Big Caps are useless with smps?
Hi Alkasar, thanks for the input. Some questions for you:
When you say bigger caps are not better with the smps, could you tell us why?
You say they can harm, what do you mean? Is it due to possible damage to smps?
Let's address both issues:
#1 is about sound quality
#2 is about equipment function and reliability
I ask because I DID run A/B tests with caps on smps, and my experience was that the additional capacitance most certainly DID sound better. With Scott Sendler's reports, I'm now wondering how much the relative quality of the smps affects this.
With the smps we used, the larger caps certainly DID make a notable and beneficial sonic difference. The variable in this may be that the Meanwell is good enough to be unaffected by extra capacitance? Now you and Sendler have gotten me curious.. which means I'll have to test it...
Now, about the A/B of Virtue vs hifimediy... again, the PS issue.
I think it should be made clear that with these Tripath based amps the quality of the PS is CRITICAL... you will clearly hear differences between them ... which means:
IF you were listening to the Virtue One with that crappy little supply, it was not a valid test of the amp itself! You were hearing the crappy little supply!
Virtue is pretty up-front about telling people the 65W supply is really only good for break-in, and they sell other power supplies as recommended options.
You ask about the logic: The logic in this is they give their clients valid options, either a more robust smps OR the battery pack without making them pay for a big smps first and THEN proposing a batt-pack.
The 65W is basically a $2 throw-away, not intended for serious listening. And they make it clear this is the case.
A more valid way to compare these amps is to use the best power supply you can find and use it to power each amp under test. A comparison using one amp with a lousy supply vs a similar amp with a really good PS is mostly going to show the difference between the supplies!
In the past, I used Virtue's 30V/4A 130W smps. Sounded pretty good, but when I phased over to batteries, first SLA and then the RedWine LFP supplies, the sound got dramatically better, easily besting several megabuck amplifiers, whether of tube, mosfet or bi-polar topologies.
From this I had concluded that good batt-packs out-perform smps. However, after all the good reports I've seen here about the Meanwell, I think I'll order one and compare vs the batt-packs... I could always re-sell it, or keep it for other projects. I'll report the results here on this forum.
My objective in comparing these amps and power supplies is to determine which one has the best potential to go into my self-powered high-end speakers. (Check HolisticAudio.com)
For me cost is less of an issue, it's all about the performance. Tripath based amps really are the best I've heard so far, excepting a $35K custom-made Cary SET using special 12" tall radio tubes. And it was pretty close! The difference was not so big anyway!
I'll keep that $34,500 difference and use it somewhere else..
So far, have tested stock Sure vs Helder vs Virtue....
I have a hifimediy on order and will soon be able to compare for myself.
Also, Alkasar, in your A/B tests, were you using a modded hifimediy or stock? and what mods you have put on it?
For my speakers it would be nice if the stock or modded hifimediy outperforms the Virtue. So far, in the evaluations neither the stock Sure nor the stock Helder have come close to equaling the Virtue... no surprise there, the parts are not equal!
Oddly enough, although the Helder looks much better built than the Sure, in stock form the Sure is closer to the Virtue's sonics. And enen though the Helder is twice the price of the Sure, the Sure runs VERY hot and makes me fret about long-term reliability.
"the bigger power cap the better" is true with linear supply. Not with SMPS.
Big caps are useles with SMPS. Too big caps can harm.
Vitueone comes with huge power caps and ridiculous smps (65W/24V). What is the logic here ?
An A/B comparison of vitueone and a hifimediy T1 + meanwell shows a clear winner, for a fraction of the price
Hi Alkasar, thanks for the input. Some questions for you:
When you say bigger caps are not better with the smps, could you tell us why?
You say they can harm, what do you mean? Is it due to possible damage to smps?
Let's address both issues:
#1 is about sound quality
#2 is about equipment function and reliability
I ask because I DID run A/B tests with caps on smps, and my experience was that the additional capacitance most certainly DID sound better. With Scott Sendler's reports, I'm now wondering how much the relative quality of the smps affects this.
With the smps we used, the larger caps certainly DID make a notable and beneficial sonic difference. The variable in this may be that the Meanwell is good enough to be unaffected by extra capacitance? Now you and Sendler have gotten me curious.. which means I'll have to test it...
Now, about the A/B of Virtue vs hifimediy... again, the PS issue.
I think it should be made clear that with these Tripath based amps the quality of the PS is CRITICAL... you will clearly hear differences between them ... which means:
IF you were listening to the Virtue One with that crappy little supply, it was not a valid test of the amp itself! You were hearing the crappy little supply!
Virtue is pretty up-front about telling people the 65W supply is really only good for break-in, and they sell other power supplies as recommended options.
You ask about the logic: The logic in this is they give their clients valid options, either a more robust smps OR the battery pack without making them pay for a big smps first and THEN proposing a batt-pack.
The 65W is basically a $2 throw-away, not intended for serious listening. And they make it clear this is the case.
A more valid way to compare these amps is to use the best power supply you can find and use it to power each amp under test. A comparison using one amp with a lousy supply vs a similar amp with a really good PS is mostly going to show the difference between the supplies!
In the past, I used Virtue's 30V/4A 130W smps. Sounded pretty good, but when I phased over to batteries, first SLA and then the RedWine LFP supplies, the sound got dramatically better, easily besting several megabuck amplifiers, whether of tube, mosfet or bi-polar topologies.
From this I had concluded that good batt-packs out-perform smps. However, after all the good reports I've seen here about the Meanwell, I think I'll order one and compare vs the batt-packs... I could always re-sell it, or keep it for other projects. I'll report the results here on this forum.
My objective in comparing these amps and power supplies is to determine which one has the best potential to go into my self-powered high-end speakers. (Check HolisticAudio.com)
For me cost is less of an issue, it's all about the performance. Tripath based amps really are the best I've heard so far, excepting a $35K custom-made Cary SET using special 12" tall radio tubes. And it was pretty close! The difference was not so big anyway!
I'll keep that $34,500 difference and use it somewhere else..
So far, have tested stock Sure vs Helder vs Virtue....
I have a hifimediy on order and will soon be able to compare for myself.
Also, Alkasar, in your A/B tests, were you using a modded hifimediy or stock? and what mods you have put on it?
For my speakers it would be nice if the stock or modded hifimediy outperforms the Virtue. So far, in the evaluations neither the stock Sure nor the stock Helder have come close to equaling the Virtue... no surprise there, the parts are not equal!
Oddly enough, although the Helder looks much better built than the Sure, in stock form the Sure is closer to the Virtue's sonics. And enen though the Helder is twice the price of the Sure, the Sure runs VERY hot and makes me fret about long-term reliability.
Hi Jack, long post
I went straight to the point because it was adressed earlier in this topic.
lets explain. In linear supply, power caps have two roles : filtering + energy tank. Caps are refilled at twice the mains power, ie 100Hz. They have to hold energy for the amp for about 8ms before they get recharged. So the bigger the better.
An SMPS like the small bricks or the meanwell s-350 works at 25kHz. More modern SMPS like the Connexelectronic or better meanwells work a 100+ kHz. The amp power caps filtering role is obviously gone and the "refilling" is done 1000+ faster. No need fo big cap anymore. To give an example : I have seen 300W poweramps powered by a connexelectronic SMPS with 2x470µF. They work very well
Why can big cap harm ? the SMPS are usually regulated (not all are). Means it adjusts the current to its "consumeer" instantaneous need to keep stable voltage. A pwer cap acts as an "intermediary" between the SMPS and its real consumer, the amp.
The power cap hides the real consumer instaneaous current need.
The bigger the cap, the more it hides.
Worst case, the smps regulation will work sub-optimal.
Worse worse case, the SMPS and cap start oscillating and you lose a lot of power or destroy regulation. Depending on the smps, it requires a very big capacitance to reach this worse worse case, theoretically possible.
I went straight to the point because it was adressed earlier in this topic.
lets explain. In linear supply, power caps have two roles : filtering + energy tank. Caps are refilled at twice the mains power, ie 100Hz. They have to hold energy for the amp for about 8ms before they get recharged. So the bigger the better.
An SMPS like the small bricks or the meanwell s-350 works at 25kHz. More modern SMPS like the Connexelectronic or better meanwells work a 100+ kHz. The amp power caps filtering role is obviously gone and the "refilling" is done 1000+ faster. No need fo big cap anymore. To give an example : I have seen 300W poweramps powered by a connexelectronic SMPS with 2x470µF. They work very well
Why can big cap harm ? the SMPS are usually regulated (not all are). Means it adjusts the current to its "consumeer" instantaneous need to keep stable voltage. A pwer cap acts as an "intermediary" between the SMPS and its real consumer, the amp.
The power cap hides the real consumer instaneaous current need.
The bigger the cap, the more it hides.
Worst case, the smps regulation will work sub-optimal.
Worse worse case, the SMPS and cap start oscillating and you lose a lot of power or destroy regulation. Depending on the smps, it requires a very big capacitance to reach this worse worse case, theoretically possible.
the Virtue versus Hifimediy comparison was :
Hifimediy T1 ALPS + meanwell S-350, power supply set at 30V. No tweak. This version had a 4700µF power cap. cost <$100
versus
Virtue One.2 + 24V 130W "power brick" (not the basic 65W power supply). No tweak. $350
Both amps are not mine and it was not my system.
We paid attention to adjust levels carefully. At low volume, no difference could be made between the two amps. When volume pot reached half-way, the difference became obvious in favour of T1. Power was much better managed by T1 and the highs were cleaner.
I have my personal ideas to correlate what our ears heard with the amp design. Listed here with no order, i can't tell which comes first :
1. design of both amp is the same. No suprise that they sound similar.
2. 130W is not sufficient power supply margin. The big caps in the virue one may hide a bit, but have limitation.
3. power a tk2050 at 24V you don't get the most out of it. Distorsion comes earlier.
4. the STA-508 power chip in T1 performs better than the regular TP2050 (used in the virtue and most other tk2050 amps)
The meanwell s-350 is by far not the best SMPS out there, but it does the job for little $$. We could not test the meanwell on the virtue because lack of connector ! I am sure this would have narrowed the differences.
To me, an easy upgrade to the virtue is to use a 30V more powerfull regulated SMPS instead of the stock 24V brick.
I am also sure that the T1-510A version, using STA-510A power chip, would have made the difference even more obvious.
I have the pleasure to own a custom made T2 with 2x 510A power chip, and it performs much better than the regular T2
Hifimediy T1 ALPS + meanwell S-350, power supply set at 30V. No tweak. This version had a 4700µF power cap. cost <$100
versus
Virtue One.2 + 24V 130W "power brick" (not the basic 65W power supply). No tweak. $350
Both amps are not mine and it was not my system.
We paid attention to adjust levels carefully. At low volume, no difference could be made between the two amps. When volume pot reached half-way, the difference became obvious in favour of T1. Power was much better managed by T1 and the highs were cleaner.
I have my personal ideas to correlate what our ears heard with the amp design. Listed here with no order, i can't tell which comes first :
1. design of both amp is the same. No suprise that they sound similar.
2. 130W is not sufficient power supply margin. The big caps in the virue one may hide a bit, but have limitation.
3. power a tk2050 at 24V you don't get the most out of it. Distorsion comes earlier.
4. the STA-508 power chip in T1 performs better than the regular TP2050 (used in the virtue and most other tk2050 amps)
The meanwell s-350 is by far not the best SMPS out there, but it does the job for little $$. We could not test the meanwell on the virtue because lack of connector ! I am sure this would have narrowed the differences.
To me, an easy upgrade to the virtue is to use a 30V more powerfull regulated SMPS instead of the stock 24V brick.
I am also sure that the T1-510A version, using STA-510A power chip, would have made the difference even more obvious.
I have the pleasure to own a custom made T2 with 2x 510A power chip, and it performs much better than the regular T2
First many thanks for your explanation regarding the caps and their optimum purposes. It is no wonder that in the end the quality of powersupply management decides about the performance - preconditioned that we have similar amp designs.
Though that I find the virtue amps are still a bang for the buck when they combined with better powersupplies. Because they are finished in cases, look pretty good and you can get them in different colors. A real good alternative for people who don't want anything to do with DIY. Mine virtue 2 classic also increased unbelievable in performance with a more current PS.
In my case with the virtue 2 I made especially in lower levels the finding that the amp sounded more relaxed and well defined in bass. Maybe the bigger 20000 > cap is responsible ? I cannot compare now with the T2 because I sold the virtue 2 months ago (including the standard factory SMPS
). Of course not because of the sound quality - I am more in a DIY trip nowadays and own different Tripath boards to compare.I have the pleasure to own a custom made T2 with 2x 510A power chip, and it performs much better than the regular T2
Is it just only the change of powerchip or maybe more ???
I myself want to experiment also with other caps ...
Originally Posted by alkasar View Post
>I have the pleasure to own a custom made T2 with 2x 510A power chip, >and it performs much better than the regular T2
Is it just only the change of powerchip or maybe more ???
I myself want to experiment also with other caps ...
Change input caps is the "usual" tweak
If you read french, I can point to a forum thread where people tried many different options, incluidng russian k75-10 and Obligato copper caps. It's a matter of preference in the end T2 with 2x510A sounds definitively better because the power chip is better. It's the same board, although some components values on the board need to be adjusted. I can now push up the volume much further before distorsion makes the highs crisp.
T2 510A can be powered up to 40V and then deliver 2x200W in 4 ohms. I have not tried yet more than 36V as I wait for a new power supply.
I needed a T2 architecture. As my speakers are 4 ohms, a T1-510A or a T4 are not a good fit. T2-510A really is
Al
How to get T2 with 2x 510A?
Did you get the T2-510A from hifimediy? If so, how did you get them to make it for you, can we do the same? Or did you mod them yourself for this change in chips?
It sounds like 2x510A would be ideal for me as well, my speakers are also 4 ohm and they really do sound better if the amp can drive massive current.
It's really too bad you were not able to compare with the same power supplies, I agree with your comment about the "larger" Virtue brick not having enough power. The batteries really did sound so much better it made me wonder about smps in general.
So yes, now I think I'll try a Meanwell and compare to the SLA and RedWine Black Lightning batt-packs. I've seen the 25kHz versions on ebay, but where do you get the 100kHz type?
Hi Alkasar, thanks for the further comments. It turns out I do read French so, yes please could you send a link to the french site?Change input caps is the "usual" tweak
If you read french, I can point to a forum thread where people tried many different options, incluidng russian k75-10 and Obligato copper caps.
T2 with 2x510A sounds definitively better because the power chip is better. It's the same board, although some components values on the board need to be adjusted.
T2 510A can be powered up to 40V and then deliver 2x200W in 4 ohms.
I needed a T2 architecture. As my speakers are 4 ohms, a T1-510A or a T4 are not a good fit. T2-510A really is
Al
Did you get the T2-510A from hifimediy? If so, how did you get them to make it for you, can we do the same? Or did you mod them yourself for this change in chips?
It sounds like 2x510A would be ideal for me as well, my speakers are also 4 ohm and they really do sound better if the amp can drive massive current.
It's really too bad you were not able to compare with the same power supplies, I agree with your comment about the "larger" Virtue brick not having enough power. The batteries really did sound so much better it made me wonder about smps in general.
So yes, now I think I'll try a Meanwell and compare to the SLA and RedWine Black Lightning batt-packs. I've seen the 25kHz versions on ebay, but where do you get the 100kHz type?
Hello,
I am affraid no.
Some components will have to be replaced for the higher voltage and gain setting, but more important : 517B dissipates more heat and needs larger heatsink.
If you need little more power, you could go with 510A and match the T1-510A board. Still some components to replace, but easier.
Bonjour
here we go : Tweaks modules tripath hifimediy tk2050 - TDG
First post has links to the useful posts.
After some discussions about the changes to make, Hifimediy built it for me. Obviously he can redo
I run it daily for over a month now, it's very stable. Powered at 36V it delivers significant more power than T2-505 (and, in my opinion, sounds even better)
As I said, I was not yet able to power it at 39V as expected. I wait for a new psu.
The only drawback is the heat. 2x510A at 36V get hot. I wonder at 39V. Unfortunately the board is quite crowded and I can not see a way to replace heatsink by a larger one, and I don't want to add a fan. I may have to limit voltage due to heatsink.
BTW, Connexelectronic just added on his website a 300W smps single voltage. High quality as usual, affordable, low profile 100x100x38mm, available at many voltages : a perfect fit for theses amps
Connexelectronic
Hi,
Which components need to be changed, are these only capacitors on the psu line? Why has to be gain changed, have the 517B and 508 different input sensitivity?
Regarding heat dissipation I won't have a problem as the board will dissipate heat through aluminum block to the chassis.
not only power supply line but also output filter caps & resistors. Also a couple of other resistor values need to be adjusted : the 5V regulator, led and I may forget about few others.
All the amps using tc2000 input chip have the same input sensitivity However, modulator feedback & gain should be adjusted according to power voltage. tk2050 datasheet explains page 12 http://www.datasheetcatalog.org/datasheet/Tripath/mXyzwqtx.pdf . Typically increase Rfbc with higher voltage, which increases also gain.
interesting. how do you do that ? can you show a picture ?
Thank you very much for the explanation. So changing to STA510A can be much more easily made.
I can't show you the picture as I haven't done this yet. But this aluminum block would be the same size as it is now heatsink. It would be bolted from bottom side to PCB and on the top side to chassis or something that has bigger heat dissipation.
mravlca said:
alkasar said:
I can't show you the picture as I haven't done this yet. But this aluminum block would be the same size as it is now heatsink. It would be bolted from bottom side to PCB and on the top side to chassis or something that has bigger heat dissipation.
What values?
Could you share the values you specified for these components? Since the 2x510a you got is stable, I'd be interested in getting something similar. I can probably figure decent values to work with from the spec sheet, but it would be most useful to start with values that have already proven to work well.
Also, is the 510a functionally equivalent to the TK20250, same gain settings, etc?
Thanks!
Could you share the values you specified for these components? Since the 2x510a you got is stable, I'd be interested in getting something similar. I can probably figure decent values to work with from the spec sheet, but it would be most useful to start with values that have already proven to work well.
Also, is the 510a functionally equivalent to the TK20250, same gain settings, etc?
Thanks!
I am also sure that the T1-510A version, using STA-510A power chip, would have made the difference even more obvious.
I have the pleasure to own a custom made T2 with 2x 510A power chip, and it performs much better than the regular T2
Hi Alkasar, in looking over your comments about the paralleled 510a, I noticed that hifimediy have released a T3 using paralleled 517B output sections.
Have you had a chance to compare between 510A and 517B?
Although I am aware of the 3020, I am trying to evaluate only amps that have the 2000 type controller in them, as Vinnie at RedWine Audio told me that it is the best sounding of the Tripath controllers and his amps are among the best ever, so I trust his ears and acumen. Big $$$ but good sound!
So my thinking is maybe to jump to the T3 from the beginning...
hello,
@liberas : I have a Valab23 stepped attenuator with smd resistors (log, 50K) excellent quality. good fit with T2
@erozsolt : in my optinion, go for T1-510A with associated power supply.
@jack :
yes T3 has paralleled sta517B. its a mono design. Bernie7 on this forum built an amp with two of these boards.
It has plenty of power
I did not need that much power and wanted a smal design, 4ohm load capable, with single smps. Thus the T2-510A custom build.
I never listened to a 517B board. I would love to compare it with a ta3020 amp as they play in the same league regarding power
I have listened to other popular tripaths ta2020 and 2022 and I agree that tc2000 based amps have best sound. (Ta2020 is not bad at all, but really lacks power). TA3020 is supposed to have the next evolution of tc2000 inside (a modded tc2001 actually). It has very good reviews actually. It may be my next try
T2-510A board :
We agreed with hifimediy on the feedback resistor Rfbc adjustment. This is R12/13/14/15 on the board. They are now 19.2k instead of 14k. With this change, T2-510A gain is 20v/v instead of 15 before.
Hifimedy made other housekeeping changes (resistor on the 7805 regulator, relays, capacitors voltage, etc.) I can't list for sure all the changes he made but he knows.
@liberas : I have a Valab23 stepped attenuator with smd resistors (log, 50K) excellent quality. good fit with T2
@erozsolt : in my optinion, go for T1-510A with associated power supply.
@jack :
yes T3 has paralleled sta517B. its a mono design. Bernie7 on this forum built an amp with two of these boards.
It has plenty of power
I did not need that much power and wanted a smal design, 4ohm load capable, with single smps. Thus the T2-510A custom build.
I never listened to a 517B board. I would love to compare it with a ta3020 amp as they play in the same league regarding power
I have listened to other popular tripaths ta2020 and 2022 and I agree that tc2000 based amps have best sound. (Ta2020 is not bad at all, but really lacks power). TA3020 is supposed to have the next evolution of tc2000 inside (a modded tc2001 actually). It has very good reviews actually. It may be my next try
T2-510A board :
We agreed with hifimediy on the feedback resistor Rfbc adjustment. This is R12/13/14/15 on the board. They are now 19.2k instead of 14k. With this change, T2-510A gain is 20v/v instead of 15 before.
Hifimedy made other housekeeping changes (resistor on the 7805 regulator, relays, capacitors voltage, etc.) I can't list for sure all the changes he made but he knows
.Does anyone have any experience with the SMPS300R from Connexelectronic? Or generally with any SMPS from Connexelectronic? Would you recomment it over the Meanwell one?
Do I need anything else then one board of ALPS-T1-T510A and one PSU for example from Meanwell or Connexelectronic? I mean a case and some plugs of course, but anything more?
@alkasar: Thanks!
Do I need anything else then one board of ALPS-T1-T510A and one PSU for example from Meanwell or Connexelectronic? I mean a case and some plugs of course, but anything more?
@alkasar: Thanks!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.