Or - Honey , did I just made World's first amp biased through little bugs in rails ???
Yup , we had some good time , having fun in Most Greedy Boy, of them all... or (there is no) DEFiSIT of Papa's Koans - diyAudio thread .
While we were contemplating of various things , mostly how to get PapaSits cheap and in unlimited quantity , we tossed more than few ideas how to make DEF amps - pure or little less pure , using parts we can lay our greedy hands on ... (Antoinel is playing with his own ideas in his own threads but , I'm not so talented to keep up with everything ..... or just too lazy , dunno )
Trying sims with various depletion parts (old stash from Semisouth Was Here era ) , Shaded mosfets , Japanese SITs etc.
More than few sims exchanged , more than few interesting ideas ; I believe few amps can be made ditto from LTSpice sshots shown there .
One of ideas was to keep , by any cost , sources of parts in question non-degenerated , as one of important "details" in Papa's purest DEFiSIT specimen(s) ; however , as zillion of sims showed , there is no way to keep source of any of these parts (or - to be more precise - compound pairs) clean , simply because THD spectra of said combos is simply not satisfactory , and some shaping of part behavior is easiest to do with source resistance , and strategic arrangement of same , regarding output node.
Anyway - anyone interested in all these examples , can read there.
One of real special moments was indra1's finding , brainstorming about biasing issue sans source resistors, that there are interesting Hall effect based chips , intended for current sensing ....... fact of , frankly , I was strictly and utterly ignorant .
So , again , I owe indra1 not just one big Thanks.
As some of you already know - M2 is my fave Papamp; not best one , not best made , not best sounding , not most versatile ...... but schematic of M2 is worthy of square mile big Poster ....... breathtaking example of wit , humor and brilliance . Besides that , it sounds bloody well .
Yup , we also had some special and good time , in F4 Beast Builders thread ..... about another specimen of Papa's enormously rewarding use of quality and fun time ; one of main characteristics of said amp (PL XA25) is non-existence of source resistors in output . That means pure square-law of transfer characteristics of output stage ..... in layman terms - Sparrow's fart is commanding to welding apparatus
.
I also have Babelfish XA25 in making , but low on schedule stairs .
So ..... indra1's idea was tempting .... tnx to meine Bruder Generg , I obtained ACS723LLCTR-05AB-T2 chips without too much fuss ..... already done some extensive simming of some possible iterations of M2 and XA25 bstrd
We know that work with big puck mosfets demands not just big heatsinks but also seriously big Cojones ; Mighty Moi , ZM , being jut tinytinytiny version of Papa (as we all are trying to be ) , recanted of pucks , and resorted for usual small bug output parts , already seen in regular M2
OK , 'nuff writing for first post , spell checker in my Firefox is already unconscious , not able to cope with ZMEngrish 🙂
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 ?
EDIT, TRADE post here : Babelfish M25 , AKA M2 on steroids , AKA M2-XA25 bstrd - kit packages
EDIT: pcbs matched to DiyAudio UMS , final schm and pcb sshot posted in #118, Babelfish M25 , AKA M2 on steroids , AKA M2-XA25 bstrd
Edit on 04.02.2025. final Babelfish M25/SET schematic; practicaly, very close cousin to SissySIT (42), few minor differences; post #243, https://www.diyaudio.com/community/...bstrd-aka-m2-gone-berserk.325038/post-7920166
Yup , we had some good time , having fun in Most Greedy Boy, of them all... or (there is no) DEFiSIT of Papa's Koans - diyAudio thread .
While we were contemplating of various things , mostly how to get PapaSits cheap and in unlimited quantity , we tossed more than few ideas how to make DEF amps - pure or little less pure , using parts we can lay our greedy hands on ... (Antoinel is playing with his own ideas in his own threads but , I'm not so talented to keep up with everything ..... or just too lazy , dunno )
Trying sims with various depletion parts (old stash from Semisouth Was Here era ) , Shaded mosfets , Japanese SITs etc.
More than few sims exchanged , more than few interesting ideas ; I believe few amps can be made ditto from LTSpice sshots shown there .
One of ideas was to keep , by any cost , sources of parts in question non-degenerated , as one of important "details" in Papa's purest DEFiSIT specimen(s) ; however , as zillion of sims showed , there is no way to keep source of any of these parts (or - to be more precise - compound pairs) clean , simply because THD spectra of said combos is simply not satisfactory , and some shaping of part behavior is easiest to do with source resistance , and strategic arrangement of same , regarding output node.
Anyway - anyone interested in all these examples , can read there.
One of real special moments was indra1's finding , brainstorming about biasing issue sans source resistors, that there are interesting Hall effect based chips , intended for current sensing ....... fact of , frankly , I was strictly and utterly ignorant .
So , again , I owe indra1 not just one big Thanks.
As some of you already know - M2 is my fave Papamp; not best one , not best made , not best sounding , not most versatile ...... but schematic of M2 is worthy of square mile big Poster ....... breathtaking example of wit , humor and brilliance . Besides that , it sounds bloody well .
Yup , we also had some special and good time , in F4 Beast Builders thread ..... about another specimen of Papa's enormously rewarding use of quality and fun time ; one of main characteristics of said amp (PL XA25) is non-existence of source resistors in output . That means pure square-law of transfer characteristics of output stage ..... in layman terms - Sparrow's fart is commanding to welding apparatus
.I also have Babelfish XA25 in making , but low on schedule stairs .
So ..... indra1's idea was tempting .... tnx to meine Bruder Generg , I obtained ACS723LLCTR-05AB-T2 chips without too much fuss ..... already done some extensive simming of some possible iterations of M2 and XA25 bstrd
We know that work with big puck mosfets demands not just big heatsinks but also seriously big Cojones ; Mighty Moi , ZM , being jut tinytinytiny version of Papa (as we all are trying to be ) , recanted of pucks , and resorted for usual small bug output parts , already seen in regular M2
OK , 'nuff writing for first post , spell checker in my Firefox is already unconscious , not able to cope with ZMEngrish 🙂
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 ?
EDIT, TRADE post here : Babelfish M25 , AKA M2 on steroids , AKA M2-XA25 bstrd - kit packages
EDIT: pcbs matched to DiyAudio UMS , final schm and pcb sshot posted in #118, Babelfish M25 , AKA M2 on steroids , AKA M2-XA25 bstrd
Edit on 04.02.2025. final Babelfish M25/SET schematic; practicaly, very close cousin to SissySIT (42), few minor differences; post #243, https://www.diyaudio.com/community/...bstrd-aka-m2-gone-berserk.325038/post-7920166
Last edited:
I'm not going to post sim schematic , instead of that I'm jumping ditto to Eagle schematic and resulting pcb patter screenshot
Disclaimer - Universal Whatnot Mounting Spec .... is for Sissies
I resorted for Papa's FW format ; I'll not gain any Cojones with that , but it's a sort of Kudos to SR Kitchen Table Enterprise , most probably ....... I just like that idea 🙂
Disclaimer - Universal Whatnot Mounting Spec .... is for Sissies
I resorted for Papa's FW format ; I'll not gain any Cojones with that , but it's a sort of Kudos to SR Kitchen Table Enterprise , most probably ....... I just like that idea 🙂
Attachments
Input buffer comprised of bunch of 2SK2145BL , in form of same-sex JFets buffer , well known from .... dunno ;
attached two pictures , taken from PapaBorbely's JFet frontiers articles ;
first one example of same sex JFet buffer ; there I practically have 6 up , 6 down (2SK2145BL being dual JFet in sissy package) , where Iq is measly 21mA 🙂
yup , cascoded to some 12V or so at drains
Second pic is Papa Borbely's implementation of tube White Cascode in JFet world ; I took provision of parts needed on pcb , so I can use kosher same sex buffer , or try slightly different THD spectra , which WC is bringing on table ; mostly opposite phase of dominant 2nd , speaking of buffer only
Though , for my case , modulating resistor in upper drains is different value than shown in pic.
attached two pictures , taken from PapaBorbely's JFet frontiers articles ;
first one example of same sex JFet buffer ; there I practically have 6 up , 6 down (2SK2145BL being dual JFet in sissy package) , where Iq is measly 21mA 🙂
yup , cascoded to some 12V or so at drains
Second pic is Papa Borbely's implementation of tube White Cascode in JFet world ; I took provision of parts needed on pcb , so I can use kosher same sex buffer , or try slightly different THD spectra , which WC is bringing on table ; mostly opposite phase of dominant 2nd , speaking of buffer only
Though , for my case , modulating resistor in upper drains is different value than shown in pic.
