ackcheng: I think SMPS can work fine in Audio amplifiers. The key is to maintain a low impedance out of the power supply. And you should also consider, that if you have a feedback regulated SMPS, the feedback loop's properties is affecting the sound. So i would make a non-feedback push pull type.
And why not use a huge oversized ferrite core, for example ETD59, which can comfortably handle 3-4 kVA. It still weighs just 300g plus a bit of copper.
And why not use a huge oversized ferrite core, for example ETD59, which can comfortably handle 3-4 kVA. It still weighs just 300g plus a bit of copper.
And here a new picture of a sample amplifier, built into a ATI research 738U.
As you can see it's very simple to build.
As you can see it's very simple to build.
Hi Lars,
Do my eyes deceive moi or.. are there a few extra trace areas next to either side of the LED on the left module?
Funny thing is, it looks as though they're floating because there's a shadow under them....?
Do my eyes deceive moi or.. are there a few extra trace areas next to either side of the LED on the left module?
Funny thing is, it looks as though they're floating because there's a shadow under them....?
thanks Lars,
Would you recommend your virtual 4 pole filter too?
http://www.lcaudio.com/index.php?page=45
Would you recommend your virtual 4 pole filter too?
http://www.lcaudio.com/index.php?page=45
Lars Clausen said:
Lars,
Forgive me in advance if I am ignorant, but is 20000uf really enough for two 200w modules?
I am guessing that because the PS provides limitless current 20000uf is OK?
Anyway, I am planning to pull the trigger on dual monos, just curious.
Like in many other amplifier types with variant power takeoff, like Class A/B, the size of capacitors in uF is not the most important figure.
More important is a good stiff transformer. Even if you are only taking off like 220 VA ( -> 200 Watts out) a 630 VA transformer can improve your midrange and top. (?). I can't give you a reasonable logic explanation, but that's just the way it is.
As for the capacitors it is important to use low impedance types (< 10 mOhms). You can hear the difference on caps, and the ones i use are fairly neutral, even if they are not the most expensive high end types.
If you want to build a cost-no-object amplifier, i can recommend Siemens/Epcos SIKOREL types. 10.000uF 63V will do nicely.
More important is a good stiff transformer. Even if you are only taking off like 220 VA ( -> 200 Watts out) a 630 VA transformer can improve your midrange and top. (?). I can't give you a reasonable logic explanation, but that's just the way it is.
As for the capacitors it is important to use low impedance types (< 10 mOhms). You can hear the difference on caps, and the ones i use are fairly neutral, even if they are not the most expensive high end types.
If you want to build a cost-no-object amplifier, i can recommend Siemens/Epcos SIKOREL types. 10.000uF 63V will do nicely.
classd4sure said:
C'mon Lars... what are you trying to hide? lol
They look like a pair of green ERO film caps lay down on my TFT🙂
The ones on the right are stood up, could be wrong though😀
I've totaly missed this thread until today, nice looking modules Lars!
Good call, they're the same green as the traces, thought never occured to me that they might be caps.
TFT?
TFT?
classd4sure said:
I can't see those caps on the photo's on Lars website, maybe they are just added decoupling?
I can't belive I missed this thread
TFT, just the pc monitor
Sorry to change the subject but seeing as though caps was mentioned on Lars last post, are the Sikoral caps as good or better than the 4pole T-nets?
classd4sure said:
Your probably right Chris!
I see Lars uses the LM6172's on input for these modules😎
....RS-components used to sell Siemens Sikorel caps. But , only to confuse us, manufacturers like to merge, or be subject to buy-outs
😉 Thus Siemens seems to be EPOCOS, and Swedish Rifa is now...?
Krummer Caps
Are these the same Mundorf MLytic?
Who designs the first Class D amp without any capacitor at all?😎
😉 Thus Siemens seems to be EPOCOS, and Swedish Rifa is now...?
Krummer Caps
Are these the same Mundorf MLytic?
Who designs the first Class D amp without any capacitor at all?😎
You mean EPCOS.
Rifa is BHC.
I read on this one site most caps are made by the very same 2 to 3 different manufacturers anyway.
At least this is the case for the Solen /Mundorf etc audiograde designer caps.
Might try and dig it up later so I can post it here.
I'll make you a class d amp with no caps, the real question is, would you really want it?
Rifa is BHC.
I read on this one site most caps are made by the very same 2 to 3 different manufacturers anyway.
At least this is the case for the Solen /Mundorf etc audiograde designer caps.
Might try and dig it up later so I can post it here.
I'll make you a class d amp with no caps, the real question is, would you really want it?
classd4sure said:
Or rather the other way around. BHC is part of Evox-Rifa since 2002.
http://corporate.evoxrifa.com/?sc=1856
And Evox-Rifa was a merger of.... I don't have to tell you that, right? 😉
Sorry, EPCOS of course.
But are there so much interest in high quality audiophil capacitors
or resistors ? Without much demand, profit may be low,and even
much regarded brands, often seems to use "off the shelf" components.
But the question remains, how much does investments in better
components pay off?
With SMD , it may be harder to evaluate this , as there are less options to choose from, and harder to swap components,or am I wrong? On the other hand , shorter signal paths and less parasitic capacitances, may compensate for this.
You can´t think of a discrete built CPU fast enough by todays standard , can you? And with more integration, we do-it-yourselfers may rightly feel a bit like outsiders?
But are there so much interest in high quality audiophil capacitors
or resistors ? Without much demand, profit may be low,and even
much regarded brands, often seems to use "off the shelf" components.
But the question remains, how much does investments in better
components pay off?
With SMD , it may be harder to evaluate this , as there are less options to choose from, and harder to swap components,or am I wrong? On the other hand , shorter signal paths and less parasitic capacitances, may compensate for this.
You can´t think of a discrete built CPU fast enough by todays standard , can you? And with more integration, we do-it-yourselfers may rightly feel a bit like outsiders?
- Status
- Not open for further replies.