Superlative Mono UcD400HG with 1kVA Toroidal

[gsattler]

Hello all,

I read through the power supply design guides on TNT Audio and Rod Elliot's websites, and began to think that I should use an Avel Lindberg 800VA or 1kVA transformer for each UcD400 module. This is to power a pair of cost no object UcD400 mono amps for my home stereo use. I realize that the guides on those websites were made for Class AB amps, but am trying to find a good reason to not spend the additional $10-20 for the increase in price over a 500VA toroid. Is there any real drawback from an electrical perspective over using too large of a transformer since I'm already planning on building a soft-start circuit? After doing lots of reading on these boards it seems that most people doing mono with the UcD400 end up with 2x500VA. Someone please give me some good reasons for not buying a larger transformer other than weight or cost. I will be building custom cases on my CNC at work, so size is also not an issue.

On a similar note but regarding capacitors, Bruno Putzeys recommends 2x10,000uf per UcD400, and some mono builders are doing 4x10,000uF per UcD400. Rod Elliot's guide has a capacitance calculator that yielded just over 84,000uF per mono amp with 1V P-P ripple voltage, 400wpc into 4 ohms with his recommended overhead and AC/DC conversion factors. I also recall reading that Jan-Peter and Bruno at Hypex thought that additional capacitance negatively affected the sound of the amp. I was planning on comparing Cornell Dublier 380LX to Panasonic TS-HA capacitors in an ABX test on my first build. Any opinions on the amount of capacitance that is required/useful or comments on my selection of capacitors?

[rha61]

here is a good link for capacitors comparison :

http://highend.puslapiai.lt/?tipstricks/7

i've used Sikorel for years and they sound great ( i've tested Elna Audio and Philips/BC )

[Corax]

Single-ended class-D power amplifiers tend to power-supply pumping effects. That means the reactive energie will flow back to the power-supply and will effect (i.e. raise) the rail-voltage(s).

I've experienced myself an over-/under-voltage lockout at my protection circuit and it disappeared after I increased the capacity of the bulk capacitors.
As a rule of thumb, the pumping effect is less 'threatening' (voltage overshot) with larger bulk capacitors.

After my modifications I could not (subjectively!) detect any negative effects on the sound. If this would be the same case with your UcD400 is something you have to find out yourself.

It's primarily also a question of money, because added bulk capacitors will definately get a bunch of bucks out of your wallet.

Take also care about the rectifier(s) if they are capable of handling such high capacitor values (keyword: peak forward current).

Some sort of inrush current limiters (or soft-start circuits) are also mandatory as you've already planned to include it in your design.
I use them always in particular when a toroid larger than 300VA is involved.

[gsattler]

rha61, were the capacitors in that test seamlessly ABX tested or AB tested after switching out the power supplies. The human brain has an incredible ability to experience changes it wants to or thinks are there. Numerous studies (http://www.thewinetrials.com/ is one great example) have shown that in blind, seamless AB tests that people have much less expensive tastes, and we convince ourselves that the money we spend on things will make them better. I will keep the classic audiophile recommendations about capacitors in mind as I go forward, but I will definitely do ABX testing to see if the extra cost produces any noticeable gains.

Corax, a soft-start circuit is definitely in the plans, as well as an dual bridge rectifier comprised of 25A soft recovery diodes. Pumping is indeed the primary reason why I am making these amps mono instead of stereo. The caps I have in mind are paralleled 80V 10,000uF Cornell Dublier 380LX capacitors that have excellent ripple current and ESR numbers even when compared to Epcos Sikorel. I am also planning on testing some Panasonic TS-HA 80V 10,000uF capacitors with similar ripple current and ESR numbers since they seem to be the cheapest "audiophile recommended" capacitors. As a reference, my most direct references of power supply capacitors in commercial gear is as follows:

My buddy has a Krell KSA-300s, 300W/8Ohm per channel, class A. That amp uses Vishay/Sprague 36DY capacitors, 4x68,000uF.

I have a Parasound Halo A23 125W/8ohm per channel, class AB. That amp uses Elna LP5 capacitors, 2x10,000uF.

Both of these capacitors are bulk grade high ripple current type capacitors, nothing special. Neither of these capacitor types are very expensive on the open market, particularly once the volume gets to the 50 or 100 piece quantity. If these amps sound as good as they do, then my theory is that much too much thought and concern goes into capacitor brand or model. In terms of electrical theory, any capacitor rated for use in high ripple current power supplies should perform excellently if the ESR & ESL is low. I hope to remain dubious until proven otherwise.

[Corax]

Since the pumping-effect mainly occures at low frequency output singnals (bass signals with high output values) there are methods to lower/suppress/avoid such issues.
In case it would be possible to synchronize the clock signals of the PWMs generated, which in turn is impossible for self-oscillating designs like the UcDs, there's a chance to use the reactive energy in the other output, as commonly done in BTL designs where the other tail uses this energy. By doing so the current will only flow through the load and the two tails of the bridge (proper supply-rail wiring assumed) but no additional current will be introduced to the rails, thus no voltage pumping will take place in such designs.

For selection of the correct/best capacitor, concerning the manufacturer, I agree almost completely, in particular what you've written in your last paragraph.
I doubt myself that all the other as 'mystic', 'magical' or 'the only perfect' published capacitors (like 'slit-foils' or 'T-network') are really worth what they promise. They only cost a hell and as long as I cannot hear the difference I stuck with the (older) traditional capacitors design. ;-)

[rha61]

if you feel happy with theory and specifications , that 's not a problem

twenty years ago , i've been very surprised by the big difference between two very good specifications amps
after that , i've experimented a lot , and electrolytic caps are one important reason of the solid state amps sound , and film caps dielectric have their own coloration ( the best are teflon ones)
no need of ABX test to hear that , you just need good speakers

[gsattler]

Yes, I can understand how you would say that a difference can be heard, but memory for sound is short, so without an ABX, can you really tell if the quality was actually different or you just thought it was different? My argument with non-immediate sequential listening tests spaced by the few minutes it takes to switch gear around is that there is no scientific method to the process. Were the amplifiers of the same circuit design? Were the speakers of the compared systems the same? Were the sources the same? Does the room have the same acoustic properties? There are so many factors at play here that you cannot know whether it is some other contributing factor that is the difference or if it is just your expectations playing games with your brain.

[rha61]

you dont need to spend time witèh ABX when the difference is big

about the amps , i simply replaced my usual amp with another in my system

[gsattler]

If you replaced your amp then one could argue that there are hundreds of significant differences between any two given amps. Were they of the same relative quality? The switching time required to change amps is still longer than a minute, and sound is quickly forgotten.

[classdphile]

Originally posted by gsattler
... and sound is quickly forgotten.

How does one manage to recognize their mother's voice, scent?