Babelfish XA252 / Babelfish XA252 SIT / Babelfish XA252 SET

now, targeted Iq is 1A7, but remember - due to Square Law transfer function , number isn't exactly carved in stone - it can be lowered without much penalty ( no louder KlunK!)

Nice reference, I just read "leaving class A" this morning (again!). I was wondering about the penalty if one would lower the bias, as the article mentions more advantages with increasing bias:
IMG_20220408_192403.jpg

I will at least use a transformer with 2x2 secondaries. Xa25 has one transformer, right?

Thanks for the detailed explanation. Would you have become an engineer, we had no one to nag at!:devilish:
 
  • Like
Reactions: 1 user
PSU’s are all about cost efficiency. For the price range, a single PSU is probably more than good enough, and all possible measures taken iot make it as good as possible.

but if you read Papas paper on power supplies, it is apparent that for the relentless approach, where cost is no object, external PSUs, mono blocks, and bigass inductors - even choke input supplies, is evidence of a makers dedication. This is readily apparent in the most costly of the Pass Labs amplifiers. Likely following Papas 1.000 dollars per kilo amplifier approach :)

We DIYers have the advantage of putting countless hours of effort into these projects. There are not many drawbacks in an oversized PSU, as long as it is constructed properly. Besides cost and weight. There are equally few objective technological advantages of an undersized PSU. But there are probably plenty of subjective advantages. And then there ate the factors of cost and time, objective at that.

It costs, alot. And is heavy as hell. Attached is a pic of the first PSU I made with my bare hands. No, «I did not extrude the metal». But it was fun as hell, and runs like on batteries. Glad my back is functioning again after last years surgery. Key factors of success: no hum, and it is possible to bias the amp despite the huge supply.

yes, it could likely have been built far better. With higher voltage rated caps, eith split gnd planes, etc etc. But I like it. Cuz I did it. Not without help from Dennis and ZM. And Papa. But still :)

correction: I did extrude those ground planes :rofl:

regards,
Andy
 

Attachments

  • ACCDEB3F-71D6-4F78-8E20-FA0B93971AF3.jpeg
    ACCDEB3F-71D6-4F78-8E20-FA0B93971AF3.jpeg
    554.3 KB · Views: 106
  • Like
Reactions: 1 users
Nice reference, I just read "leaving class A" this morning (again!). I was wondering about the penalty if one would lower the bias, as the article mentions more advantages with increasing bias: View attachment 1042619
I will at least use a transformer with 2x2 secondaries. Xa25 has one transformer, right?

Thanks for the detailed explanation. Would you have become an engineer, we had no one to nag at!:devilish:
I believe ZM’s point was that the square law behaviour of the devices in this circuit sorta defies physics, and so a little lower Iq won’t imply the same amount of penalty as in an output stage with for example source resistance. No?
 
  • Like
Reactions: 1 user

Plott

Member
Paid Member
2021-09-12 6:34 pm
Vienna, Austria
I didn't made any measurements at 1A of Iq nor I intend to do that (lazy or not wanting to promote idea?)

though, I believe my friend when he's saying that there is no difference to speak of, with his sensitive speakers

he's having exactly the same as shown here: https://www.zenmod.in.rs/sonido-15-way-speakers-by-zm-woland-edition/

amp is made to operate in bracket of 30Vdc to 35Vdc per rail

if you go for 30vdc, there is opening to use 35Vdc caps in cap bank, they will endure wiith ease short unbiased ampp condition during power up sequence, when rails are unloaded
if you go for 35Vdc, you must resort to 50Vdc caps as minimum

taking in account usual multiplication factor of 1.25 (AC to DC), applicable to heavy loaded rectification/filtering block, that calls for 24Vac secondaries in first case, or 28Vac secondaries in second case

give or take half a volt, who cares

now, targeted Iq is 1A7, but remember - due to Square Law transfer function , number isn't exactly carved in stone - it can be lowered without much penalty ( no louder KlunK!)

so, 35Vdc rails, 1A7 Iq ; that's 119VA; doable with Modu 5U/500 (Babysitter during Summer just for ease of sleep and prolonged caps life)

say that I'm Chicken, so no less than 300VA of Donut power per channel

will you go with two of that size or one double - your choice

if Cheapskates of Yore, producing Tube amps in numbers, could use common cathode resistors for both OS channels, there is no reason one can't use common PSU for both amp channels - it increases organicity of sound and it enlarges Mono Base ( coherence of Soundscape )

matter of choice, but let's make it informed (choice)

:)

- sorta important detail - advisable to use separate secondaries/one bridge per secondary arrangement; as with majority of my little amps, also in case of Babelfish XA252 SET and Babelfish XA252 SIT - dynamic loading of rails is not uniform - result of complementary connected output devices not exactly being complementary, so having different/wakoo transfer characteristic

let's phrase it differently - whenever you have proper mosfet in one rail and SIT or Schaded mosfet in another rail, proper mosfet is having greater xconductance than it's partner

that resulting in greater dynamic loading of rail where proper mosfet is situated

so, if there is usual power xformer arrangement - center tapped secondary, entire rails loading difference will transpose to Center Tap ( used as Audio Gnd) , and xformer - happy stupid critter - is subjected to see it as nothing else as intermittent DC

We all know that xformers don't like DC and completely normal consequence is nothing else than buzz ........ and spoiling of xformer transfer characteristic

spoiling is going to result in spoiling AC coming from transformer, putting garbage back in rails, so some smearing of supposedly beautiful Soundscape ......

go figure

wouldn't be easier that I really ended as Construction Engineer, instead of messing with multilegged fuses in Papaland ?

:rofl:
Those speakers are great, "want-have" :deerman:
 
Those speakers are great, "want-have" :deerman:

for you, nothing easier than calling Istvan, Sonido

https://sonido.hu/kapcsolat.php?lang=hu

I have nothing but strong recommendation for their fullrange and extended range drivers

planning to pony up for their 10" FC drivers, even if God knows that I really don't need more speakers (twice,trice....)

somehow, my speaker related greediness is hardest to tame