Expert on MOSFETs
Archiver
Posts: 46,084
Hi Pat M,
I lost your email, but here is my response. I see two
potentially fatal problems with what you described.
The first is the inductive kickback from the motor.
Even though most modern MOSFETs have an intrinsic
(internal) protection diode, there can be massive
inductive spikes which will kill your MOSFET. You will
need two fast-recovery diodes - rated about the same
as your motor current - to protect your MOSFET. One
will connect accross your motor (reverse biased, of
course (line (cathode) towards positive)) and the
second across Drain to Source on your MOSFET also RB.
This will "clamp" the reverse voltages and protect
everything.
The second problem I see is driving the gate of the
MOSFET. There are two possible issues here. The first
is the voltage level of the FET. Most MOSFETs require
about 12 volts Gate-to-Source to fully saturate. While
it may work a light loads, when pulling amps, it won't
saturate and could go due to heat. There are
logic-level FETs that will fully saturate at just 5
volts, but I'm not sure if that's what you have. You
could add a buffer to do the level translation. Next
is the very famous oscillations. The silent FET
killer. With FETs high imput impeadance, the are prone
to oscillate in the RF range. Make sure you have a 47
ohm resistor from the pin of the Stamp to the Gate of
the FET. You may also want to try a 0.01uF cap from
Gate-to-Source to present a load at high frequencies.
Last but not least, use a heatsink. It doesn't take
much to help.
DD
__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more
http://taxes.yahoo.com/
I lost your email, but here is my response. I see two
potentially fatal problems with what you described.
The first is the inductive kickback from the motor.
Even though most modern MOSFETs have an intrinsic
(internal) protection diode, there can be massive
inductive spikes which will kill your MOSFET. You will
need two fast-recovery diodes - rated about the same
as your motor current - to protect your MOSFET. One
will connect accross your motor (reverse biased, of
course (line (cathode) towards positive)) and the
second across Drain to Source on your MOSFET also RB.
This will "clamp" the reverse voltages and protect
everything.
The second problem I see is driving the gate of the
MOSFET. There are two possible issues here. The first
is the voltage level of the FET. Most MOSFETs require
about 12 volts Gate-to-Source to fully saturate. While
it may work a light loads, when pulling amps, it won't
saturate and could go due to heat. There are
logic-level FETs that will fully saturate at just 5
volts, but I'm not sure if that's what you have. You
could add a buffer to do the level translation. Next
is the very famous oscillations. The silent FET
killer. With FETs high imput impeadance, the are prone
to oscillate in the RF range. Make sure you have a 47
ohm resistor from the pin of the Stamp to the Gate of
the FET. You may also want to try a 0.01uF cap from
Gate-to-Source to present a load at high frequencies.
Last but not least, use a heatsink. It doesn't take
much to help.
DD
__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more
http://taxes.yahoo.com/
Comments
Original Message
From: PH [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=oEGdMjj1-SNYh6R3jBFE3Qhi0NnN2_m8K2Vs5ToQLxG1gfhGEVuHaB4-AfzW6G9a7p1ABS1ApsWGKAljkaMR]drdiode2002@y...[/url
Sent: Tuesday, February 25, 2003 7:45 PM
To: basicstamps@yahoogroups.com
Subject: [noparse][[/noparse]basicstamps] Expert on MOSFETs
> Even though most modern MOSFETs have an intrinsic (internal)
protection diode, there can be massive inductive spikes which will kill
your MOSFET.
We have used Mosfets in pulsed applications for years and ALWAYS use an
external diode to protect them. The intrinsic diode is somewhat
delicate. We have used (sometimes stacks of) relatively fast diodes
like the EGP10G or FE6D for this application, as the fastest diode will
cut on first<G>. If the fastest diode is in the transistor, there is no
protection<G>. This has been adequate for us (so far).
>The second problem I see is driving the gate of the MOSFET. There are
two possible issues here. The first is the voltage level of the FET.
Most MOSFETs require about 12 volts Gate-to-Source to fully saturate.
This is extremely good advice. It's good to look at the data sheets.
Some FETS only take +/- 15 Volts gate to source. A simple 2N2222 and a
1 kOhm resistor to a higher Voltage makes a good driver if you are not
after speed.
> Next is the very famous oscillations. The silent FET killer. With FETs
high imput impeadance, the are prone to oscillate in the RF range. Make
sure you have a 47 ohm resistor from the pin of the Stamp to the Gate of
the FET. You may also want to try a 0.01uF cap from Gate-to-Source to
present a load at high frequencies.
We use a 10 Ohm resistor and a ferrite bead right on the gate of the
transistor. But we are looking for speed. The other fella's version is
very good.
While a circuit may not be oscillating now, it may later when you aren't
watching it. We lost 5 parallel 200 Amp (each) fets yesterday due to
this fact. That or something else changed and we didn't see it<G>.
Woof