Hello all.
I just wanted to post my impressions on the bypass of the input caps in my UcD400 modules. Since I use a transformer based volume control passive pre, any DC that might be coming from my sources is suppressed (that was, of course, verified with a meter). I just removed the caps and jumped the connections (it was hard to find a piece of wire thin enough...). This is a must do tweak for anyone using these modules. As described by matjans, there's a definite veil removed from the sonic presentation. The sounds becomes more alive and dynamic. Lower level detail, overall transparency and imaging are well improved. I'd say all aspects of musical reproduction are improved. A good film cap should have similar effects, and as someone else had described, it doesn't have to be installed in the PCB itself, it can be placed in the input wires, or at the RCA or XLR inputs.
I've ordered a few Panasonic FC caps to replace the 2 big 470uF/100V caps. The FC caps have almost identical specs to the Vishay/BC Component 136 RVI series well praised by Lars from LC Audio (I haven't been able to find these for retail here in the US). The original caps in my modules are Nichicon VR series. They have decent specs, but the Panasonic and Vishay have much higher ripple current and lower impedance. While checking the UcD400 board, I noticed that these are bypassed by small surface mount caps, so I won't be adding an extra bypass for these.
I'll post my results once I install the caps. But I'd definitely recommend removing or replacing the input caps.
J.
I just wanted to post my impressions on the bypass of the input caps in my UcD400 modules. Since I use a transformer based volume control passive pre, any DC that might be coming from my sources is suppressed (that was, of course, verified with a meter). I just removed the caps and jumped the connections (it was hard to find a piece of wire thin enough...). This is a must do tweak for anyone using these modules. As described by matjans, there's a definite veil removed from the sonic presentation. The sounds becomes more alive and dynamic. Lower level detail, overall transparency and imaging are well improved. I'd say all aspects of musical reproduction are improved. A good film cap should have similar effects, and as someone else had described, it doesn't have to be installed in the PCB itself, it can be placed in the input wires, or at the RCA or XLR inputs.
I've ordered a few Panasonic FC caps to replace the 2 big 470uF/100V caps. The FC caps have almost identical specs to the Vishay/BC Component 136 RVI series well praised by Lars from LC Audio (I haven't been able to find these for retail here in the US). The original caps in my modules are Nichicon VR series. They have decent specs, but the Panasonic and Vishay have much higher ripple current and lower impedance. While checking the UcD400 board, I noticed that these are bypassed by small surface mount caps, so I won't be adding an extra bypass for these.
I'll post my results once I install the caps. But I'd definitely recommend removing or replacing the input caps.
J.
impressions
Yes, using electrolytics for signal coupling is not a good thing.
What transformer are you using? This could be the very best setup and one of the few setups that are safe to use DC coupling with. Going to DC coupling also lowered the noise floor and that is some of the benefit you are hearing.
I have used the FC caps and find them to work well and sound good. Should be a good match, let us know.
Roger
Julien_M said:
Yes, using electrolytics for signal coupling is not a good thing.
What transformer are you using? This could be the very best setup and one of the few setups that are safe to use DC coupling with. Going to DC coupling also lowered the noise floor and that is some of the benefit you are hearing.
I have used the FC caps and find them to work well and sound good. Should be a good match, let us know.
Roger
Did someone do blind ABX tests on the input caps issue, or is this just "percieved" as being better?
PS. Not trying to be picky, but there's always a bit of placebo effect. I found the UcD's to be better also, but I heard the difference with going back to regular amps. Didn't do blind tests either... Did hear things in recordings I couldn't hear with the old amp though.
Can't believe that removing the caps would be another large step forward.
PS. Not trying to be picky, but there's always a bit of placebo effect. I found the UcD's to be better also, but I heard the difference with going back to regular amps. Didn't do blind tests either... Did hear things in recordings I couldn't hear with the old amp though.
Can't believe that removing the caps would be another large step forward.
Re: impressions
I agree Roger, electrolytics are really bad in the signal path. But it seems that every piece of solid state audio gear has them somewhere in the signal...
My pre is a NOH kit from Bent Audio. It uses Stevens and Billington transformers.
Bent Audio site is: http://www.bentaudio.com/ , but I don't think they still offer the kit.
When I decided to go with a TVC pre, I was using Jensen IsoMax transformers between some components and noticed the sonic improvement they brought. But the Stevens and Billington transformers are even better. They're really transparent.
The way I wired my NOH, all the grounds are isolated between each other. That's a good way to know what's polluting the signal. My DVD player is a major offender, but it can't contaminate my other gear while not in use.
And, when I can, I float the components ground with a 10 ohm resistor with a small film cap and 2 diodes in parallel. That seems to eliminate any noise from the earth ground that may pollute the signal.
