Babelfish M25, AKA M2 on steroids, AKA M2-XA25 bstrd AKA M2 gone Berserk

Official Court Jester
Joined 2003
Paid Member
I've put some edit lines on end of first post :

edit :

Few lines about intention of experiment:

1. to pursue possibilities of using small bug cased 2SK2145BL Toshiba JFets , of present production , instead of more and more rare Toshiba complementary oldies or always scarce LS JFets ........ in both usual SE configuration and in Papa Borbely's White Cathode Follower transpose in JFet realm ;

1. to make possible usage of Edcor , Jensen and Cinemag xformers of final user's choice ;

3.to pursue possibilities of current sensing in rails , to avoid source resistor in output , further allowing:

4. to pursue changes in OS current capabilities , allowed with unspoiled square law transfer characteristics of mosfets used , enabling significantly enlarged A class envelope (4 Ohms load - A class up to 40-ish Watts , then Klunk! up to 50W , 2 Ohm load -A class up to 20-ish Watts , then clunk up to Wattage you are brave enough to invoke ..... so less Klunk!!-ing in your sound ; also lower THD comparing to regular M2 .



So , 4 or more intentions realized , Mighty ZM fully satisfied .

Am I same as Brave Mickey The Tailor ?
 

Attachments

  • Brave Mickey.jpg
    Brave Mickey.jpg
    17.7 KB · Views: 829
Official Court Jester
Joined 2003
Paid Member
well ........ it'll be swell ......... but no

proper trio for upper and proper trio for lower position

Idss in say 5% range for trio , difference in summ Idss between trios , say, max 15%

even then , Temp Co (for DC offset) is way superior than with complementary version
 

Attachments

  • WP_20180716_21_02_44_Pro.jpg
    WP_20180716_21_02_44_Pro.jpg
    109.4 KB · Views: 845
I didn't

I got it exactly as is ...... using it for years for same purpose

hey , I even did Babelfish J2 prototyping on same htsnk :)

back in 2010.

OK ZM, :D but you know that size does matter.. sometimes :male:
I just used some surplus aluminum board.

I am trying to get a grip on how the M2 circuit works. :confused:
Can give me a tip on those components that I do not understand? Optocoupler seems for against overcurrent through the power FETs. Per overcurrent it makes the voltage to open the Q5 and then closing the FETs. But is this just for protection or to continuously stabilize the bias precisely? What is the effect on the sound?

I thought I need to understand the M2 to follow you further.
You see that I am not electronic engineer however I would to taste the charm that you cherish about M2..
 

Attachments

  • M2 with pinout by ZM.jpg
    M2 with pinout by ZM.jpg
    119.5 KB · Views: 1,098
Official Court Jester
Joined 2003
Paid Member
OK ZM, :D but you know that size does matter.. sometimes :male:
I just used some surplus aluminum board.

I am trying to get a grip on how the M2 circuit works. :confused:
Can give me a tip on those components that I do not understand? Optocoupler seems for against overcurrent through the power FETs. Per overcurrent it makes the voltage to open the Q5 and then closing the FETs. But is this just for protection or to continuously stabilize the bias precisely? What is the effect on the sound?

I thought I need to understand the M2 to follow you further.
You see that I am not electronic engineer however I would to taste the charm that you cherish about M2..



damn ....... it took me longer to find it than to write it again ....... but sorta sleepy , so better to search for than to read my own brainfarts tomorrow ;

(edit: I'm also not electronic engineer , but that's not preventing me of making my self a fool in public , on regular basis :rofl: )

anyway , it is here :

Official M2 schematic

to quote my self :

optocoupler is there , functioning in this way:

LED is fed with voltage across two source resistors
series resistor for LED is to keep current and voltage for LED in safe area

LED is shinin' on transistor base , defining conductivity through it

transistor is part of current path from + to - PSU rails , and it's maintaining proper voltage between mosfet gates

(current through mosfets rise , LED shining brighter , transistor more open/lesser effective resistance , so lesser voltage between gates so current through mosfets decline)

3300uF across transistor is to neutralize process of normal current fluctuation through mosfet source resistors , so LED is having in same manner fluctuating voltage , inducing same fluctuating of light to transistor base

that way voltage between mosfet gates is held pretty solid and steady

so imagine that cap as wide open solid fork ,placed with fork pins between two gates

on whichever side (pin) of fork you're pushing/pulling , other side is following , always staying exactly same apart

in case of M2 , fork is pushed/pulled on lower end , through C2

you can connect it to other end of C3 , result will be the same

I bypassed C3 with 1uF polycarbonate just because I can , will sleep like a baby , knowing that I'm done everything that fork stays solid even on highest frequencies**

that's exactly intended role of C3 - to not conduct signal , but to represent solid pushrod to signal ; that trick/part of circ is sometimes called level shifter

**not that's for certain , but sleep is important



(all ref. to schm in post #1 ; sketch of exact forks missing there , so I'm enclosing it , just in case ; second picture just because I'm Enormous and my Ego is Enormous)

:clown:

edit: linked schm from #13 , so you don't need to jump from tab to tab :

510732d1445768234-official-m2-schematic-m2-pinout-zm.jpg

587158d1482372268-official-m2-schematic-fork-jpg


587159d1482372268-official-m2-schematic-einstein-quotes-imagination-8-jpg
 
Last edited:
Official Court Jester
Joined 2003
Paid Member
and today is first day of evaluation on Dynaudio Contour 1.8 Mk2

4R impedance .....

after just 30min (after initial warm up) it is clear that Babelfish M25 is damn well improvement over regular (Papa's Koan) M2 ......... and that still even without trying/engaging (supposedly superior ) White Cathode Borbely JFet buffer ...... and that one is promising to have objectively lower distortions , driving that nice Edcor Iron

and , we are speaking here about properly made M2 - dual mono (each channel having 250VA Donut and 4 x 33mF caps) with proper wiring inside ...... while T-Bed is having common supply for both channels , with old and tired caps and sooooo crappy wiring ........ and with flimsy 1A6 fuses inline with speakers

:)
 

Attachments

  • WP_20180720_11_55_29_Pro.jpg
    WP_20180720_11_55_29_Pro.jpg
    76.1 KB · Views: 399
  • WP_20180720_11_55_41_Pro.jpg
    WP_20180720_11_55_41_Pro.jpg
    78 KB · Views: 334
  • WP_20180720_11_55_50_Pro.jpg
    WP_20180720_11_55_50_Pro.jpg
    67.2 KB · Views: 303
  • ZM-Genius.jpg
    ZM-Genius.jpg
    27.5 KB · Views: 277
Last edited:
Official Court Jester
Joined 2003
Paid Member
Russellc asked , over in Official M2 schematic thread :

Kicking self in head, are these be berserking M2 (M25) boards going to be available?
Russellc



so :

this days .....thinking about it ....

it can be done ,though not exactly for cheap , considering matching and soldering tiny smds and pretty much low numbers nature of venture

so , viable offer could be :

A:
-pcbs with presoldered most critical components (all smd ) - input buffer and current sensing chips with surrounding filtering caps ,
-my beloved MKC caps for signal route electrolyt (10uF Silmic or Nichicon) bypass
-matched BD 139/140 cascoding bjts


B:

-same as above ,
-all (MF 0W6) resistors
-all multiturns
-all electrolytic caps (Pana, Nichicon, working on genuine Silmic for signal route)
-4 micas for outputs
-pinholders for few resistor positions ( regular buffer or Borbely's WCF option)
-IRFP outputs (damn , looking for Harris mnfctrd 9240)
-2 optocouplers
-2 7805 regs

.......... practicaly all parts on pcbs , excluding mosfet bolts and pcb bolts and signal xformers (too complicated for me to source them from anywhere in larger quantities)

to repeat 3 important things :

1.input buffer can be arranged in two ways : either as regular SE (this one having same THD spectra as complementary Toshiba buffer having 1 up and 1 down) , OR Borbely's WCF JFet buffer ; easy to change between two options, if two resistors positions are having soldered pin holders and one cap position having the same.
WCF buffer is having lower THD , but opposite K2 phase .... so flipping speaker's wires is more than advisable , to choose best for your auditory equipment

2. practically 4 options for signal autoformer (pinouts for all present on pcb):

-Edcor PC series 600:15K (same one as in original M2) , thus having gain of 6V/V

-Edcor PC series 600:600 , thus having gain of 4V/V (edit - only for this one it'll be needed few exacto cuts of traces and few tiny wire bridges soldering ;same physical layout of pins as previous one , but different electrical connection needed ..... very hard to accommodate without too much complication )

- Jensen JT-123-FLCPH (same one as in F6) , thus having gain of 4V/V

-Cinemag CMOQ-4LPC (same one as in SIT-3) , thus having gain of 4V/V

3. most important benefit of all this Berserking is unspoiled square law transfer characteristics of mosfets used , enabling significantly enlarged A class envelope (4 Ohms load - A class up to 40-ish Watts , then Klunk! up to 50W , 2 Ohm load -A class up to 20-ish Watts , then clunk up to Wattage you are brave enough to invoke ..... so less Klunk!!-ing in your sound ..... resulting in significantly lower THD comparing to regular M2 .



so , what you say ?


You can call it XA25 for Sissies , if you want :rofl:
 
Last edited: