Refer to post #264.
Zen Mod, I tried this circuit all over again with new home etched PCB and exact values, parts etc., just two days ago. The amp biased at 2.5 Amps, measured with an Ammeter in the positive rail.
Sounded forward and a bit dry, compared to M2. I also felt that it was a bit compressed on dynamics compared to M2 output stage. Just my two cent. Thanks.
Last edited:
Dear Generg, I don't know what happened, I just found that (after upgrading to LtSpiceVII) my statement in #175 is only correct on inverting systems (fundamental phase close to 180 deg.). On non-inverting system with (fundmental phase close to 0 deg.), a positive 90 phase on second harmonic like your left picture on #174 corresponds to negative phase. You can confirm your system with fundamental null tool jcx shared at Spice simulation post #361.
Sorry ...
Honney , I created The Monster!
Well , it works .
It's really easy to work on newly made test bed (T-Bed , it deserves own name) ...
except one tiny change (value of resistor in parallel with opto-diode) and surprising fact that is not normal inserting mosfets on pcb , while PSU is fully powered on (damn phone distraction)
, really everything is as prescribed on back of (LTSpice) napkin 
'morrow mounting Edcors (both Jensen and Cinemag pinouts present on pcb ), wiring another channel ,mounting lazy 12cm fan on heatsink face ....... then struggle with EMU1212 setting for FFT measurements
maybe I will enjoy in listening few days , then measurements
Iq is rock-on stable during extended power on-of cycling , DC offset too
Iq monitored on 0R11 , part of CRC ........
Well , it works .
It's really easy to work on newly made test bed (T-Bed , it deserves own name) ...
except one tiny change (value of resistor in parallel with opto-diode) and surprising fact that is not normal inserting mosfets on pcb , while PSU is fully powered on (damn phone distraction)
, really everything is as prescribed on back of (LTSpice) napkin 'morrow mounting Edcors (both Jensen and Cinemag pinouts present on pcb ), wiring another channel ,mounting lazy 12cm fan on heatsink face ....... then struggle with EMU1212 setting for FFT measurements
maybe I will enjoy in listening few days , then measurements
Iq is rock-on stable during extended power on-of cycling , DC offset too
Iq monitored on 0R11 , part of CRC ........
Attachments
Babelfish XA25 is in making , pcbs populated and everything put on shelf ..... waiting
however , as zillion man and machine hours sez - I rejected Hall chips sensing for pucks - there is simply not enough gain in bias servo , when made with hall chips ditto to optocoupler(s) , so I resorted to current mirror scheme (my best shot posted in Beats thread)
yeah , I believe it is possible making bstrd of M2 and Beast - buffer+autoformer feeding adequately biased pucks ....... but somehow ugly capacitances of pucks are rearing a head ........ is it buffer+autoformer too weak , that's question only bench test can answer
btw. there are just two hall sensor chips - one in each rail
however , as zillion man and machine hours sez - I rejected Hall chips sensing for pucks - there is simply not enough gain in bias servo , when made with hall chips ditto to optocoupler(s) , so I resorted to current mirror scheme (my best shot posted in Beats thread)
yeah , I believe it is possible making bstrd of M2 and Beast - buffer+autoformer feeding adequately biased pucks ....... but somehow ugly capacitances of pucks are rearing a head ........ is it buffer+autoformer too weak , that's question only bench test can answer
btw. there are just two hall sensor chips - one in each rail
The one aspect that scares me from using the pucks is the uber high gm. I'm getting some scary spikes in Ltspice on CFP topology. Both Hugh Dean and Paul Bysouth also reported fatal intermittent oscillation on high gm Fairchild parts and recommend the use of drain to gate RC snubber. Still patiently waiting for more info from guys with better equipments.
it works, and it sounds reallllly good
took just one fast and dirty phone pic
even connected few days ago finished final iteration specimen of Iron Pumpkin SE , but didn't took that picture
was tired , wasted at least 3 hours of work trying to solve ookup with right channel - replaced two Hall chips , opto ......... scratched my head from all possible and impossible directions ....... only to found one bloody 1N4148 , taken from ancient box (never again!!!) had wrongly labeled stripe
anyway , life is good ;
two good Dreks finally made/confirmed , in 3 days
took just one fast and dirty phone pic
even connected few days ago finished final iteration specimen of Iron Pumpkin SE , but didn't took that picture
was tired , wasted at least 3 hours of work trying to solve ookup with right channel - replaced two Hall chips , opto ......... scratched my head from all possible and impossible directions ....... only to found one bloody 1N4148 , taken from ancient box (never again!!!) had wrongly labeled stripe
anyway , life is good ;
two good Dreks finally made/confirmed , in 3 days
Attachments
Last edited:
