Lateral mosfet driving question - Hitachi amplifiers

[fojica]

Hi all,

Recently I've looked through Hitachi amplifiers, their mid-to-high range from '80s, HA-xxxx and HMA-xxxx.
What I found intriguing is that in their "bigger models" they've used small capacitors in parallel gate resistances of output mosfets. The "exceptions" are HMA-6500, where there are no caps across the gate resistors and HA-8700, where on 2SJ49's in only one gate resistance with 1nF parallel cap - for both mosfets. I've checked also parts list for HA-8700, indeed only one cap is listed on "negative side", for 2SJ49.
I could not figure out the exact reason for doing this way - ok, cap across the resistor can help maybe a bit at high frequencies, but it's not worse from stability point of view ? From what I know reducing gate impedance with raising frequency would not be the best thing to do for a mosfet if you want to avoid oscillations, generally speaking.
Below are some shots from schematics described.
In case of using caps across gate resistors as Hitachi did - how to check for oscillations without a scope (mosfets temperature, zobel burning) ?
Hitachi mosfets are almost unobtainable now - I'm wondering if can we apply this "trick" with the C||R for current production Exicon/Semelab/Alfet/Renesas mosfets (single-die, not dual-die) ?
Any comments and suggestions are welcome.

Thanks,
Tibi

[tommy1000]

I think that the gate capacitors eliminates stray inductance since the TO-3 power mosfets are not mounted on the PCB but instead wired to the PCB. They may also compensate for the different gate capacitance of the P and N mosfet.

Hope this helps.

[jez]

It's because the P and N channel mosfets have vastly different source to gate capacitances.. It's to try and equal them.

[fojica]

thanks for answers guys
I think that elimination of stray inductance could be the main reason, because for Cgs equalization I've seen caps added on N-ch between G and S...P-ch have anyway a bigger Cgs, I see no reason to add another cap just to increase it further.
Also please note that gate resistors are different values for N-ch and P-ch...this would be the second method in quest for equal "slew rate" between N-ch and P-ch...also mentioned on Exicon app notes.

regards,
Tibi

[MiiB]

does a cap in series increase capacitance..???

[fojica]

@MiiB
my guess is no...we have caps in both N-ch & P-ch, usually bigger for P-ch (also bigger Cgs). but there is also an "exception", although cap value doesn't appear in the attach: for HA-7700 both caps are equal, 100pF (C726L & C727L) - indeed, for equal gate resistors

[MiiB]

my idea is that the 100pF cap helps secure equal drive also at higher frequencies...similar to different gate stoppers at the P and N channel.

[fojica]

well...I've tried adding 33pF for N-ch and 47pF for P-ch...no accoustical change spotted so far.
recently I've seen another idea on some Albs schematic....some kind of buffer-resistor between VAS (drivers) and usual gate resistors (attached). R18 and R19 are 8.2 to 10ohm, R20-R25=100ohm. Output FETs are 2SK134/2SJ49, T3-T6 are MJE340/350.
any opinions on this ?

regards,
Tibi

[AndrewT]

R18 & R19 are to help prevent the Zeners blowing up the driver stage with all it's carefully matched components that took many resources to work well.

[fojica]

I didn't took into account this way but it's good to have learned it
thank you, AndrewT