Attachments
as you can see, even if hardly - it's crowded there , I decided to make possible use of 3 different types of xformer :
-same Edcor as used in original M2 ( and I did all protowork with that one)
-Jensen repeater ( is that one used in F6 ? lazy to check datasheets and type in this moment)
-Cinemag CMOQ4-LPC (also repeater) , same as Papa used in DEFiSIT ;
first one giving gain of 6V/V , two later ones giving gain of 4V/V ;
I'm pretty sure that Edcor is also having repeater coil in same form factor as their 600:15K , if one is to choose cheapest variant ; though , Edcors are cheap , but not bad , in any case.
-same Edcor as used in original M2 ( and I did all protowork with that one)
-Jensen repeater ( is that one used in F6 ? lazy to check datasheets and type in this moment)
-Cinemag CMOQ4-LPC (also repeater) , same as Papa used in DEFiSIT ;
first one giving gain of 6V/V , two later ones giving gain of 4V/V ;
I'm pretty sure that Edcor is also having repeater coil in same form factor as their 600:15K , if one is to choose cheapest variant ; though , Edcors are cheap , but not bad , in any case.
Attachments
Last edited:
Then , to finally make long ago dreamed of - Test Bed ...... I'm so happy with new toy , that I even have a name for it - T-Bed
long screws situated at sides are good as handles and also as support , when you toss T-Bed on any side ...... no possibility to pinch or brake any part against the bench
long screws situated at sides are good as handles and also as support , when you toss T-Bed on any side ...... no possibility to pinch or brake any part against the bench
Attachments
Last edited:
then later , time to divide boyz against the men ........ LTSpice and Back of Napkin Sissies against ....... just lucky ones ....... as I am 🙂
bench time ; observe direct feed from siggene to buffer then to output stage , sans xformer ; why wasting perfectly good and new Edcor on possibly not good shebang
bench time ; observe direct feed from siggene to buffer then to output stage , sans xformer ; why wasting perfectly good and new Edcor on possibly not good shebang
Attachments
Last edited:
OK , it sings and it behaves.... I can solder Edcors in , I can connect it to my precious spks and listen to some muuuusic!
two problems - need some air on heatsink ( will do that later ...... tomorrow) and - my speakers are easier to drive than some of headphones , so no bloody use and benefit of square law output behavior
two problems - need some air on heatsink ( will do that later ...... tomorrow) and - my speakers are easier to drive than some of headphones , so no bloody use and benefit of square law output behavior
Attachments
Next day , to solve few problems ........ naah , I can solve just one , lacking in not easy to drive speakers department
so , tossed 12cm fan , small 5+0V6+0V6 reg , and that's it
later I added switch to fan circ , just to ensure shorter warm up
rails are at healthy 25Vdc , Iq first set on inline 0R1 for each channel separately , later removed 0R1s , re-checked at 0R11 which are part of CRC in filter
yup - Iq is 1A7 , per channel ..... I don't wanna depend of some hypothetic square law ( I can't see it ) , so better more than less
that resulting in nice for summer 25 x 2 x 1A7 x 2channels =170W of heat
heatsink with lazy fan at Papa's Crikey , it can cruse all day long
IR measuring gun-like thingie with red dot is at home .... who needs that , anyway
so , tossed 12cm fan , small 5+0V6+0V6 reg , and that's it
later I added switch to fan circ , just to ensure shorter warm up
rails are at healthy 25Vdc , Iq first set on inline 0R1 for each channel separately , later removed 0R1s , re-checked at 0R11 which are part of CRC in filter
yup - Iq is 1A7 , per channel ..... I don't wanna depend of some hypothetic square law ( I can't see it
) , so better more than lessthat resulting in nice for summer 25 x 2 x 1A7 x 2channels =170W of heat
heatsink with lazy fan at Papa's Crikey , it can cruse all day long
IR measuring gun-like thingie with red dot is at home .... who needs that , anyway
Attachments
few days later , I had on mind solving/trying few things :
-trying it on some harder to drive spks
-solving problems I had with my THD Spectra measuring rig . I coudn't make it to work after putting "new" EMU1212 and giving old Audigy to friend .
I wanted and needed bal in/bal out , but from sme reason bloody thing is not working ; box with at least 4 various Audigy cards is simply impossible to find in all mass of boxes on zillion of shelves I made crowded in years .
EMU is simply having too much channels , so my brain can't cope with it ...... or ?
anyway , first problem first .... at least while I'm not having solution for second one ; I already posted few pics (in http://www.diyaudio.com/forums/pass-labs/314822-greedy-boy-defisit-papas-koans.html#post5244007) of listening session/having fun occasion , we had at my friend Woland's ......
