150A 1000V bjt based amp possible?

dadouzzu

Member
2007-12-10 10:28 am
Hello,
I found in a old ups 2 pair of npn power transistor.
I would build a quasi complementary class AB amplifier based on this transistor.
is it possible?

Fuji electric 2DI150D-100
Ic=150A, 300A (1ms)
Vce=1000V
Pd=1000W per transistor, isolated case
beta=100-3000.

Storage time = 15us with Ib1=3A , Ic=150A ib2=-9A.
are too slow?

thanks, sorry 4 my bad english
 

dadouzzu

Member
2007-12-10 10:28 am
hfe vary from 100 to 3000
there are 3 transistor in darlington configuration.
the termal resistance is 0.125°C/W

My problem is the storage time of 15us :xeye:
whith Ib=-9A!!!!!

is good for audio application?

with a 1ohm resistor from base to emitter, what time is necessary to turn off?
 

dadouzzu

Member
2007-12-10 10:28 am
p.s.i dont know the linearity because this is a switching product
i tryed to add datasheet but is too big (380k):whazzat:


if tstg of bjt is too high for audio app. i can dismantle an other ups, with a 1200V 300A 1800W igbt.
ton max = 1us
toff max = 1.5us

it's a good idea? igbt for audio?:xeye:
it is necessary Vge=10V to turn on this igbt...
it's possible to use a quasi complementary class ab configuration with this igbt?

thanks
 

fizzard

Member
2006-11-02 8:12 pm
Most larger motor drives operate below 3KHz.

With 1000 W disipatation you'll have a very small safe operating area relative to the current and voltage.

If it's BJT it should have alright linearity, even the switching ones are okay.

IGBT's could work. I used one as a pass device in a linear regulator. As far as I could tell it behaved just like a MOSFET. I've never seen them in audio though, except switching power supply.
 

dadouzzu

Member
2007-12-10 10:28 am
fizzard said:
Most larger motor drives operate below 3KHz..

With 1000 W disipatation you'll have a very small safe operating area relative to the current and voltage.

If it's BJT it should have alright linearity, even the switching ones are okay.

IGBT's could work. I used one as a pass device in a linear regulator. As far as I could tell it behaved just like a MOSFET. I've never seen them in audio though, except switching power supply.

I think in original application these bjt work at 100hz (square wave output U.P.S.)
how can i calculate turn off time (tstg+tf) with Rbe=1ohm?
if i dont use Rbe there are any problem for audio application?

thanks
 
djk said:
The semiconductors are the least expensive portion of the total cost of an amplifier.
by far.
the output stage should account for no more than 10% of the total cost and more likely <5%.
the voltage amp and PCB accounts for another 5 to 10%.
the heatsink for about 20 to 30%.
the transformer and PSU about 30 to 40%.
the case and hardware another 20 to 40%.

An expensive case could easily double the cost of a project.
 

dadouzzu

Member
2007-12-10 10:28 am
AndrewT said:
by far.
the output stage should account for no more than 10% of the total cost and more likely <5%.
the voltage amp and PCB accounts for another 5 to 10%.
the heatsink for about 20 to 30%.
the transformer and PSU about 30 to 40%.
the case and hardware another 20 to 40%.

An expensive case could easily double the cost of a project.

i have output transistor
i can build the pcb, optically (photoresist) or mechanically(CNC)
I have heatsink
I have 1.5 kw (50hz ) toroidal trasformer and i woul modify windings for correct output voltage (+/- 100V), 4 home appl;
or dsPIC with high & lo side driver to drive power mosfet or igbt (i have) to drive HF transformer (to buy) winded with litz wire, 4 car application.
I can build ( cheap) case.

anyway if i study,project & build my own amplifier, is 4 my personal satisfaction.
Build an amlifier based on other person schematic brings me no satisfaction & the cost is too hight compared with commercial product.

again sorry 4 my bad english. L.


p.s. in a "infinity" amplifier schematics i found tip35/36 complementary bjt.
These bjt have a storage time = 3us with Ib=-1/10Ic
My problem is:
how can i calculate turn off time with Rbe = 1ohm?
how turn off time influence audio response in audio amplifiers?

thanks
 

wg_ski

Member
2007-10-10 5:21 pm
dadouzzu said:


i have output transistor
i can build the pcb, optically (photoresist) or mechanically(CNC)
I have heatsink
I have 1.5 kw (50hz ) toroidal trasformer and i woul modify windings for correct output voltage (+/- 100V), 4 home appl;
or dsPIC with high & lo side driver to drive power mosfet or igbt (i have) to drive HF transformer (to buy) winded with litz wire, 4 car application.
I can build ( cheap) case.

thanks

Then you have everything expensive you need. Except the output transistors. The ones you have just won't do. If it were any of us, we would buy 16 pairs of MJ21193 and 21194 at a whopping cost of around $140 (you can't even buy an EP2500 for that) for the whole works and use that tranny and heatsink you have. You don't want to copy someone else's design? Fine. It will be easier to design yourself using the correct parts than unsuitable ones.
 

dadouzzu

Member
2007-12-10 10:28 am
megajocke said:
Can you show us the SOA graph for the transistors? If you don't have it/can't find it, the storage time is the least of your problems. Like djk already said, it is a switching part, probably with no DC SOA.

yes, switching product.

I think there are sufficient SOA for +/-100V application.

with 1ms pulse second break down curve start over 300V

DC, infinite heatsink:

150A until 6V

-20Db/dec until 300V (I think) 3A

-40Db/dec until 800V

termal time constant is about 0.1s :cool:

with sink at 85°C SOA is about 1/2

there are no need of termal pad.
Commercial pad have termal resistance very high for to3 case.
at last 0.5°K/W