D
Deleted member 550749
The ones on stock at Mouser have different hfe classification, KSA1381ES vs KSC3503DS
Oh yeah , 100mhz/20W -SOA looks the same. So , even with a smaller package ... they might be good drivers for a single pair outputs..
The 2SC4793 with it's larger package lists as 2/20W vs 1.5/10W (TTC011).
2-3 pairs OP , I'd still go with MJE15032 (8A/50W).
If you can find the same binning ... they are the very best VAS devices (or pre-drivers).
OS
The 2SC4793 with it's larger package lists as 2/20W vs 1.5/10W (TTC011).
2-3 pairs OP , I'd still go with MJE15032 (8A/50W).
No , If you have them. I can't order them , and won't pay more $$ or fall prey to Ebay fakes.
If you can find the same binning ... they are the very best VAS devices (or pre-drivers).
OS
Protection , Protection.
Time to design the 2 "business ends" of the Arcwelder protection scheme.
- Arcwelder soft-start. Would be nice to to have all AC (and only AC) on it's own board.
A little education before we commit to a PCB - https://neurochrome.com/pages/the-ultimate-guide-to-soft-start-design and ...
EMI filters - https://www.electronicproducts.com/all-about-emi-filters/ .
Might as well add anything "ANAL" to the design. We already will just have (+5/+12/ ground) and 3 control signals... that would be inrush/main relay
and AC sense. Just a single 6 pin connector to interface with the control unit.
Mount this board right at the rear panel (AC plug) ... all by itself.
- Arcwelder SS relay , Jwilhelm's circuit is perfect. I used it before - perfect ! 2 connectors , one for the rail reference and one for the control signals.
All that remains is the (digital) control signal PCB with the Arduino NANO. I debated on how to split this circuit up. Inspiration of actually
building a setup like this and looking at PASS amps made me come to this conclusion.
Also , if I want future "extra OCD" (ESP32) , I can just re-design a more sophisticated digital board , all the "dumb linear interfaces" remain on the existing PCB's.
OS
Time to design the 2 "business ends" of the Arcwelder protection scheme.
- Arcwelder soft-start. Would be nice to to have all AC (and only AC) on it's own board.
A little education before we commit to a PCB - https://neurochrome.com/pages/the-ultimate-guide-to-soft-start-design and ...
EMI filters - https://www.electronicproducts.com/all-about-emi-filters/ .
Might as well add anything "ANAL" to the design. We already will just have (+5/+12/ ground) and 3 control signals... that would be inrush/main relay
and AC sense. Just a single 6 pin connector to interface with the control unit.
Mount this board right at the rear panel (AC plug) ... all by itself.
- Arcwelder SS relay , Jwilhelm's circuit is perfect. I used it before - perfect ! 2 connectors , one for the rail reference and one for the control signals.
All that remains is the (digital) control signal PCB with the Arduino NANO. I debated on how to split this circuit up. Inspiration of actually
building a setup like this and looking at PASS amps made me come to this conclusion.
Also , if I want future "extra OCD" (ESP32) , I can just re-design a more sophisticated digital board , all the "dumb linear interfaces" remain on the existing PCB's.
OS
Attachments
Last edited:
Oh , by the way - I'm about to build ... did not abandon this project !
Working on the rest of the protection system (below) and
I will post the SYMASUI , EASYAMP TH , and the INFIDEL
- 5 IPS's.
I want to have a complete setup for a complete OEM+ quality amp . NO sloppy DIY "kit".
Working on the rest of the protection system (below) and
I will post the SYMASUI , EASYAMP TH , and the INFIDEL
- 5 IPS's.
I want to have a complete setup for a complete OEM+ quality amp . NO sloppy DIY "kit".
Attachments
Symasui will be next... (just a BOM). Interest of the Maplin (I did not know) and Hitachi topology has surfaced recently.
I just thought it was the "symasym" DIYA project that I was familiar with.
Maplin/Hitachi were real simple like the "HH" English V600-800 LATFET models. (below)
This seem to be a "lazy" design (that works).
Just a few more parts equals something that could match or BEAT the mighty Wolverine.
