The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: firstname.lastname@example.org (Bill Allison)
Subject: Re: Help - Power mosfets - difficult load
Date: 22 Oct 2002 00:28:03 -0700
References: <email@example.com> <3DB2C5F9.C89A683B@rica.net> <firstname.lastname@example.org> <3DB47CA5.51C80BFA@rica.net>
NNTP-Posting-Date: 22 Oct 2002 07:28:03 GMT
John Popelish wrote in message news:<3DB47CA5.51C80BFA@rica.net>...
> Bill Allison wrote:
> > John
> > Thanks for your input - responses below...
> > John Popelish wrote in message news:<3DB2C5F9.C89A683B@rica.net>...
> > >
> > > The full blown way to do this whole process is to also sense the motor
> > > average current, take the square root of it (since the torque of a
> > > series wound motor is proportional to the square of current), and
> > > control that value in a closed loop, whose setpoint is the output of a
> > > force control loop based on the line tension measurement.
> Correction, here. Since the torque is about proportional to the
> square of current, you want ot measure the current and square that
> measurement to provide a torque feedback to implement a torque
> > > Then you
> > > tune both PIC controllers (the amplifier that forces the measured
> > > variable to equal the setpoint value) so that both loops operate
> > > stably.
> > >
> > > But back to your fet concerns: I agree that the lead frame on TO220
> > > devices are a bit svelte for 75 amps. At those peak currents, I would
> > > probably go for a heavier package. but stud mounted devices aree impression
> > > expensive and hard to find. I R makes some high current fets in the
> > > SO-227 (big flat pack with screw terminals) rated for 180 ampsin it's
> > > continuous (and 100 volts) but cost $35. Before spending that, would look at devices in the TO-247 (bigger than the TO-220, but
> > > similar) package. For instance, the IRFP064N is rated for 98 amperes
> > > continuous and 55 volts @ $13.20 from Digikey.
> > Thanks John - I'll look into the above
> I got those prices from http://www.digikey.com
> > > You will also need a healthy schottky diode rated for similar current
> > > to act as the clamp across the motor when the fet turns off (unless
> > > you use a second fet as a synchronous rectifier). For instance, the
> > > 175BGQ030 in what they call a PowIRTab package (a TO-220 with the
> > > three leads merged into a sheet of metal) is rated for 175 amperes
> > > continuous at 30 reverse volts and goes for about $6. You will have a
> > > hard time blowing that on a brief overload.
> > Now I suspect i'm going to show my ignorance / inexperience... I was
> > under the impression (from the IRF app notes / spec sheets) that the
> > mosfet's integral avalanche capability via the p-n diode which is
> > inherent in it's construction was capable of safely passing approx the
> > same current as rated in the forward direction?
> The fet is in series with the motor, but the schottky diode should be
> in parallel with the motor, to carry the current when the fet opens.
> This allows the inductance of the motor to act as a current fly wheel
> to average the on voltage and off voltage over a pulse period.
> Without the parallel diode, the motor inductance will produce large
> voltage across the fet and destroy it. The fet body diode is no help
> here, since the motor inductive voltage adds to the battery voltage,
> rather than reversing it.
Ah - right, I'm with you now. Thanks for the link.
> Take a look at the standard buck switching regulator (the motor fills
> in as both the regulator inductor and load resistance):
> You will also have to pay attention to filtering the ripple current on
> the battery to keep noise and losses low.
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup