please give me a hand with my circuit about electronic load~~

[tom1987]

thanks for your help!
i have working with the circuit for a long time,my company need a electric load to test some parameters of the transformers .
the voltage of this transformer secondary side is from 10v to 40v(rms),the current is 6A(max rms),so i design this circuit.


after adding some compensating components,it can work successfully,unfortunately the mosfet goes bad suddenly after working
for a while.i think the cause of this situation is loop compensation,but i can not control it at allnow.
please help me.
PS:mosfet IRFP4232,dac tlv5618,opamp op37

 

[dick_freebird]

That's a fairly traditional looking E-load, I'd begin with measuring
the frontside case temperature (the side furthest from the heat
sink) and check carefully whether the FET is rated for the steady
state power dissipation you're asking of it, SOA violations, etc.
You could look at the gate voltage behavior under conditions that
induce failure but I bet the destruct mechanism is simpler than
that.

240W RMS is not exactly mellow, and your line might have some
spike or something that punches out the FET (forget RMS - put
a 'scope to VIN+ and let sit there for a full shift in infinite envelope
mode and see what that worst case looks like). Especially in an
industrial environment when you're being asked to shoestring
test equipment, what else has your manufacturing management
skimped on? Maybe power quality?

 

[tom1987]

thanks for your tips.My e-load is tested at i=3a(peak),u=40v(peak),and the mosfet's Pd>300w,Vds(max)=250v,so i think SOA is not the main reason.i have many questions about the loop of this circuit.the Ciss of IRFP4232 is about 8nf,so the FET contributes a -3db low pass point at about 40k.I want to know what should i do to deal with it?

 

[dick_freebird]

Pdiss rated is at some specified case temp. That makes it all
about the heat sink. Your average soldering iron makes 400C
with 40 watts after all.

Since the input changes slowly (?) I don't think think you need
a very sporty control loop - in fact I expect the right answer
is a very slow one that will ignore any ripple and spikes and
keep the average current constant.

 

[tom1987]

thanks very much for your help
1: i put the FET together with the heat sink,and the e-load is working about 60c,so i have to believe some of FET's features have changed in this heat which is the
reason to explain the unexpected circuit oscillation.but it is out of my ability,i don't know detailed rule about FET which is working at linear region.if possible, can you can give me any manual such as application notes?
2: the output wave of the transformers is just sine wave,i just find the control loop can not deal with the rapid changes in the voltage such as the crossing point of sine wave,so i want to need a very sporty control loop to avoid waveform distortion,after all i need e-load work at CR mode.