advertisement


Howto: TeddyCap

teddy_pardo

Trade: Teddy Pardo
For those of you who bought the TeddyReg boards and would like to build a full power supply I recommend using the following circuit for the smoothing capacitors between the rectification diodes (or bridge) and the TeddyRegs:


HTML:
----| R1 |----------| R2 |----------| R3 |----------
              |               |               |
             ---             ---             ---
 in           C1              C2              C3           out (TeddyRegs)
             ---             ---             ---
              |               |               |
 ---------------------------------------------------

R1, R2, and R3 are 0.47R 1W resistors
C1, C2, and C3 are 2200-4700uF capacitors

This configuration will smooth the noisy ripple coming from the rectification diodes and will give you best performance with the TeddyRegs.

One such circuit can feed several TeddyRegs.

Cheers, Teddy
 
Teddy, thanks ... your contribution to this place (along with others) is v. much appreciated.

Can you pls explain why you choose to do it this way instead of use a nice big fat 15000uf cap?

Uz
 
Can you pls explain why you choose to do it this way instead of use a nice big fat 15000uf cap?

Uz

Here is how the ripple looks with a large good quality BHC capacitor:

IMG_3956.jpg


and here is how the ripple looks like at the output of this circuit:

IMG_3954.jpg


A picture worth a thousand words... ;)
 
Teddy, thanks for those, i'll certainly reexamine my psu design /implementation....what's the scale on those 'graphs? Do you recommend any particular caps for your smoothing?
Uz
 
Thank you so much! I was about to order a transformer, rectifiers and caps based on the single large capacitance idea. Now I shall revise my plan.
 
Would be interesting to see a comparison with the equivalent capacitance and resistance and a single order filter i.e. 1.5R and 15000uF.

Dan
 
I was considering doing this in my PSU, but since I already have 15,000uF caps, can anyone comment on what is kind of a mid-way between the two designs mentioned above: just an RC (or LC) filter. The resistor value is independent of the capacitor value, right?

Do you have any suggestions for inductor values? (I actually tried to simulate using Duncan's PSU Designer, but I couldn't get LC filters to run, only RC).
 
Inductors can work but they are more suitable for power amps where resistors cannot be used. For pre amps, better use resistors.

Another advantage of this design is that you don't need huge transformers, a 50VA will be good enough. Large capacitors with low ESR require large transformers to charge them, here if you use small capacitors you don't need it.

Bottom line: smaller in size, costs less, and performs better...

Teddy
 
I was considering doing this in my PSU, but since I already have 15,000uF caps, can anyone comment on what is kind of a mid-way between the two designs mentioned above: just an RC (or LC) filter. The resistor value is independent of the capacitor value, right?

Do you have any suggestions for inductor values? (I actually tried to simulate using Duncan's PSU Designer, but I couldn't get LC filters to run, only RC).

The 1.5R // 15000uF will not filter the highest frequencies as strongly, however if will roll off much sooner to compensate. As HF noise is already strongly filtered by the TeddyReg, there won't be much in it either way.
 
Oh... I already have 15,000uF caps and I'm not sure what my transformer is acutally, it says 700mA for the 25-0-25 output, two of them. Does that make it ~70VA, or is it based on the primary voltage, in which case it would be ~160VA...?

Don't bigger caps just take longer to charge?
 
I'd imagine whatever's cheap. Metal film, wirewound... both good choices. I'd be tempted to over-engineer and use 2W resistors, but in reality it'll make little difference unless your preamp draws a lot of current.

You don't need anything fancy: these are 21p and I'm sure you could get cheaper :)
 
Good thread... gosh i learn so much here ... well at least it gives me a place to start reading! Beth Orton is with me tonight, bless her pointy-nose and hauntingly beautiful voice....
 
Hmmm... Interesting. Measurement is one thing, and effect on the sound is another. Is it possible that classical power supply without R1,2,3 sounds more dynamic and involving ?
Does anybody compared sound of both solutions?

:) Tomek
 
OK, so is it going to make any difference if the resistors are inductive or non-inductive - since I believe there are both kinds?
 
Not directly related, but this is the best set of pictures I could find that illustrates various filters in the power supply (http://www.clarisonus.com/Research Reports/RR002-DCFilterTest/RR002-DCFilSupplyTest.html). Look at the RC vs CRC filter - wow. I can see why Teddy went with the RC vs the usually large single cap or the CRC filters I've seen around. Thanks for you contributions as always, Teddy.

Do you have any data on diodes? I know the mur860 and msr860 are really popular, but Cree 1a Schottky diodes are less than $1 USD each at Digikey. Most people seem to dismiss them due to the pricing of the higher current models, but the low current model seems ideal for the current that the Teddyregs draw (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=CSD01060A-ND).
 
I built two different versions of teddyreg based power supplies, and the second one sounded so much better than the first that I had to find out why so I could bring the first one to the same level of performance.

They were constructed using :

1st- 48Vct 3A trafo / block bridge / BHC 15000uF / strip board teddyreg (like how to thread picture) using carbon resistors

2nd- 48Vct 0.5 A trafo / schottky / cap3 (3x2700uF+2x home made +-10uH air core) / strar ground pcb teddyregs using metal film resistors

What I did was build a third teddycap, but I made it modular so I could swap components easily. After all this, I also tried using wimas 10uF as output cap instead of tants. Here are my findings (% represents perceived importance of change):

Transfo : very slight difference (2%), the smaller one a bit less bloated and more agile. Surprisingly the smaller one seems to offer slightly better bass extension.

Bridge/ caps : biggest difference (70%), cap3 better in every aspect. More detailed, more insights, single cap sounded veiled and bloated in comparison.

Teddyreg : good difference (20%) surprisingly, PCB version has punchier bass, and cleaner highs. Maybe this was due to the different resistors used ?

Output capacitor : noticeable difference (8%), Wima weird at first but became better than tants after burn in. Removed some muddiness, better everywhere but less "bouncy"

I also played with cable, started with a flashback interconnect, switched to thick power cord which gave more bass extension but sounded bloated and veiled a bit. I then tried canare speaker cable (why not ? ) and this gives me what I wanted. IE nice bass extension while keeping all details and nuances. All cables (or maybe the solder) needed a few days burn in to sound their best

Next step for me is to swap inductors with resistors on the cap3 boards as per teddys findings.

Hope this helps someone.

Eric

The final Teddycap :


cap3.jpg
 
Do you have any data on diodes? I know the mur860 and msr860 are really popular, but Cree 1a Schottky diodes are less than $1 USD each at Digikey. Most people seem to dismiss them due to the pricing of the higher current models, but the low current model seems ideal for the current that the Teddyregs draw (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=CSD01060A-ND).

Don't know what they sound like, but SiC diodes have a very big forward voltage - 1.8V, versus about 0.4V for a Silicon schottky. They are normally used in higher voltage circuits where this is neither here nor there, but might do something strange. At the very least you will lose a volt or so of raw voltage. On the plus side there is no stored reverse charge.
 


advertisement


Back
Top