Hello,
This is one of the best threads alive on the forum! Thanks for that.
I was wondering when someone pops out a jfet input cfa..
I´ve been doing similar kind of "rumble" with Dunlaps Krill ops and various different VGS:s. An interesting journey which will likely continue with new toys presented in this thread...
J
This is one of the best threads alive on the forum! Thanks for that.
I was wondering when someone pops out a jfet input cfa..
I´ve been doing similar kind of "rumble" with Dunlaps Krill ops and various different VGS:s. An interesting journey which will likely continue with new toys presented in this thread...
J
Ah, not so quick. Isnt snake oil a highly prized commodity in Japan?
In truth, a symmetrical jfet FE with folded cascode is a REALLY nice sounding amp. Also high bandwidth and pretty stable. I believe curl used no feedback and a servo, but a little feedback never hurt.
hee, hee
not trying to namedrop or start forest fires, just figured it would be easier to find a representative schematic if more info provided for a search.
it's a slight variation on comp-quad input + vas that works well and sounds good.
mlloyd1
not trying to namedrop or start forest fires, just figured it would be easier to find a representative schematic if more info provided for a search.
it's a slight variation on comp-quad input + vas that works well and sounds good.
mlloyd1
hee, hee
not trying to namedrop or start forest fires, just figured it would be easier to find a representative schematic if more info provided for a search.
it's a slight variation on comp-quad input + vas that works well and sounds good.
mlloyd1
Please post such a schema , I'll call it the "snakeyamp" IPS.
Guaranteed I'll make it work with this OPS , as long as I don't have to run
the VAS at 30ma .
OS
I don't like FET's , harder for timbuktoo builders to source , expensive ..
matched pairs are problematic.
BJT's like the 1015/1815 .... buy 10 each for 2$ , 2-3 good matches will be yours for under 5$
OS
Ah, not so quick. Isnt snake oil a highly prized commodity in Japan?
Awww ! Build this next one (below 1-2).
This "primitive" little collection on 6 BJT's is a 25ppm wonder !
Member Jkeutmann gave me the idea of dual trimmers (R5/R12)
for the dual CCS's.
This allows both VAS current adjustment and offset by a "unbalanced'
trimming.
The models (992/1845 or 1015/1815) show about .1 ma difference
between genders. This is also a result of the VAS gender mismatch.
Careful matching of BOTH the input pair/VAS will minimize this.
Since the input pair of this one does all the "work" , I'm looking for
a to-92L pair(the big ones) with 120 Vceo / low Cob/1A+ Ic to
do the job for the 2 larger OPS's (60-85V rails).
PS - this amp "floats " on the oil ....
OS
Attachments
Last edited:
ALL the circuits require at least DMM "halfar$e" matching.
Wolverine examples would be the CM/LTP ... you could trim the "offset" to
compensate for a gross mismatch ... but fidelity might suffer.
As these pairs are the same gender (NPN-CM and PNP-LTP) your
5$ 20 piece purchase would most likely produce mostly matches.
On the CFA , being totally dependent on that 1 pair at IP - both offset
and performance would suffer. P and N would be harder to match ...
the reason for the 2 trimmers !!
PS - these circuits will accept almost anything - they just have to be the same
"anything" ....
OS
I don't like FET's , harder for timbuktoo builders to source , expensive ..
matched pairs are problematic.
BJT's like the 1015/1815 .... buy 10 each for 2$ , 2-3 good matches will be yours for under 5$
OS
Not hard to source and expensive if someone already have it.
Fet IPS please, nothing fancy maybe (LS)J74/(LS)K170 (I bet buzzforb have a lot of them
) ? Diy Audio Store have stock. Or maybe noisier J103/K246 ?For Fet lover
Didiet
ok, here's two, assuming the posts both attach.
first is borbely's old servo 100 design with bipolar inputs & mosfet output.
second is hafler xl-280 (also borbely, i guess )with jfet inputs and mosfet outputs.
i circled the pseudo folded cascode on the hafler so you can see what i meant.
mlloyd1
first is borbely's old servo 100 design with bipolar inputs & mosfet output.
second is hafler xl-280 (also borbely, i guess
)with jfet inputs and mosfet outputs.i circled the pseudo folded cascode on the hafler so you can see what i meant.
mlloyd1
Please post such a schema , I'll call it the "snakeyamp" IPS.
Guaranteed I'll make it work with this OPS , as long as I don't have to run
the VAS at 30ma .
OS
Attachments
ok, here's two, assuming the posts both attach.
first is borbely's old servo 100 design with bipolar inputs & mosfet output.
second is hafler xl-280 (also borbely, i guess
i circled the pseudo folded cascode on the hafler so you can see what i meant.
mlloyd1
I downloaded the full hafler 280 schema
.. I will add this to the "spooky".Quite a hybrid .... leach front end /hafler VAS/H/K EF3 OPS !!
"spookysnakeamp"??
OS
OS, one problem with using hundred-watt drivers in an EF3 is that due to their large capacitance, if the amp oscillates, the predrivers will readily saturate into them. When this happens the predrivers act like rectifiers, pumping up output stage bias. I think this is very dangerous behavior.
One thing you could do to help with this is to give the predrivers at least 54mV of degeneration, this way they will never be more than 150% eager to drive the outputs, until rectification begins. This won't stop the rectifier action, but it should provide a degree of safety as long as oscillation is not severe. Adding reverse schottkeys across the emitters and degeneration will limit the predriver output current by overloading the VAS.
One thing you could do to help with this is to give the predrivers at least 54mV of degeneration, this way they will never be more than 150% eager to drive the outputs, until rectification begins. This won't stop the rectifier action, but it should provide a degree of safety as long as oscillation is not severe. Adding reverse schottkeys across the emitters and degeneration will limit the predriver output current by overloading the VAS.
Attachments
Hi Kevin,
I know this is a little off topic but I am curious. What does Eagle's Electrical Rule Check and Design Rule Check miss? It seems to pick up all my errors and then some.
regards