UcD400 Q & A

Jump to Latest
B

Bruno Putzeys

Member
Joined 2004
hypex.nl
Pierre said:
Bruno, what can you consider a maximum safe absolute temperature, for example, if you measure with a sensor "glued" to the "T" heatsink near the mosfets? (I am talking about UcD400)

Best regards,
Pierre
Click to expand...

Depends what you call "safe". Here's the absolute worst case calculation.

The MOSFET chips have a maximum rated chip temperature of 175ºC. The thermal resitance to case is 1K/W. The spacer adds another 1K/W. At 175ºC junction temperature, Ron is max. 107mOhm. The overcurrent protection trips at around 20A. This is a peak power loss of 40W. During this peak (discounting thermal inertia, remember we're doing a worst-case calculation), the chip could become 86K hotter than the heat sink. 175ºC-86K=89ºC.
 
E

ewildgoose

Member
Joined 2004
Couple of quick questions now that I am building a new powersupply for my UCD400:

Capacitors: Any preference between BCC ALP22A and ALC10C, both at 10,000uF 63V. Both seem to be about the same price with similar specs. I seem to recall seeing that Hypex were using the ALC10C version in your PS though - any reason?

There will be 4x 10,000 per channel in this design. Sound OK? (Easy 8 ohm load)


Rectifiers. I'm going to be using a dual bridge rectifier board which takes discreet diodes. I have been recommended either HFA08TB60 or MUR 1560:

http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=273612&N=401
http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=930933&N=401

However, the MUR 1560 is much, much, less money. Are the "International Rectifiers" really that much better?

Scanning Farnell I also find a couple of other interesting diodes which are cheaper and have apparently better specs than the HFA above. How do I read the specs, would either of these diodes be suitable for a UCD supply:

MUR1520
http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=708306&N=401

RHRP860
http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=516491&N=401

Am I right in thinking that if I go for either MUR1560 or HFA08TB60 then I shouldn't use bypass caps on the diodes? The advice I have seen is that these fast recovery diodes are better without the caps...? I don't own the equipment to test both ways so looking for advice...

Thanks for any help on these basic questions

Ed
 
E

ewildgoose

Member
Joined 2004
MBR10100 looks great, thanks

By the way Timo, you say "no switching noise", but your design shows bypass caps on all your diodes? Does anyone have any comments on whether the MBR10100 should be bypassed or not?

Any thoughts on the difference between BCC ALP22A and ALC10C?


Different question: I am taking another UCD400 amp round to the local hifi circle meet this weekend. I have one of the early UCD400 boards with no coupling caps.... Do I risk blowing the amp up by letting people switch between it and another amp for A-B testing without switching it off first (and letting it power down)... I have destroyed a couple of Zappulse boards like this so I am very cautious now.

The guy holding the meet has measured his pre-amp DC out at 0.2mv, so I assume this won't cause any problems.

Advice appreciated

Thanks, Ed W
 
B

Bruno Putzeys

Member
Joined 2004
hypex.nl
ewildgoose said:
Different question: I am taking another UCD400 amp round to the local hifi circle meet this weekend. I have one of the early UCD400 boards with no coupling caps.... Do I risk blowing the amp up by letting people switch between it and another amp for A-B testing without switching it off first (and letting it power down)... I have destroyed a couple of Zappulse boards like this so I am very cautious now.
Click to expand...
UcD is unconditionally stable under no-load conditions, even when clipping. It's unnecessary to power it down to unplug the speakers but one would power it down before plugging/unplugging the inputs.

The oldest 400 modules already have overvoltage protection, so it won't break down when DC is applied to the input.
 
E

ewildgoose

Member
Joined 2004
As a workaround the guy who is hosting is hopefully going to make up some leads with blocking caps in them at the amp end. This way I hope that we can unplug at the pre-amp end without any problem.

So am I "safe" to avoid turn on and off if I add the blocking caps... I realise it's good advice to turn them off generally, but you understand the issues with A-B testing. Just wondering if I am missing anything other than the DC blocking being an issue with swapping input leads while the unit is active...?

Thanks

P.S. I digested several other threads about tweaking the UCD400, but was there a general consensus on good blocking caps? I am about to purchase some WIMA MKS2 2.2uF caps, which I hope aren't bad...? Any other suggestions or is it basically that any polyester cap is fair game?
 
