Optimal supply design for UCD and Zappulse modules

[maxlorenz]

Hi Mike:

Yeah, I recently moved from a 100 year old victorian mansion that I had rewired totally. Had a few home runs for audio.

You're kidding! It should have had very good acoustics.

Also been trying to get a big 500VA+ R core, but I've found no off the shelf units.

Only 400VA and from far East:
http://eshop.diyclub.biz/product_info.php?cPath=152_74&products_id=60

I'll likely be talking about all this a year from now as "things to do".

At least you know how to do them wright!!
(Not my case)
Wait till you see my next speakers

Cheers.
Mauricio

 

[Portlandmike]

Hi Mike:

You're kidding! It should have had very good acoustics.


Only 400VA and from far East:
http://eshop.diyclub.biz/product_info.php?cPath=152_74&products_id=60


At least you know how to do them wright!!
(Not my case)
Wait till you see my next speakers

Cheers.
Mauricio

Acoustics were great in both my listening rooms. My living room was 32 x 16ft and opened into two other room making it basically 32x32. That, and it unloaded up a huge stare to the 2nd floor.

It was quite good, but it took quite abit of acoustic output to get it loud. Acoustics actually really sucked, but since that's were I make my living, I eventually got it right.

Best room was the one I rebuilt for audio. Imagine a standard 2x4 stud wall with 4x4 oak sistered to each stud. It was cross braced and rocked with 5/8inch rock, glued with mastic to the studs and 4x4's. That treatment was only done in the source half the room. The other end of the room was left standard stud wall on purpose, to give some losses. The room had amazing stage and punch!


I digress. I miss my house, sorry.

Best Regards,

Mike

 

[maxlorenz]

Acoustics actually really sucked, but since that's were I make my living, I eventually got it right.

You should see my all-concrete 3.3*6.6m living room
Bass traps were my entry to DIY

I had to build a 2.7*5*6.7 "multi-pourpose" room for my secondary system

You got mail

M.

 

[fram]

New UCD PSU prototype

Hi all,

Here is my new PSU cards (prototype) for UCD.
It is including T-network BHC caps to the main board and an auxiliary board as a small regulated +12V/-12V PSU for input buffer stage (need a very small transformer).

I will let you know my opinion about it in the next week.

An externally hosted image should be here but it was not working when we last tested it.

Franck

 

[maxlorenz]

OK, time to refloat this thread, that should be the most important thread for UCD amps, IMHO.

I had HF noise issues (see appendix (*)) with my UCD400ST monoblocks, wich use 40mF of Epcos Sikorel capacitance
Sikorel are fast and transparent but are probably not as good as Jensen "four pole" on HF.

Our dear Rha61 adviced me months ago about pi-filtering (pi-filter, what's that?) . We discussed previously this possibility here but I believe nobody actually tried it. Me neither... untill yesterday...

I am no expert on inductors as you shall see...
Rha61 recommended this configuration:
10mF//100uF->4uH->10mF, with a DCR of at least 0.5R for the inductor.
Noone adviced me about wich good inductor to employ , and, as I didn't want to spend lot of $ on Hi quality units, I decided to use my beloved 10uH shotgun ferrites (Epcos, N30 material) in series with 0.5R to imitate the "ideal" inductor. The 100uf cap is a common type electrolytic.

Well, I like the change so far...much cleaner highs wich led to more natural mids and good LF, without loss of dynamics. I even think that bass is better. This was preliminary impression with a couple of CDs. This is the best sound I got from my monoblocks to date.

What do you think?
(appart than advicing me to change hobby, I mean... )
Which inductors shall I pick up?

Thanks,
M

-----------------------------------------------------------------------------------
*I am very good at producing HF noise
While my M-Audio SuperDAC is at the ICU, i use a filterless TDA1543 DAC directly fed through I2S. With film cap or BG N type cap as output DC blocking cap, my system had no HF noise but led some "air and space" to be desired. I swaped them for a pair of BG N type, antiparalleled (sorry JosephK) and the soundstage expanded and much more detail and slam appeared...with also HF noise
Several tries to get rid of noise both on DAC and UCD amp power supply, failed. Particularly, "snubbering" UCD's PS did not succed as, while taking away some noise, they also made things slower and boring...
-----------------------------------------------------------------------------------

 

[classd4sure]

Hi,

Have you at all tried any level of mains input filtering .... even just a common mode choke? How about other common mode chokes on the signal inputs and perhaps more so, on the power rails.

There's alread a CLC filter on the UCD module I don't think I'd bother with adding another.

I have experienced the same sort of thing as you though. What I recently did seems to have improved things... just ensuring all wires are nicely twisted with as minimal loop area as possible.

Different 470uF decoupling caps can make a difference but as you've noticed anything you do that qwells the HF response will also attenuate the level of "air" which really makes for a natural sounding analytic type amp... which I prefer as well.

With that in mind it seems best to concentrate on reducing the generation of that noise at the source, rather than trying to filter it later.

However what I did recently got it about 8/10's of the way there.. for a stereo amp.

Comon mode chokes on the rails for each amp and maybe on the mains input may be worth trying at this point for me at least.

Cheers,
Chris

 

[Hans L]

Rha61 recommended this configuration:
10mF//100uF->4uH->10mF, with a DCR of at least 0.5R for the inductor.

Just a small note: uF equals mF. Since you're mixing them here, is that what you mean too?

Cheers,
Hans.

P.S.: hats off for your drive to experiment and contribute

 

[classd4sure]

mF is not = to uF

there's F, mF, uF, nF, pF

F and mF are for the sake of argument never used. 10 000uF as common in supplies is sometimes referred to as being 10mF though, which is still ok, in worst case you'll see people erroneously call it 10k which is absurd, but they mean 10k uF.

Anyway in this case he means 10 000uF//100uF.... which is different to say the least.

 

[Hans L]

mF is not = to uF

there's F, mF, uF, nF, pF

F and mF are for the sake of argument never used. 10 000uF as common in supplies is sometimes referred to as being 10mF though, which is still ok, in worst case you'll see people erroneously call it 10k which is absurd, but they mean 10k uF.

Anyway in this case he means 10 000uF//100uF.... which is different to say the least.

See for instance: http://www.justradios.com/MFMMFD.html

Short forms for micro farad include MFD, mfd, uF, MF and UF

Although m should stand for milli and u for micro, mF (or was it specifically 'MF' and 'mfd?) is used in different ways (US vs Euro or something?), which is why I checked to see what was meant. I normally avoid any possible confusion by expressing all cap values common to audio in uF, nF or pF...

 

[classd4sure]

See for instance: http://www.justradios.com/MFMMFD.html


Although m should stand for milli and u for micro, mF (or was it specifically 'MF' and 'mfd?) is used in different ways (US vs Euro or something?), which is why I checked to see what was meant. I normally avoid any possible confusion by expressing all cap values common to audio in uF, nF or pF...

Thanks for the link, I see your point. Those aren't naming conventions I've been faced with, or have ever seen yet, and I've been able to discern cap values alright for the most part. Could it be that it is "old news" of times long since past since it has to do with restoring old radios? "micromicrofarade"..... no thanks!

I could not agree with you more about sticking with more conventional naming.

 

[SSassen]

I've recently designed and built the following PSU for the v6.1 UcD400 which allows me to supply all parts of the module with its own supply. It uses a custom wound torodial transformer to supply all needed voltages.

It uses a double rectifier bridge using BYV79 diodes, four T-network capacitors, and has seperate sections for the supplying the opamp and the MOSFET driver. Naturally a amp-on circuit has been implemented as has a DC protection using two 16A relais.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

 

[ghemink]

I've recently designed and built the following PSU for the v6.1 UcD400 which allows me to supply all parts of the module with its own supply. It uses a custom wound torodial transformer to supply all needed voltages.

It uses a double rectifier bridge using BYV79 diodes, four T-network capacitors, and has seperate sections for the supplying the opamp and the MOSFET driver. Naturally a amp-on circuit has been implemented as has a DC protection using two 16A relais.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

Hi Sander,

That supply looks great, I guess many would be interested as an even higher end alternative to the Hypex HG supply. Do you sell it?

Gertjan

 

[SSassen]

Hi Gertjan,

I'm not really planning to, I had the same question asked in the Dutch DIY forum I frequently visit. I'm just showing off really

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

 

[maxlorenz]

Hi,
Thanks for your replies.
I didn't want to confound you with that "mF" unit. Chris is right, 10mF=10000uF... I promise I won't do it again.


Chris wrote:

Have you at all tried any level of mains input filtering

Yes, I use "balanced technical power" transformer, but now I moved it to my "secondary" system. I used to have also a choke+Xcap type filter. Both improved some aspects and make worse others...
I will try again snubbering the UCD transformer's secondaries to PS ground.

Thanks for your other tips.