He's having Poor Man's RCA speakers - Tann DC in Gorgeous and Voluptuous and Luxurious MLTL cabs , dreamed/constructed by ZM , MIghty , Who Else .
🙂
So - after painstakingly long warm up time ( I didn't had that fan switch at a time , nor brain to realize necessity of it) , Babelfish M25 finally came to sound ....... it singed dozen of tunes , without special excitement from anyone .......
Reason , later I reckoned - two reasons:
- they're all bloody hooked on Papamps , so nothing new there,
- bloody Tanns in MLTL are also easy to drive , so no shine for M25 Cojones
-trying it on some harder to drive spks
-solving problems I had with my THD Spectra measuring rig . I coudn't make it to work after putting "new" EMU1212 and giving old Audigy to friend .
I wanted and needed bal in/bal out , but from sme reason bloody thing is not working ; box with at least 4 various Audigy cards is simply impossible to find in all mass of boxes on zillion of shelves I made crowded in years .
EMU is simply having too much channels , so my brain can't cope with it ...... or ?
anyway , first problem first .... at least while I'm not having solution for second one ; I already posted few pics (in http://www.diyaudio.com/forums/pass-labs/314822-greedy-boy-defisit-papas-koans.html#post5244007) of listening session/having fun occasion , we had at my friend Woland's ......
He's having Poor Man's RCA speakers - Tann DC in Gorgeous and Voluptuous and Luxurious MLTL cabs , dreamed/constructed by ZM , MIghty , Who Else .
🙂
So - after painstakingly long warm up time ( I didn't had that fan switch at a time , nor brain to realize necessity of it) , Babelfish M25 finally came to sound ....... it singed dozen of tunes , without special excitement from anyone .......
Reason , later I reckoned - two reasons:
- they're all bloody hooked on Papamps , so nothing new there,
- bloody Tanns in MLTL are also easy to drive , so no shine for M25 Cojones
Attachments
back to reality , day or two after
I found what's wrong with EMU - lack of brain combined with defective LCH output
switched to RCH output and all is well 🙂
of course that I made a mistake , while making new interface box wit bal in/bal out capability - I put it in plastic box .... forgetting that steel is better , esecially when dealing with lower level signals
one thing more to re-do
that time I'm going to put some protection zenners , at side looking back in sound card ; I don't need them , but who knows what's going to happen if I stumble on cable , while going on/from bench
I found what's wrong with EMU - lack of brain combined with defective LCH output
switched to RCH output and all is well 🙂
of course that I made a mistake , while making new interface box wit bal in/bal out capability - I put it in plastic box .... forgetting that steel is better , esecially when dealing with lower level signals
one thing more to re-do
that time I'm going to put some protection zenners , at side looking back in sound card ; I don't need them , but who knows what's going to happen if I stumble on cable , while going on/from bench
Attachments
Last edited:
now , battery of tests at 1W
remember - crappy PSU (old caps - see that garbage left of 1K) , intentionally crappy wiring
due to lowish levels involved and my overall lazyness , I must accept sort of lower dBFS levels ....... don't know no one able to make me some decent gain stages with uber OP chips ..... 🙂
1W/8R (see - same figure as Papa's original M2 ..... easy load)
1W/4R
1W/2R
remember - crappy PSU (old caps - see that garbage left of 1K) , intentionally crappy wiring
due to lowish levels involved and my overall lazyness , I must accept sort of lower dBFS levels ....... don't know no one able to make me some decent gain stages with uber OP chips ..... 🙂
1W/8R (see - same figure as Papa's original M2 ..... easy load)
1W/4R
1W/2R
Attachments
I was thinking today - I'm really lazy to make superduper OP gain stage , to obtain 6db thingie , and put in between EMU output and interface box , to be able to get more watts for scope; I'm going to use precious Iskra repeater as autoformer , with 6db gain , and see what I'm going to get
I'm anyway more interested in comparison of THD on higer power levels than in absolute values
so , I tried exactly that : 6db gain autoformer ditto on EMU output , so you can see loop-back test is ........ Steel, ZM , Steeel!!!!! , almost joke
so - ignore again everything left of 1K , take rest with pinch of salt
I'm anyway more interested in comparison of THD on higer power levels than in absolute values
so , I tried exactly that : 6db gain autoformer ditto on EMU output , so you can see loop-back test is ........ Steel, ZM , Steeel!!!!! , almost joke
so - ignore again everything left of 1K , take rest with pinch of salt
Attachments
Last edited: