Babelfish XA252 / Babelfish XA252 SIT / Babelfish XA252 SET

[Zen Mod]

Few years ago, we had enormous fun in funny named thread : F4 Beast Builders

Well, you can call it Group of Hysteric Housewives Brainstorming 😛

Result of that was that I had myriad of back of napkin sketches, and after these I started simulating ...... and simulating ... and simulating

After myriad of sims ( out of sheer enthusiasm I even made pcb files for some of iterations), I finally had made nice smallish channel pcbs, containing OS having different-ish approach to biasing than Papa's original XA25, while FE was in principle same as Papa's, but with few beautiful quirks, of which I was damn proud (yup, Mighty ZM reinventing Da Wheel, time and time again)

And they were sitting, populated, on my shelf ........ always being pushed to bottom of list; I knew they would work, but somehow I wasn't content with overall picture

Now, recently - besides revising biasing mechanismus for SissySIT ( resulting in R.3) and few other amps, I also made FE called Plethora of Pinjatas ( or just Plethora, short) and found interesting to combine it in sitting Babelfish XA25 project

Plethora Fe having both SE and Bal inputs as matter of choice, called to adding another "2" in name , so result is Babelfish XA252

Plan is also to have iteration of DEF OS ( real SIT, no emulation in form of Schaded Mosfet), thus addition of "SIT" in name, so Babelfish XA252 / SIT

Find enclosed cumulative schematic, and will do edits of this post, if and when needed

OK1-OK2-TH1 .......... 10A limiting circuit, Savior of Our Souls, blatantly stolen from Pa 😛


amp is phase correct , Papa's fave gigglywiggly graphs sez that 2nd is negative phase

all measurements made with SE input, neg leg input grounded


edit on 18.07.2021.

Babelfish XA252 SIT - post #64 Babelfish XA252 / Babelfish XA252 SIT / Babelfish XA252 SET

Babelfish XA252 SET - post #119 Babelfish XA252 / Babelfish XA252 SIT / Babelfish XA252 SET


edit on 18.12.2022.

separate schmtc for Full Mos, SET, SIT iterations, post #624

ignore notes on schematics about "do not mount" - just follow schematics

read all down for separate schmtcs, Extra-carefully made set of files, for each of 3 amp iterations

short setting procedure:
*********************
P1 is for DC Offset, P2 is for Iq


do not forget to have grounded neg input , presuming you're using it SE input, but also ground positive input, temporary

no load at output

P1 irrelevant for start how it's set

P2 set to max value - check with probes across unpopulated Rt, that one being in parallel with P2; that way pre-set is min Iq

I don't care for left-right orientation of trimpots - never taking care about that when making pcbs - too much times in my service work I

expected some logic with trimpots orientation, just to be confused and from then on, I'm just using DMM as guide

leave JP1 open for start

advisable to put temporary fuses in rails - say 2A

if you have variac, use it

of course, regular praxis - one DVM across rail sense resistor group ( + or -, irrelevant) - Iq observer

second DVM at output, offset observer

power up, have an eye on Iq observer;

DC offset for start - good even if few volts, but normal to expect few dozen of mV

now, fiddle with P2, slowly, to get half of intended Iq

observe DC offset is it still OK


give it time to develop some heat in heatsinks, then increase Iq to 75%


if everything is stable, close JP1 ( alive, no need to power off for that), try to minimize DC offset as much you can

if that is done, open JP1 ( alive, no need to power off for that), wait for some time to get closer to temp equilibrium, increase Iq to

final desired value


that's it

try few lazy powering off - powering On sequences, observe behaviour

if everything is OK, power off, remove temporary fuses, wire rails properly

proceed/repeat procedure with second channel
*******************

edit on 22.12.2023.

set of files for last iteration, Post #1386 - consider it as hair-splitting changes, gilding da lilly, whatever; ZM playing in Sandbox...

read down.....

Edit on 07. Jan. 2024.

Extra-carefully made set of files, for each of 3 amp iterations, post #1462 https://www.diyaudio.com/community/...2-sit-babelfish-xa252-set.373443/post-7560120

 

[Zen Mod]

Now, for clarity sake, there is edited schm, parts needed for SIT iteration omitted

few more funny pics, too

 

[Zen Mod]

I did planned before to make it, but finally made T-Bed 2 , as proper cradle for prototyping this

from one of pics, you can clearly see that ZM is Chicken , executing first power up and settings with IRFP240/9140 and 2A fuses in rails...... did it with smallish Iq, and later - when I was happy camper, changed to Biguns

 

[Zen Mod]

Graphs , 8R load, BFR=22R

FE is made as more or less basic LTP thingy; though, I did wanted to have possibility of altering THD level by changing OLG in simple and predictable way - which I made in also basic way - tearing LTP halves ( each halve having own CCS) and introducing common degeneration resistor ( practically changing length of effective Tail of LTP) , which I then named as Beauty Factor Resistor, short BFR

as you will see, changing value of that resistor is chhanging OLG, thus changing amount of overall NFB, and as result we have lesser or higher THD level

Luckily, Spectra is not changing, and that was a goal

now - set of measurements with highest possible OLG - BFR s 22R

Rails are 30Vdc ( amp projected to work with rails of 30-35Vdc with no changes)

Iq 1A7

DC offset and Iq practically not changing in temperature domain

While varying mains voltage (nominal 230Vac) between 200Vac and 255Vac - DC offset is changing some 20mV, while Iq is practically not changing

Good enough for me

Load for these graphs is 8R; each graph is having title captured, so you can see what/how much

 

[Zen Mod]

Graphs , 8R load, BFR=22R, continued

too many pics for one post, so continued

due to 30Vdc rails, clipping smidge above 40W

with 35Vdc rails, clip near 55W

who needs that ?

 

[Zen Mod]

Graphs , 4R load, BFR=22R

everything the same, load 4R

 

[Zen Mod]

Graphs , 4R load, BFR=22R, continued

again

I didn't investigated where is bloody clipping

Meant that 70W is already enough

 

[Zen Mod]

Graphs , 8R load, BFR=510R, continued

Everything the same, but BFR changed to 510R, that being opposite extreme to 22R - practically lowest OLG which I considered

 

[Zen Mod]

Graphs , 8R load, BFR=510R, continued

more of these

 

[Zen Mod]

Graphs , 4R load, BFR=510R

everything same as previous, load changed to 4R

 

[Zen Mod]

Graphs , 4R load, BFR=510R, continued

more.....

 

[twitchie]

Fugly! Me likey!

When is swap meet thread coming?

 

[twitchie]

Impressive - still lower THD at 50W into 4R than many measly amps do at 1W and it's still able to stay H2 dominant. Again, many of the ones I measures start to become more H3 at higher output.

Have you done any SIT testing yet?

Do you like this one better than Singing Bush? Any comparisons you can share?

 

[Zen Mod]

Twitchie - I did made some listening - as always - decisive step should I proceed with some construction - with Ikebana style prototype, mono

Have some excuisite mono recordings, which I know damn well, and - if my foot is thumping and I feel like crying - that's it - damn thingie is worthy further pains

( lying, as always, there is no pain in this)

so - yes, I did listen and its worthy , but usual ZMAudiophooling commence tomorrow...... I mean , regular sissy stereo approach

well, it's spooky

didn't heard SIT iteration, but I have no any suspicion how it'll sound

so, few days of fun with full MOS, including visit to my Set of Spare Ears (hopefully tomorrow) then SIT tryout/presentation/visit to SSE

edit - no way to compare this Beast to Singing Bush Beast - really two sides of same coin ........ if you can call Coin territory marking same level of high quality

yup, ZM humble , as always

 

[twitchie]

...

Rails are 30Vdc ( amp projected to work with rails of 30-35Vdc with no changes)

Iq 1A7 ...

Chance for redemption for all the failed F5T builds and won't leave dry taste in your mouth

That's also big consideration for building this over BAF 50W (or SB) is no need for 63+V caps but is it still possible to sound like big powerful SET amp?

Only MZM knows 😀

 

[wwwtttwww]

impressive low THD+N at full power.

 

[twitchie]

Big problem now is I will have to go find or build big horribly inefficient speakers to burn all those watts

 

[flocchini]

Nirvana 😉

Mighty ZM - All I can say is WOW - what a roll you are on 😀

 

[Zen Mod]

Chance for redemption for all the failed F5T builds and won't leave dry taste in your mouth

That's also big consideration for building this over BAF 50W (or SB) is no need for 63+V caps but is it still possible to sound like big powerful SET amp?

Only MZM knows 😀

no, even if I'm sure that regular Greedy Boy is predestined to love both, these two are never going to sound, well, similar

breed thread is there, but difference in THD level is just too big ........... in fact, yes - this one can sound as big powerfulllllllll SET amp-cruising at 1% of potential

which, funny thing, is exactly how I think that Singing Bush is best used

impressive low THD+N at full power.

name of the game was to make Babelfish of XA25

so, that is it

Nirvana 😉

Mighty ZM - All I can say is WOW - what a roll you are on 😀

yup Bubba, tnx 🙂

 

[wg45]

I like it very much. My big, old, inefficient speakers are trembling at the moment.

Congrats ZM! You are rolling in amplifiers these days.

I blame Pa.