Hans L wrote:

P.S.: hats off for your drive to experiment and contribute

You're welcome As you see, I only try to implement what was suggested to me as a quieter PS for this very sensitive amp.

My goal is to attract the experts to know if I'm on the right track (wich apparently is the case, based on sound) and get some tips about best inductor options, as there are many choices, sizes and prices. That is, before I buy a scope and spectrum analizer

Maybe a Black Gate for the 100uF cap is worthy? It's only money

Last night I listened again (very tired and sleepy, after a nightmare job-day). I found again a clearer presentation than before, with sweeter mid and highs. I think there's still HF noise (i am like a bat) but the mod shifted it to a higher level where it is better tolerated. What before sounded shhhh, now sounds as ssssss. Violin particularly, is enjoyable now. I have still a problem with the source, sorry.

Regards,
M

 

[ghemink]

Hi Gertjan,

I'm not really planning to, I had the same question asked in the Dutch DIY forum I frequently visit. I'm just showing off really

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

Maybe you can just offer the PCB?

Gertjan

 

[Bgt]

I think there's still HF noise (i am like a bat) but the mod shifted it to a higher level where it is better tolerated. What before sounded shhhh, now sounds as ssssss. Violin particularly, is enjoyable now. I have still a problem with the source, sorry.M

Max, you get hiss from your DAC??? This shouldn't be.
What speakers do you use? I mean are they over 100db efficiency??I use normal psu's and my amps. are dead quiet, also with my ear on top of the speakers. Ofcourse they are only 92db eff.
Bert

 

[deandob]

Sander,

If you could share the PCB that would be great. I have T-power caps and diodes for a double rectifier bridge but have not designed the PCB yet - if I could use your PCB as a base it will save me an hour or so PCB designing (not the most enjoyable part of DIY for me).

Regards,
Dean

 

[maxlorenz]

Bert wrote:

Max, you get hiss from your DAC??? This shouldn't be.

I'm affraid so...during a crisis of pride I sold all my commercial stuff and now I only have DIY sources ...so I have no "standard" to wich compare it. Anyway, appart HF noise, the rest of the frequencies AND soundstaging improved significantly with "pi-filter". Elegance...
Speakers are Tannoy Definition 500 (93db sens.). Dead quiet at idle.
Second system (active) with UCD180 to DIY Tannoy Autograph enclosures with Beyma coaxials , are 105db sens. for the tweeter and >100db for woofer. Reasonably quiet at idle, minimal HF noise when playing music through moded DDDAC1543. (Tweeter is -3db at 17Khz).

I am waiting to build our mate Ecdesigns' Ultimate DAC.

The other tweak I used recently (and that few are willing to try) is "technical balanced power" with a small EI-Tx for the transport. Very good

Thanks for your interest.
M

 

[ghemink]

Bert wrote:

I'm affraid so...during a crisis of pride I sold all my commercial stuff and now I only have DIY sources ...so I have no "standard" to wich compare it. Anyway, appart HF noise, the rest of the frequencies AND soundstaging improved significantly with "pi-filter". Elegance...
Speakers are Tannoy Definition 500 (93db sens.). Dead quiet at idle.
Second system (active) with UCD180 to DIY Tannoy Autograph enclosures with Beyma coaxials , are 105db sens. for the tweeter and >100db for woofer. Reasonably quiet at idle, minimal HF noise when playing music through moded DDDAC1543. (Tweeter is -3db at 17Khz).

I am waiting to build our mate Ecdesigns' Ultimate DAC.

The other tweak I used recently (and that few are willing to try) is "technical balanced power" with a small EI-Tx for the transport. Very good

Thanks for your interest.
M

I want to try balanced power, not only for the source but also for the amps. A bit difficult to find suitable high power balanced transformers. International shipping costs are very high. ALthough I guess there should be some Japanese suppliers, I have not found any yet.

Gertjan

 

[lucpes]

Anybody tried LM338 to build a regulated supply? 5A (7A peak) should be more than enough for UcD180 with 'normal' loudspeakers. In fact, you can parallel more to obtain a higher current capability.

An externally hosted image should be here but it was not working when we last tested it.

from http://myweb.tiscali.co.uk/nuukspot/decdun/gainclone6.html

(24V schematic in there can be adjusted for say +/-28V operation) (no, it ain't too small for most setups).

Paralleling voltage regulators (basicly adding a 5A schottky after each reg - should be other solutions, too):

An externally hosted image should be here but it was not working when we last tested it.

from http://www.dream-models.com/eco/electrics-bec1.html