10K is 1/2 ppm , gain is 80 DB (below 3/4)
1K gain is 120db with PPB (billion'th) , can't really even measure it with the simulator ... 0 % ??
I'm going to order "Hellraiser" CFA as my first choice boards. I remember them from slewmaster.
They had such a crystal clear sound. Not really bright , but very detailed.
In the BASS department , CFA feeds back EMP to the emitters of the negative FB point , correcting for the speaker (big woofer)...
These "Symasui's" , just to export a couple to the Wolverine people . My test setups could not measure
this design. Wolverine team even runs up on the DAC/noise floor limitation. These should be the same.
They DO clip better (last pix) , these are just the Wolverine input stage with a balanced differential VAS - perfect balance.
OS
I just thought it was the "symasym" DIYA project that I was familiar with.
Maplin/Hitachi were real simple like the "HH" English V600-800 LATFET models. (below)
This seem to be a "lazy" design (that works).
Just a few more parts equals something that could match or BEAT the mighty Wolverine.
10K is 1/2 ppm , gain is 80 DB (below 3/4)
1K gain is 120db with PPB (billion'th) , can't really even measure it with the simulator ... 0 % ??
I'm going to order "Hellraiser" CFA as my first choice boards. I remember them from slewmaster.
They had such a crystal clear sound. Not really bright , but very detailed.
In the BASS department , CFA feeds back EMP to the emitters of the negative FB point , correcting for the speaker (big woofer)...
These "Symasui's" , just to export a couple to the Wolverine people . My test setups could not measure
this design. Wolverine team even runs up on the DAC/noise floor limitation. These should be the same.
They DO clip better (last pix) , these are just the Wolverine input stage with a balanced differential VAS - perfect balance.
OS
Attachments
Yeah , If I ran either a Wolverine/Symasui IPS through a BD139/140 class A headphone pre-amp/headphone output stage ....
I could really blow away any ACP+ , match that with "Jung super regulator" +/- 15V supplies wow !! -150thd plus -150 PSRR.
Edit - that is the Symasui test - .1V (+10db like a preamp) 2nd harmonic down @ 160db 1K ... noise floor territory -
OS
I could really blow away any ACP+ , match that with "Jung super regulator" +/- 15V supplies wow !! -150thd plus -150 PSRR.
Edit - that is the Symasui test - .1V (+10db like a preamp) 2nd harmonic down @ 160db 1K ... noise floor territory -
OS
Last edited:
Just a BD139/140 pair biased into class A with 22R emitter resistors. Then you feedback 4.7K (instead of 22-27K) to NFB.
You would want the class A the be of the current regulating variety with a touch of thermal feedback.
Not exactly like the below schema , but generally a diamond buffer (fast) based with some form of thermal FB.
(op-amp) is replaced with the IPS.
Edit - I might add a dual line-stage to the Arcwelder family - plug in some symasui's.
OS
Attachments
Just a pair LM7824/LM7924 with ample decoupling and a little 24-0-24 trafo and either the spook or hellraiser IPS
would typically be -120db+ separation and -135db THD + noise.
Would easily beat anything out there ... you would have to spend 10X the money just to (be able) to test it.
OS
would typically be -120db+ separation and -135db THD + noise.
Would easily beat anything out there ... you would have to spend 10X the money just to (be able) to test it.
OS
Attachments
The apt 1 was the inspiration for my folded/Hawksford cascode (with the leds) that is on the spooky and hellraiser,
APT 1 just uses fixed diodes with no local FB loop - it works but is inferior.
The APT 1 also uses a level shifter to achieve a symmetric VAS , the hitachi (symasui method) is "alien tech" compared
to that.
Still , the APT 1 was one of the topologies that i worked from , it was also a triple EF ... yummi.
OS
APT 1 just uses fixed diodes with no local FB loop - it works but is inferior.
The APT 1 also uses a level shifter to achieve a symmetric VAS , the hitachi (symasui method) is "alien tech" compared
to that.
Still , the APT 1 was one of the topologies that i worked from , it was also a triple EF ... yummi.
OS