guys hack it out on what one or two values I should try to buy, then I'll worry about minimums and that.
I personally wouldn't rule out 63V caps, even for a UcD400.
Especially if this LTC3703 looks as good in reality as on paper 🙂
Guys, that part might just be the ideal regulator for single ended class D!
One for each rail, after the big iron, so not off line for several good resons.
Just trying to decide if its best to configure it as a buck or a boost. Both could work, the buck (step down) having a much faster transient response (beyond the audio band) but the boost might have some advantages to in that the fet rating will depend on a regulated voltage rather than a potentially pumped (and ideally not with lots of cap for xfrmer reasons)
Regards,
Mike
ZL caps
I put in a question to Rubycon USA as to what volume it would take to order direct. Some manufactures have surprisingly small requirements, worth a shot. I also asked about stock in the US. This was for part # 63 ZL1500M 18×40
Do you guys think this part can be fitted to the UcD400? It looks real close. That is my main interest although a bank of them could work very well for storage. It would be expensive and awkward with so many being needed.
Roger
I put in a question to Rubycon USA as to what volume it would take to order direct. Some manufactures have surprisingly small requirements, worth a shot. I also asked about stock in the US. This was for part # 63 ZL1500M 18×40
Do you guys think this part can be fitted to the UcD400? It looks real close. That is my main interest although a bank of them could work very well for storage. It would be expensive and awkward with so many being needed.
Roger
Rubycon
Mike,
Can you send me email as your forum profile has mail disabled so I can't email you. I have a source for the Rubycons but only for 120 of the 1500uF 63v then the next size order is 1000.
I'd suggest the lots of 1500uF 63v for the power supply (eg. I need 40 for a 5 channel amp!) and the 470uF 100v or 680uF 63v for the onboard power decoupler cap.
Regards,
Dean
Mike,
Can you send me email as your forum profile has mail disabled so I can't email you. I have a source for the Rubycons but only for 120 of the 1500uF 63v then the next size order is 1000.
I'd suggest the lots of 1500uF 63v for the power supply (eg. I need 40 for a 5 channel amp!) and the 470uF 100v or 680uF 63v for the onboard power decoupler cap.
Regards,
Dean
I've a few comments at this point of the game.
If you haven't tried it yourself, it is evident that the capacitance value has a big enough effect with frequency response. A bigger cap in this location gives improved bass and attenuated highs, a smaller cap will have produce the same bass not but you'll experience less power behind it (very thin sound) while the highs will seem to be boosted. 470uF seems to give the best overall response, based on the stock value.
Therefore I'd recommend not attempting anything in the range of >1000uF because it will suffer a great lack of air.
Looking out for the UCD180 interests, the biggest that fits is 12.5mm diameter, therefore that leaves us with the biggest possible cap being 680uF 63V 12.5mmX40mm,
I'd also like to try 63V 560uF 12.5mmX35mm, and
63V 470uF 12.5X30mm.
While it is not enough of a margin of safety for me, I'd also be interested in trying 50V 820uF 12.5X35mm.
That should allow to enough of a range to tune for the best overall frequency response.
I think UCD400 owners might appreciate a margin of safety on these caps, which don't feature a rated surge voltage, and are much smaller in value than the standard singular supply cap.
Supply caps ought to be considered seperatly I think, if it makes too much of an overal cost difference.
Regards,
Chris
If you haven't tried it yourself, it is evident that the capacitance value has a big enough effect with frequency response. A bigger cap in this location gives improved bass and attenuated highs, a smaller cap will have produce the same bass not but you'll experience less power behind it (very thin sound) while the highs will seem to be boosted. 470uF seems to give the best overall response, based on the stock value.
Therefore I'd recommend not attempting anything in the range of >1000uF because it will suffer a great lack of air.
Looking out for the UCD180 interests, the biggest that fits is 12.5mm diameter, therefore that leaves us with the biggest possible cap being 680uF 63V 12.5mmX40mm,
I'd also like to try 63V 560uF 12.5mmX35mm, and
63V 470uF 12.5X30mm.
While it is not enough of a margin of safety for me, I'd also be interested in trying 50V 820uF 12.5X35mm.
That should allow to enough of a range to tune for the best overall frequency response.
I think UCD400 owners might appreciate a margin of safety on these caps, which don't feature a rated surge voltage, and are much smaller in value than the standard singular supply cap.
Supply caps ought to be considered seperatly I think, if it makes too much of an overal cost difference.
Regards,
Chris
Re: ZL caps
Roger,
18mm will fit a UcD400, but its almost touching if not touching the Tbar for the heat sink.
that said, its likely a bigger deal for life than running 63V since it will be hotter to much hotter depending on your heatsink.
I would bet not a sole on this forum will ever leave the amp alone long enough for a cap to go bad, and if so, its a pretty trivial thing to change.
Chris,
Interesting that they don't have a surge rating on these caps.
Also, thanks for the input on capacitance vs bass. So you got around to trying a 470uF FC in your amp?
I've got some 63V 470uF FC's I haven't tried, and I guess you just saved me the time.
I'll keep in mind that we will need 63V 12.5mm caps for UcD180.
btw, what is the pitch required?
I'm starting to wonder if your results with the FC, 680uF, ?V might have something to do with the package, i.e. did you used the 12.5mm ones and I tried the 16mm and 18mm ones?
Brono onced mentioned that the case size changes the sound, and in some families its the big ones that are best and in others the small ones. Just a thought!
Thanks
Has anyone tried using two stock caps or M series yet?
I think the concern over lead length is over blown (if kept to a resonable minimum, as the traces, at least on the UcD400, are about an inch long before the fets.
t.
About the LM6172, again, haven't listened yet, but I did put them in a dead 400 6.1 and powered them up. It was stable and well behaved with square waves. That's as far as I've got.
mike
sx881663 said:
Roger,
18mm will fit a UcD400, but its almost touching if not touching the Tbar for the heat sink.
that said, its likely a bigger deal for life than running 63V since it will be hotter to much hotter depending on your heatsink.
I would bet not a sole on this forum will ever leave the amp alone long enough for a cap to go bad, and if so, its a pretty trivial thing to change.
Chris,
Interesting that they don't have a surge rating on these caps.
Also, thanks for the input on capacitance vs bass. So you got around to trying a 470uF FC in your amp?
I've got some 63V 470uF FC's I haven't tried, and I guess you just saved me the time.
I'll keep in mind that we will need 63V 12.5mm caps for UcD180.
btw, what is the pitch required?
I'm starting to wonder if your results with the FC, 680uF, ?V might have something to do with the package, i.e. did you used the 12.5mm ones and I tried the 16mm and 18mm ones?
Brono onced mentioned that the case size changes the sound, and in some families its the big ones that are best and in others the small ones. Just a thought!
Thanks
Has anyone tried using two stock caps or M series yet?
I think the concern over lead length is over blown (if kept to a resonable minimum, as the traces, at least on the UcD400, are about an inch long before the fets.
t.
About the LM6172, again, haven't listened yet, but I did put them in a dead 400 6.1 and powered them up. It was stable and well behaved with square waves. That's as far as I've got.
mike
Hi Mike,
Lead pitch for the 180's is 5mm. I dont' know offhand what it is exactly for the 400.
Regarding Bruno's comments and your experienced results perhaps differing with mine. We have fairly different modules and implementations.
Bruno's comments relating to case size which I believe you're refering to were more towards differences in orders of magnitude, in that if a smaller cap of a certain variety sounds good, then the same cap in a much bigger case size (for power supplies) likely wouldn't, and vice versa. He'd found that untrue with black gates though. Point being.... experiment, and that's why I think a variety of sizes that will work will be very nice to try if at all possible, this is of course running on the assumption from most of us that the ZL caps actually do sound good. I for one have never heard one.
In keeping with my experimentation theme, I have not saved you the time. I've done a few experiments with different cap values on a homebrew UCD and the results were the same. Now having gone over to 820uF FC's on the module up from the stock 470uF's the increase in bass went along with previous tests. The lack of 3D and coloration is I guess a characteristic unique to the type of cap over the size alone.
That's not at all to say a 10 000uF 100V FC won't have a different sort of character as Bruno had experienced before, but I highly doubt it's of any interest given the case sizes of interest here.
So, I'd predict that if you put the 470uF FC caps in instead of 680uF or whatever it is you have, you'll experience a shift in response or Q towards the higher frequency spectrum.
470uF leaves the bass too weak, 820uF leaves the highs too weak, it's a shame we can't put a variety in place and tweak it accordingly, but it's easy to understand why Hypex does not do so. Again it stands to reason everything must be optimal before that sort of change makes an audible difference apparent, I'm guessing, or also maybe the UCD400 and newer versions of it are more robust against this type of change given the differing values in the CLC rail filters, would be worth looking into .... if you had the means.. or JP could chyme in and laugh saying we're all wasting our time and money because we're barking up the wrong tree?
I believe I will be ordering a few 136 RVI caps to try in this case this group buy falls flat or I can't afford them when the time comes.
As far as any caps touching the t-sink, I'd only be worried about the heat alone, not so much noise coupling as the T-sink is well bypassed.
If a bigger diameter cap is found to be what works best, I will be the first to take a dremel to the T-sink, and won't lose any sleep over it.
BTW, the 12.5mm FC caps on my 180 DO touch it, hard to avoid as there's so very little clearance, I doubt it makes them run much hotter to have to be concerned with, I still haven't got these suckers hot yet.
Regards,
Chris
Lead pitch for the 180's is 5mm. I dont' know offhand what it is exactly for the 400.
Regarding Bruno's comments and your experienced results perhaps differing with mine. We have fairly different modules and implementations.
Bruno's comments relating to case size which I believe you're refering to were more towards differences in orders of magnitude, in that if a smaller cap of a certain variety sounds good, then the same cap in a much bigger case size (for power supplies) likely wouldn't, and vice versa. He'd found that untrue with black gates though. Point being.... experiment, and that's why I think a variety of sizes that will work will be very nice to try if at all possible, this is of course running on the assumption from most of us that the ZL caps actually do sound good. I for one have never heard one.
In keeping with my experimentation theme, I have not saved you the time. I've done a few experiments with different cap values on a homebrew UCD and the results were the same. Now having gone over to 820uF FC's on the module up from the stock 470uF's the increase in bass went along with previous tests. The lack of 3D and coloration is I guess a characteristic unique to the type of cap over the size alone.
That's not at all to say a 10 000uF 100V FC won't have a different sort of character as Bruno had experienced before, but I highly doubt it's of any interest given the case sizes of interest here.
So, I'd predict that if you put the 470uF FC caps in instead of 680uF or whatever it is you have, you'll experience a shift in response or Q towards the higher frequency spectrum.
470uF leaves the bass too weak, 820uF leaves the highs too weak, it's a shame we can't put a variety in place and tweak it accordingly, but it's easy to understand why Hypex does not do so. Again it stands to reason everything must be optimal before that sort of change makes an audible difference apparent, I'm guessing, or also maybe the UCD400 and newer versions of it are more robust against this type of change given the differing values in the CLC rail filters, would be worth looking into .... if you had the means.. or JP could chyme in and laugh saying we're all wasting our time and money because we're barking up the wrong tree?
I believe I will be ordering a few 136 RVI caps to try in this case this group buy falls flat or I can't afford them when the time comes.
As far as any caps touching the t-sink, I'd only be worried about the heat alone, not so much noise coupling as the T-sink is well bypassed.
If a bigger diameter cap is found to be what works best, I will be the first to take a dremel to the T-sink, and won't lose any sleep over it.
BTW, the 12.5mm FC caps on my 180 DO touch it, hard to avoid as there's so very little clearance, I doubt it makes them run much hotter to have to be concerned with, I still haven't got these suckers hot yet.
Regards,
Chris
classd4sure said:
Lead pitch for the UcD400 is 7.5mm, an 18mm diameter cap is just not touching the T-shape heatsink.
What size is good (in uF) I can't tell, have not tried. Would be nice to know the inductance of those ferrite cores/beads on the UcD board, we could then calculate/simulate where the resonance frequency (and Q factor) of the LC combination would go (could also try to measure it sometime). With bigger caps, resonance frequency and Q-factor would go down but you probably do not want the resonance frequency to end up below 20kHz, especially not for an amp that drives the tweeters. For a woofer and midrange amp, bigger caps may therefore be OK, or even desired while for a tweeter a smaller one may sound better? I have no experience with that but it maybe an explanation for Chris's experiences?
Best regards
Gertjan
Gertjan,
I think it is certainly one to run with.
On my homebrew amp I actually had no type of pi filter but.... I also had no scope to see what was going on, so it might be explainable at the cap itself along with the usual parasitics, but the type of interraction you mention ought to be looked into I think.
BTW on the homebrew what I had was two 470uF's in // per rail. It gave excellent bass response and attenuated highs. Removing one of them gave a more rounded response. Nothing else was in line with them up to the mosfet, expect for an inch and a half of trace/wire. We ought to account for anything that is in line with them though, to understand it fully.
Regards,
Chris
I think it is certainly one to run with.
On my homebrew amp I actually had no type of pi filter but.... I also had no scope to see what was going on, so it might be explainable at the cap itself along with the usual parasitics, but the type of interraction you mention ought to be looked into I think.
BTW on the homebrew what I had was two 470uF's in // per rail. It gave excellent bass response and attenuated highs. Removing one of them gave a more rounded response. Nothing else was in line with them up to the mosfet, expect for an inch and a half of trace/wire. We ought to account for anything that is in line with them though, to understand it fully.
Regards,
Chris
ghemink said:
Assuming the ferrite beads have 1uH inductance, with a 470uF cap, the resonance frequency woud be about 6Khz with a +3dB peak when the caps ESR is 0.05 Ohm. With an ESR of 0.1Ohm, the peak would be reduced to +1 dB (simulated this with switchcad as I'm too lazy to calculate). If those beads are 0.1uH, the whole picture becomes a lot rosier, but then suppression at 400kHz is much less, so likely those beads have an induictance slower to 1uH. If you go to an extreme of for example 2 1000uF 63V ZL caps with ESR of 0.021 Ohm each, the resonance frequency would go to about 3Khz with a nasty more than 6dB peak!!!