I'll post my findings when I get the FC caps, it should be tomorrow or Friday...
Julien
sx881663 said:
I agree Roger, electrolytics are really bad in the signal path. But it seems that every piece of solid state audio gear has them somewhere in the signal...
My pre is a NOH kit from Bent Audio. It uses Stevens and Billington transformers.
Bent Audio site is: http://www.bentaudio.com/ , but I don't think they still offer the kit.
When I decided to go with a TVC pre, I was using Jensen IsoMax transformers between some components and noticed the sonic improvement they brought. But the Stevens and Billington transformers are even better. They're really transparent.
The way I wired my NOH, all the grounds are isolated between each other. That's a good way to know what's polluting the signal. My DVD player is a major offender, but it can't contaminate my other gear while not in use.
And, when I can, I float the components ground with a 10 ohm resistor with a small film cap and 2 diodes in parallel. That seems to eliminate any noise from the earth ground that may pollute the signal.
I'll post my findings when I get the FC caps, it should be tomorrow or Friday...
Julien
yves,
i didn't do a blind abx test (lack of modules, I only have one stereo amp) but the difference is audible. I don't think it's only a placeo effect. I was with a friend when i removed (ypassed) the elco's in the signal path and when I returned home, a roommate asked what I did with my amps. He thought they sounded different.
The rest of this post falls under the "imho" thingy...
To my ears, the ucd's without the input caps have more clarity in the highs, ride cymbals really sound like ringing metal, especially when the drummer hits the inner part of the cymbals, they really go "tinnnnnnnnnnnnnggggggggg"; soundstage is more precise and the modules sound less "warm", less mid-bass whereas the deeper bass is more defined. Somehow the amp seems to be more "musical".
Has anybody found (better) replacements for the output cap that sort-of fit the pcb footprint?
i didn't do a blind abx test (lack of modules, I only have one stereo amp) but the difference is audible. I don't think it's only a placeo effect. I was with a friend when i removed (ypassed) the elco's in the signal path and when I returned home, a roommate asked what I did with my amps. He thought they sounded different.
The rest of this post falls under the "imho" thingy...
To my ears, the ucd's without the input caps have more clarity in the highs, ride cymbals really sound like ringing metal, especially when the drummer hits the inner part of the cymbals, they really go "tinnnnnnnnnnnnnggggggggg"; soundstage is more precise and the modules sound less "warm", less mid-bass whereas the deeper bass is more defined. Somehow the amp seems to be more "musical".
Has anybody found (better) replacements for the output cap that sort-of fit the pcb footprint?
matjans said:
Maybe it's me, but I don't understand why people are futzing around with the limited space on the UcD boards when they could be more easily merely shorting out the input caps and installing whatever they like for output caps on their preamps. Am I missing something here?
sx881663, unless I'm sorely mistaken, the big 470 uF 100 volt caps are actually supply bypass and therefore not directly in the signal path.
Francois.
ghemink said:
Hi Francois,
Yes, you are missing something
He was talking about the filter cap at the output of the UcD. (the 0.68uF cap)
Best regards
Gertjan
D'Oh!
I'd forgotten about the output filter cap, but my question still stands for the input ones. Julien_M is most of the way there with parking input caps at the input leads, though.
Thanks,
Francois.
Caps
Thanks guys!
Actually if you trace the signal path out the 470's are in the path but that was not what I was referring to. I think the reference was to the 22uf coupling caps. This thread is getting long enough we need to use quotes so we don’t lose track of what we are responding to.
About the .68uf output caps, stacked film caps are not very good unless they have a lot of bias voltage on them to hold them solidly together. You actually get a smearing effect due to mechanical motion storing and releasing energy. This is why they are ok for power supply bypass but not very good for signal coupling. A properly wound cylindrical cap is far superior in this regard. This phenomenon also directly relates to microphonics. A quick and dirty test will reveal much. Charge up a cap to a known voltage then put a DVM across it, squeeze it and watch the voltage change. This is a measure of its microphonic properties. As expended energy this can even be audible. I have heard cheap Zobel caps sing like a bird while doing a frequency sweep.
What I did on my modules was to mount .68UF Auricaps with as short of leads as possible right on the back of the fast on speaker terminals. This did result in a higher level of the 500 KHz but it was a very pure sine wave and poses no interference problems. I think the extra pc trace length used contributes as much as the cap leads. I will experiment more on this later. Results are what count and they sound wonderful!
Julien,
Sounds like a nice setup. Haven’t listened to a Bent setup personally but have heard a lot of good things about it. It always amazes me how audible removing noise is. Ground leakage and loops are the worst. Using signal isolation transformers can work wonders. We did a powered line array a few years back and it was unusable without transformers to tame the noise. At 112db efficiency only a couple of millivolts were way too loud. Anyway, this is one of the reasons I am liking SMPS type power a lot, far less line related problems.
Roger
Thanks guys!
Actually if you trace the signal path out the 470's are in the path but that was not what I was referring to. I think the reference was to the 22uf coupling caps. This thread is getting long enough we need to use quotes so we don’t lose track of what we are responding to.
About the .68uf output caps, stacked film caps are not very good unless they have a lot of bias voltage on them to hold them solidly together. You actually get a smearing effect due to mechanical motion storing and releasing energy. This is why they are ok for power supply bypass but not very good for signal coupling. A properly wound cylindrical cap is far superior in this regard. This phenomenon also directly relates to microphonics. A quick and dirty test will reveal much. Charge up a cap to a known voltage then put a DVM across it, squeeze it and watch the voltage change. This is a measure of its microphonic properties. As expended energy this can even be audible. I have heard cheap Zobel caps sing like a bird while doing a frequency sweep.
What I did on my modules was to mount .68UF Auricaps with as short of leads as possible right on the back of the fast on speaker terminals. This did result in a higher level of the 500 KHz but it was a very pure sine wave and poses no interference problems. I think the extra pc trace length used contributes as much as the cap leads. I will experiment more on this later. Results are what count and they sound wonderful!
Julien,
Sounds like a nice setup. Haven’t listened to a Bent setup personally but have heard a lot of good things about it. It always amazes me how audible removing noise is. Ground leakage and loops are the worst. Using signal isolation transformers can work wonders. We did a powered line array a few years back and it was unusable without transformers to tame the noise. At 112db efficiency only a couple of millivolts were way too loud. Anyway, this is one of the reasons I am liking SMPS type power a lot, far less line related problems.
Roger
matjans said:
The 470uF/100V Panasonic FC or Vishay 136 RVI will fit perfectly. They both have a diameter of 16mm witch is the same as the original caps. The difference is in their hight, the Panasonic is 40mm and the Vishay 35mm. They'll be taller than the heat transfer "T".
The Panasonic have a slightly higher ripple current rating but a slightly higher impedance than the Vishay.
I also ordered 680uF/100V FC caps, but the diameter is 18mm, it might be a tight fit but ripple current is 15% higher and impedance lower.
Will see how it goes...
matjans said:
Sorry matjans, I got confused. I haven't explored the output caps yet. But it looks like Roger had found good results with Auricaps.
Roger,
What's the voltage rating of the Auricaps you're using?
I wonder if Solen caps, or maybe Panasonic Polypropylene caps would be more compact and easier to implement.
Julien
Julien_M said:
The Auricap I used was the 450 volt. It is about the same size as their 1uf @ 200v. The Auricap has quite heavy leads for this size of cap and correspondingly lower ESR. Not sure how much that would affect results. My only recommendation would be to use a cylindrical type for its better properties. Also a more precision piece would be nice. Generally higher voltage caps sound better but are bigger. It is necessary to fit the right compromise to the actual usage. I haven’t done the experiments yet but am getting more modules next week and am going to do extensive testing. I will post my results good or bad.
Roger
Hi,
It's funny that this hot rodding thread is only thus far touching upon coupling caps.
Why look at the sine wave when it's the sound you should be concerned with, I'm not sure I'd even expect the look of the sine wave to change just by changing the coupling caps.
Someone should at least note that while removing or bypassing the coupling caps is an obvious and admitted improvement, they _are_ there for a reason, and that reason should be observed and respected, someone is likely to remove them expecting improved sound and blow their module..
Also this thread can never provide you or anyone with a "hotrodding" recipe as the final quality is a subjective experience..
With that in mind it may be more helpfull if more people posted what they have tried, and the results they've experienced, for example, instead of blindingly throwing the most expensive caps at it, find me a set of cheap caps that sound even better...
Regards,
Chris
Chris
It's funny that this hot rodding thread is only thus far touching upon coupling caps.
Why look at the sine wave when it's the sound you should be concerned with, I'm not sure I'd even expect the look of the sine wave to change just by changing the coupling caps.
Someone should at least note that while removing or bypassing the coupling caps is an obvious and admitted improvement, they _are_ there for a reason, and that reason should be observed and respected, someone is likely to remove them expecting improved sound and blow their module..
Also this thread can never provide you or anyone with a "hotrodding" recipe as the final quality is a subjective experience..
With that in mind it may be more helpfull if more people posted what they have tried, and the results they've experienced, for example, instead of blindingly throwing the most expensive caps at it, find me a set of cheap caps that sound even better...
