schematic and bias set up info for trace elliot AH350SMX MOSFET power amp

[bassman 59]

i have trace elliot gpsmx head,thats what it says on front but rear is AH350SMX confusing to say the least
can find nothing schematic wise that indicates the use of hitachi 2sk1058 and sj162 mosfets,i need to know
bias set up,and what else is likely to of failed in the driver stage,does the bias control transistor generally fail it
seems to make sense in changing the electrolytics,i have spoken to someone who has had many years experience
with these and the awful bipolar b as he refered,does anyone have any experience of instability probs as i am informed
a major issue with the vers i have here thanks steve

[mozwell]

I hope this helps
You may not have much getting these mosfets any more

[bassman 59]

thanks mozwell i have found this a real challenge,the
mosfets in question dont seem to show on any schems,
i assume they were superseded, and going by what i have
found out so far,this does appear the case,ex i have am
informed is early vers,there must be schematics for this
exact one,maybe the later with alt fets biases up similar
regards steve

[J M Fahey]

1) looks confusing but it has an explanation: T.E. has 2 very popular preamps with 7 and 12 band EQ , and coupling one of these with half a dozen power amps, they ofer different options.
In your case what´s important is that it´s the AH350H power section, which seems to be a biampable 150W+300W power amp or thereabouts and is on page 5 of Mozwell's .PDF .

2) besides what happened to others, what is the actual problem in *your* amp?

3) in case you need them, there are modern versions equivalent to those MosFets at reasonable price.

4) good luck

[bassman 59]

thanks JMF i have failed fets x 2 one is shorted totally did not blow
apart like sometimes happens,my research so far,and speaking with one
in particular that has worked on and (scrapped) a fair number he informed
me,is that they get very unstable,i assume that is the driver section,so my
thoughts are to renew caps transistors in that cct,only other thing that concerns
me re the 2sk1058 j162 config in this power amp is what the drain current level
should be finally set to,another question has anyone had success with alternate
fets in this trace elliot output stage cheers steve

[J M Fahey]

Profusion PLC saved your life.

They sell better specs MosFets, fully compatible, for 10 U$A each, and sell direct to public.

NEC/Renesas 2SK1058/2SJ162 are 160V , 7 Ampere, 100W parts , in TO3P case

EXICON (sold by Profusion) EXC10N20R/10P20R are 200V, 8 amperes, 125W parts, same pinout, in compatible TO247 case.
Lateral Mosfet

So recheck originals, I would replace all same side even if one is still alive to ensure they share load same evenly , set that bias pot to zero, and start the amp plugged into a lamp bulb limiter, to avoid expensive mistakes.

If everything is fine, I wouls slowly raise bias until I get some 25mA per MosFet.

If everything´s fine, test it without the limiter and then consider it repaired.

ECX10P20R

[bassman 59]

hello JMF thank you for your very valued help with my trace ah350
i have tried to ref the schematic but it differs from what i actually
have,the component ref no's start with for instance tr400 plus as ex
the board i have is stamped with low numbers r13 nxt to trimmer
pot is burned out and cannot read value,on the rear of this board
etched PB300 cb issue 3B this is not a biamp board has four fets
where the other vers has six if yourself or anyone else can help that
would be great.

[J M Fahey]

You can do this:
1) I think yours is the non biamp version then, so it's just the Bass side shown, obviously doing full range.
Schematic must be basicalley the same, since it's a "textbook" application, pulled from the original Toshgibab datasheet.

So I suggest you print that sheet large, if possible in 2 sheets that you tape together, check that my theory is correct (of course may not be) and in that case go over the printed schematic, pencil in hand, crossing each label and writing the correct (matching the actual PCB one).

Once double checked, size the original .pdf so the useful part fills the screen, capture it (shift + Prt Scr) and paste it in any graphics program, label all parts as in the PCB, save it as a .gif or .png and post it here.

To begin with, you will find the blackened resistor value by yourself, and we can suggest correct part ID to check if necessary.

Of course, if you have some Adobe .pdf editor, you can edit the original schematic saving some steps, but if not, what I suggested works anywhere.

[bassman 59]

JMF thanks for your help with my traceAH350 issue,i fitted the NEC/REN plastic encapsulation M/fets,
maybe a wrong move,the top hat 135/50 in hindsight it seems would of been a better option.
so now it is time to see if it all works,BTW R13 the burned out resistor was 220n,and further to
this confusing situation,the cct schematic that applies to what i have,is the 250 watt four fet
pb200 series board,yet on the one i have it clearly states 300,i am ready to set this up now,
here is procedure i am set to follow,please correct me if wrong,lamp on supply which is variac,
disconnect drain of one fet,put dvm in series set on 300m/a range,monitor current draw and
voltage with second dvm,turning up slowly on variac until i get approx 25 ma,once set inject
1khz into power board for crossover cancellation adj,i may not have this totally correct,and will
not proceed until i have your advice,as it is clear to me know the deal with this.
i would think this thread i have started is going to be very valuable to many engineers,and to
of found one as clued up as yourself is brilliant.