For a woofer amp maybe OK but probably not good for tweeters and midranges.
Gertjan
I'm ashamed to say I have yet to check it out under any scrutiny, but can I assume that the ~470uF cap actually forms one leg of the CLC pi filter? If so, perhaps it shouldn't? Also does the PSU happen to form the other leg?
Thanks
Thanks
ghemink said:
Sorry for all the typos 🙂
so likely those beads have an inductance closer to 1uH.
The effect of low ESR caps should be easily measurable on the power lines. Thinking of using for example a 1kHz square wave in an 8 or 4 Ohm dummy load, this should give ringing at a few kHz which allows us to estimate the inductance of the ferrite beads. I do not know the series resistance of those beads either, would have a large impact on the Q-factor of the LC network. With a lot of caps like 2x1500uF Rubycon ZL with ESR of 18mOhm, having a series resistance in the order of 20-30mOhm for the beads would very significantly lower the Q-factor.
So maybe power supply ringing in the audio frequency range is no issue at all. I guess measurements are needed here.
Gertjan
cap sound
While going to a bigger value of one particular brand of cap will "generally' give more body and bass it does not hold up the same when going to a different brand. One brand of caps 100uf might sound bassier and warmer than another brands 470uf. Sometimes the voltage rating of the cap changes the sound a lot! For instance I was trying a 470uf 10V Cerafine in a certain spot (5V rails) and when I went to the 470 uf 25V Cerafine it sounded not only warmer but much more real. However the 100uf Cerafine at 100V sounded VERY slow compared to a 220 uf 50V cap (47V rails/caps were same case size). So, it pays to experiment with different brands, voltages and capacitances. A small value Blackgate has tons of bass, whereas a much larger value Rubycon ZA sounds leaner. Also Blackgate FKs sound a lot bassier than N series. Then there is the difference in sound from damping....and removing the plastic around the cap (be careful when removing the plastic as the case will sometimes have DC on it!)...just had to throw in some Voodoo he he....always, lots to play with...the never ending story.
While going to a bigger value of one particular brand of cap will "generally' give more body and bass it does not hold up the same when going to a different brand. One brand of caps 100uf might sound bassier and warmer than another brands 470uf. Sometimes the voltage rating of the cap changes the sound a lot! For instance I was trying a 470uf 10V Cerafine in a certain spot (5V rails) and when I went to the 470 uf 25V Cerafine it sounded not only warmer but much more real. However the 100uf Cerafine at 100V sounded VERY slow compared to a 220 uf 50V cap (47V rails/caps were same case size). So, it pays to experiment with different brands, voltages and capacitances. A small value Blackgate has tons of bass, whereas a much larger value Rubycon ZA sounds leaner. Also Blackgate FKs sound a lot bassier than N series. Then there is the difference in sound from damping....and removing the plastic around the cap (be careful when removing the plastic as the case will sometimes have DC on it!)...just had to throw in some Voodoo he he....always, lots to play with...the never ending story.
Hi Rick,
Well, So what did you end up using for the power supply caps, not on the board? Or is that a secret too like what kind of solder do you use? I did ask you that on the AC, but you didn't respond.
My one big question to all of you is why must we use such a big transformer for a device that doesn't pull much current? I have heard that a big transformer because of the metal inside, tends to slow down the rate of current draw, versus the other way around. Now, I unfortunately do not have a source for that piece of info. I'll wait a bit and see how this thread ultimately pans out. It's getting pretty big, and would take days to paw through it.
So Rick, what front end chip are you using on the UCD amps?
Ray
Well, So what did you end up using for the power supply caps, not on the board? Or is that a secret too like what kind of solder do you use? I did ask you that on the AC, but you didn't respond.
My one big question to all of you is why must we use such a big transformer for a device that doesn't pull much current? I have heard that a big transformer because of the metal inside, tends to slow down the rate of current draw, versus the other way around. Now, I unfortunately do not have a source for that piece of info. I'll wait a bit and see how this thread ultimately pans out. It's getting pretty big, and would take days to paw through it.
So Rick, what front end chip are you using on the UCD amps?
Ray
some secrets are not secrets
Ray,
If you look on my website you will see that I use Nichicon caps (Gold Tune) as the "starter" cap bypassed by film caps and Jensen as the higher end option...most everyone is getting Jensens. Other stuff is secret. Secrets are fun. If I were not producing an amp in a competitive environment I would be babbling on for days here....telling all....I wish I could.
Ray,
If you look on my website you will see that I use Nichicon caps (Gold Tune) as the "starter" cap bypassed by film caps and Jensen as the higher end option...most everyone is getting Jensens. Other stuff is secret. Secrets are fun. If I were not producing an amp in a competitive environment I would be babbling on for days here....telling all....I wish I could.
- Status
- Not open for further replies.