Regards,
Chris
Chris
classd4sure said:
Chris,
You have brought up several good points. I have stated that jumpering out input caps can only be done safely under certain circumstances. Like with a transformer input or caps on the preamp output. I don’t even like this as the amp might be connected to different preamp. I have never recommended bypassing coupling caps as this splits the signal path and time smears it. It is almost never a sonic improvement. Bypassing power supply caps is a different story, multiple caps are recommended as this has low impedance for a broad range of frequencies. Btw, Auricaps are not that expensive and well worth their price.
I want to justify my choice here. I knew Bruno had stated that replacing the output filter cap could lead to problems but I felt it was worth the risk. This cap has a great deal of apparent power across it and any deficiencies will turn that into real power as heat or mechanical energy. Due to the Auricaps superior design and construction I knew it to be much less of a risk of developing nonlinearities. With nearly perfect cap behavior the output residual is very nearly pure sine wave and poses no problem even though it is higher in amplitude due to inductive interaction. Sonically it does work as long as you have room for it.
Arbitrarily replacing the filter caps can lead to ringing due to interaction with the combination of high frequency switching, low ESR and the inductance of the pc traces. Possibility some beefing up of the traces could be done and then a low ESR cap could be a real improvement. I wouldn’t feel right doing this without bouncing it off of Bruno first.
That is the problem with almost any changes, everything interacts with everything else. All these changes will have to be tried and tested carefully. Like the good physician, “first do no harm”.
I am the first to admit I don’t have all the answers and have a lot to learn. Class D is brand new to me and I am looking for answers too. I have had a bit of analog experience to draw on and this helps.
Roger
sx881663 said:
Hi Roger,
Do you plan to check power supply ringing with a scope once you start fiddling with the 470uF caps? What I thought to have understand from Bruno/Jan-Peter comments is that you better stay away from the 10uF caps as these are choosen as to suppress ringing. The 470uF caps could however be replaced with lower series resistance versions. Does anybody know the value of that resistor between the 470uF caps and the 10uF caps? It says R005, does that mean 0.005 Ohm, really that low?? (would be great if it was that low). I can`t measure such low R with my DMM, can`t even really measure 0.1 Ohm, so can`t easily figure out whether it is 0.05 or 0.005 Ohm.
Best regards
Gertjan
Hi Roger,
Do you plan to check power supply ringing with a scope once you start fiddling with the 470uF caps? What I thought to have understand from Bruno/Jan-Peter comments is that you better stay away from the 10uF caps as these are choosen as to suppress ringing. The 470uF caps could however be replaced with lower series resistance versions. Does anybody know the value of that resistor between the 470uF caps and the 10uF caps? It says R005, does that mean 0.005 Ohm, really that low?? (would be great if it was that low). I can`t measure such low R with my DMM, can`t even really measure 0.1 Ohm, so can`t easily figure out whether it is 0.05 or 0.005 Ohm.
Best regards
Gertjan
The only way to measure resistance that low is the 4 wire method. This involves putting a known current through the res. and measuring the voltage drop right on the part with separate leads. I have used this method to measure things like inductors with only a few turns of wire. May have to go up to a full amp to get a reasonable reading but that represents almost no power with a few millivolts of drop.
Now that you mentioned it I do recall Bruno talking about the 10uf caps. Yes I would use my 350MHz scope to check for ringing. I am going to look at NIC, Nichicon and a few worthy other caps. I have no problem with a taller cap so will probably max it out.
Roger
Do you plan to check power supply ringing with a scope once you start fiddling with the 470uF caps? What I thought to have understand from Bruno/Jan-Peter comments is that you better stay away from the 10uF caps as these are choosen as to suppress ringing. The 470uF caps could however be replaced with lower series resistance versions. Does anybody know the value of that resistor between the 470uF caps and the 10uF caps? It says R005, does that mean 0.005 Ohm, really that low?? (would be great if it was that low). I can`t measure such low R with my DMM, can`t even really measure 0.1 Ohm, so can`t easily figure out whether it is 0.05 or 0.005 Ohm.
Best regards
Gertjan
The only way to measure resistance that low is the 4 wire method. This involves putting a known current through the res. and measuring the voltage drop right on the part with separate leads. I have used this method to measure things like inductors with only a few turns of wire. May have to go up to a full amp to get a reasonable reading but that represents almost no power with a few millivolts of drop.
Now that you mentioned it I do recall Bruno talking about the 10uf caps. Yes I would use my 350MHz scope to check for ringing. I am going to look at NIC, Nichicon and a few worthy other caps. I have no problem with a taller cap so will probably max it out.
Roger
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.