B

Bruno Putzeys

Member
Joined 2004
hypex.nl
ewildgoose said:
As a workaround the guy who is hosting is hopefully going to make up some leads with blocking caps in them at the amp end. This way I hope that we can unplug at the pre-amp end without any problem.
Click to expand...
It's not DC that's the problem... The problem is that transformer leakage current (capacitive coupling) can transpire to an open circuit AC voltage between 2 amplifier chassis equal to the mains voltage (although half is more common). So when you unplug the inputs and the signal makes contact before the ground does, you can have 100VAC at the input of the amplifier.
ewildgoose said:
I digested several other threads about tweaking the UCD400, but was there a general consensus on good blocking caps? I am about to purchase some WIMA MKS2 2.2uF caps, which I hope aren't bad...? Any other suggestions or is it basically that any polyester cap is fair game?
Click to expand...
No consensus yet. Most mylar (polyester) caps are better than most electrolytics, but the variety in sonic signatures is as large among film caps as among electrolytics. I do think Wima is a safe place to start tho.
 
E

ewildgoose

Member
Joined 2004
Bruno Putzeys said:

It's not DC that's the problem... The problem is that transformer leakage current (capacitive coupling) can transpire to an open circuit AC voltage between 2 amplifier chassis equal to the mains voltage (although half is more common). So when you unplug the inputs and the signal makes contact before the ground does, you can have 100VAC at the input of the amplifier.
Click to expand...

Eeek...

However, the amp has a high input resistance, and also overload protection, so asking the dumb question: why is 100v input a problem.... (yeah, I know it's a stupid question, but I think you can see what I mean?)

Thanks

Ed

P.S. While I have distracted you with my dumb questions, do you have any comments on my question above over when it makes sense to add bypass caps on rectifier diodes like the MBR10100 I propose to use? I assume that most people use a scope and look for switching noise? Short of buying a scope are there any rules of thumb which I might want to follow in my situation?
 
B

Bruno Putzeys

Member
Joined 2004
hypex.nl
ewildgoose said:
However, the amp has a high input resistance, and also overload protection, so asking the dumb question: why is 100v input a problem.... (yeah, I know it's a stupid question, but I think you can see what I mean?)
Click to expand...

The op amp inputs might still not like being knocked into conduction like that. It's quite possible the amp still works after this, but that it'll sound worse or have developed a DC offset.

ewildgoose said:
While I have distracted you with my dumb questions, do you have any comments on my question above over when it makes sense to add bypass caps on rectifier diodes like the MBR10100 I propose to use? I assume that most people use a scope and look for switching noise? Short of buying a scope are there any rules of thumb which I might want to follow in my situation?
Click to expand...
Hmm. You could try hunting for switching noise using a SW radio receiver tuned to an empty spot near 30MHz.
MBR10100 does not need bypass caps in the strict sense because they don't have recovery problems. If you want to be really strict about potential resonances (potential as in: probably not actual), you could add across each diode a series network 10ohm+1nF. This damps any thinkable and unthinkable oscillation the schottkies might still be involved in.
 
E

ewildgoose

Member
Joined 2004
We had our shoot-out. Some initial thoughts in the UCD v Zappulse thread here:

http://www.diyaudio.com/forums/showthread.php?postid=670402#post670402

Consensus was that the UCD400 was slightly better than the Zappulse, but BOTH amps turned some pretty marginal speakers into something that a bunch of valve heads thought sounded exceptionally good. So high praise indeed to both amp designs.
 
J

jurjaan

Member
Joined 2005
ucd400/ma240

A simple question, which might be of interest to other members new to the subject. I am considering building an amp with UCD400 to drive my Quad ESL's. Does anyone see any objections? Someone tried this before?

I heard the Van Medevoort MA240 which uses a UCD chip as well although it seems to be a diffrent one. Sounded good, but the Quads could theoratically do with some more power. Maybe Bruno knows about the MA240 and the possible differences with, or advantages of, the UCD400.

Anybody in the neighbourhood of Utrecht (The Netherlands) who is willing to give his UCD400 a try on my Quad ESL's?

Thanks, Jurjaan
 
You must log in or register to reply here.
Top Bottom