This is what I am thinking on. Course i dont haev SiT's, but Perhpas SiC's will do. Might as well add Sh'T.
So, what do we know?
We know there is a comp buffer at the input.
From Posts #61, 88, 214, 292 and of coarse #253 we know there is a 4 X 600 Ohm transformer windings. We don't however know how, where or if all 4 are used. I'll bet that 2 primaries are paralelled. In post #214 NP comments you can use 2 1:1s so? Also, I beleive that buffer will probably do OK driving 2 600 ohm windings?
And we know there is a "clasic" Quasi Comp looking Sic output stage.
We also see some (AC) feedback there?
Naturally there are other components and just because it's a classic there is no reason not to expect a little Pass twist?
We know there is a comp buffer at the input.
From Posts #61, 88, 214, 292 and of coarse #253 we know there is a 4 X 600 Ohm transformer windings. We don't however know how, where or if all 4 are used. I'll bet that 2 primaries are paralelled. In post #214 NP comments you can use 2 1:1s so? Also, I beleive that buffer will probably do OK driving 2 600 ohm windings?
And we know there is a "clasic" Quasi Comp looking Sic output stage.
We also see some (AC) feedback there?
Naturally there are other components and just because it's a classic there is no reason not to expect a little Pass twist?
Pa wrote exact Jensen type ; it's 4x150R , where you're using primaries in parallel , and secondary as drawn , so all 4 windings are used
as I wrote about my crude testing few posts above - 50R siggene doesn't sweat driving two 150R prim. in parallel
isn't enough of twist bringing to light some old techniques , combined with modern parts?
besides that - be sure that some thorough work is behind that schmtc
fact that every Toobhead is aware of them doesn't diminish measure of twist ....
in fact - I'm having old Pioneer SX1000T receiver besides me on the shelf (FM FE Nuvistors , everything else Germainum , Silicium T03 cans) where phase splitter is made exactly the same , with exactly same logic for feedback net .
whatever , it wasn't me who proposed that schmtc , but wittier Joker than me .....
yes , of course ; two identical 600:600 xformers , both primaries in parallel (in phase ) , then secondaries as drawn
in that case - it will buffer going to be happier than driving two (Jensen's) 150R primaries in parallel ?
that's strictly matter of blending with primary inductance , not matter of impedance ; xformers are - so called - 600:600 or whatever - those numbers are information of intended purpose ( each one dictating inductance and Rdc of windings ) but , generally speaking , one can find xformer specimens with similar inductance , one called 150 Ohms , other one called 600 Ohms , third one called 1K Ohms .
in this ( and many other cases ) reflected impedance to source , in this case load which buffer is seeing , is matter of input impedance of each output gate mechanismus , then these two in parallel ...... then that transferred 1:1 through winding ratio back to buffer
conclusion - ratio is what's counts most , combined with load ; only knowing (or anticipating these during construction ) we can say how's source-xformer-load combo called , in regard to how-many-Ohms
then - depending how much inductance is (so called 150, or so called 600, or so called 1K , whatever) in primary - that will reflect in left F3 , phase behavior etc.
higher range (impedance ) xformer will be safer regarding left F3 and phase behavior , but it will present more troubles in right F3 (capacitance between windings ! ) and slightly worse generator for gates (of output Jfets)
whoa ! I'm sorta spinnin' in circles now ......
Last edited:
just tested ( to refresh my weak memory) those repeater coils (pictured in one of my previous post)
exacttype is Iskra TL-1M , so called 600:600 , in fact 4 x 150R
two primaries connected in parallel , fed with siggene , 50 ohms tap
CRO across one secondary , loaded with 6K8 (estimated range of load in F6 )
it's 'ookin' linear 15Hz-58Khz , and up to 280Khz ;
there is some rise in few freqs between 58KHz and 180KHz , but nothing what appropriate RC loading can't handle
and they're primary made for telephone range
but do not forget some other wakoo implementations of telephone transmission lines - teleprinters and other crazy military purposes
considering that I'm expecting that loading impedance of each (of two ) output's gate circuits is in range of 5-10K , we can freely say that load reflected to buffer is (5-10K)//(5-10K) = 2K5-5K ; that's good enough , at least in my neck of wood
Last edited:
yes , of course ; two identical 600:600 xformers , both primaries in parallel (in phase ) , then secondaries as drawn
in that case - it will buffer going to be happier than driving two (Jensen's) 150R primaries in parallel ?
that's strictly matter of blending with primary inductance , not matter of impedance ; xformers are - so called - 600:600 or whatever - those numbers are information of intended purpose ( each one dictating inductance and Rdc of windings ) but , generally speaking , one can find xformer specimens with similar inductance , one called 150 Ohms , other one called 600 Ohms , third one called 1K Ohms .
in this ( and many other cases ) reflected impedance to source , in this case load which buffer is seeing , is matter of input impedance of each output gate mechanismus , then these two in parallel ...... then that transferred 1:1 through winding ratio back to buffer
conclusion - ratio is what's counts most , combined with load ; only knowing (or anticipating these during construction ) we can say how's source-xformer-load combo called , in regard to how-many-Ohms
then - depending how much inductance is (so called 150, or so called 600, or so called 1K , whatever) in primary - that will reflect in left F3 , phase behavior etc.
higher range (impedance ) xformer will be safer regarding left F3 and phase behavior , but it will present more troubles in right F3 (capacitance between windings ! ) and slightly worse generator for gates (of output Jfets)
whoa ! I'm sorta spinnin' in circles now ......
Do we have any confirmation from Papa that 2 x 2-winding transformers with have the same behaviour as 1 x 4 winding transformer? In an earlier post http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-44.html#post3109366 he refers to the "quadfilar windings have symmetric capacitance, so all windings exhibit the same conditions with respect to that" . It seems to me that with one transformer there would be more capacitive coupling between the transformer secondaries than with 2 transformers.
DC? Bias flows from + to - supply. Only a small offset current "may" flow thru the primary...
ZM, oh mighty one
According to Bill Whitlock, I'm sure you recognise the name, sited in a post many posts ago, and likely the engineer involved in the F6 Xfrmr, says; "the distortion performance of ANY transformer is significantly improved when the driving source impedance is less than the dc resistance of the primary. However, little is gained below about 10% of the winding dc resistance."
So, my "150" ohm primaries are about 28 ohms DC. Ohms not turns, not L, not Z, and if those Jensens are the same, and they seem to be, we have not satisfied a design requirement Bill thought important??? do you have any comment on that idea. I know I stated that already but you responded to freq resp issues instead
ZM, oh mighty one
According to Bill Whitlock, I'm sure you recognise the name, sited in a post many posts ago, and likely the engineer involved in the F6 Xfrmr, says; "the distortion performance of ANY transformer is significantly improved when the driving source impedance is less than the dc resistance of the primary. However, little is gained below about 10% of the winding dc resistance."
So, my "150" ohm primaries are about 28 ohms DC. Ohms not turns, not L, not Z, and if those Jensens are the same, and they seem to be, we have not satisfied a design requirement Bill thought important??? do you have any comment on that idea. I know I stated that already but you responded to freq resp issues instead
that's simplified sch ; I'll not disclose too much if I say that Pa certainly isn't allowing DC through windings ;
patience please
DC? Bias flows from + to - supply. Only a small offset current "may" flow thru the primary...
ZM, oh mighty one
According to Bill Whitlock, I'm sure you recognise the name, sited in a post many posts ago, and likely the engineer involved in the F6 Xfrmr, says; "the distortion performance of ANY transformer is significantly improved when the driving source impedance is less than the dc resistance of the primary. However, little is gained below about 10% of the winding dc resistance."
So, my "150" ohm primaries are about 28 ohms DC. Ohms not turns, not L, not Z, and if those Jensens are the same, and they seem to be, we have not satisfied a design requirement Bill thought important??? do you have any comment on that idea. I know I stated that already but you responded to freq resp issues instead
moi already stated what's carved in stone
in post regarding crude test of Iskra repeaters ; their Rdc is 18 Ohms per winding (so around 9 ohms when primaries are paralleled )what can I say - some rules o' thumb are more critical than other ones
we always must look at broader picture - what's reflected impedance , is there DC across primary ...
I'm far of being xformer xpert , just trying to get why they're working so well - either driven with proper buffer or even ditto from DAC chip output
Last edited:
