Motor Control needs a snubber??

[Fred]

I'm using a DPDT relay and a fet to activate a small PM DC motor. The
fet provides the ground and the DPDT swaps the polarity so the motor can
run in either direction. The motor can draw about 7A full load and it's
about 1 1/2 ft distant from the control PCB.

The Problem: I'm plowing fets and TVS diodes and can't seem to get a
handle on the cause as it only happens every once in a while.

See attached Schemo. Q1 provides the DC ground (or not) under uP
control. Q2 swaps the polarity. D1 is a dual polarity TVS diode
that is located on the PCB rim across the motor cable to snub ransients.
D2 snubs transients at the fet Q1 drain. D3 is an alternate path for
motor current to go beside thru D2.

Failure modes:
1) D1 pops short. The battery is shorted to the drain of Q1. When Q1 is
activated the fet pops violently.
2) Q1 fails but D1 is still in tact.

Note: D2 & D3 never have failed.

It appears that the motor transients are too powerful for D1 or perhaps
a snubber circuit is necessary across the motor. Does anyone have a link
to an article or book that would help me resolve this problem.

TIA,

Fred

 

[James Meyer]

On Tue, 26 Oct 2004 16:47:49 -0500, Fred <fps@pmr-iowa.com> wroth:

Failure modes:
1) D1 pops short. The battery is shorted to the drain of Q1. When Q1 is
activated the fet pops violently.
2) Q1 fails but D1 is still in tact.

Note: D2 & D3 never have failed.

It appears that the motor transients are too powerful for D1 or perhaps
a snubber circuit is necessary across the motor. Does anyone have a link
to an article or book that would help me resolve this problem.

TIA,

Fred

D1 is doing nothing in the way of protection. It only fails and then
takes Q1 out with it. Get rid of D1.

The contacts in a DPDT relay do not always open and close like you think
they should. Sometimes one pair can switch before the other. When that
happens, you can get a short directly from the 24 Volt supply to the drain of Q1
with nothing to limit the current. The relay isn't failing because Q1 acts like
a fuse to protect it. Use the micro to insert a delay interval into the signal
to the gate of Q1 every time the direction command changes to give the relay
time to complete its switchover. Turn Q1 off for an interval before you change
direction and keep it turned off for a while after you change direction.

Jim

 

[Fred]

James Meyer wrote:

On Tue, 26 Oct 2004 16:47:49 -0500, Fred <fps@pmr-iowa.com> wroth:


Failure modes:
1) D1 pops short. The battery is shorted to the drain of Q1. When Q1 is
activated the fet pops violently.
2) Q1 fails but D1 is still in tact.

Note: D2 & D3 never have failed.

It appears that the motor transients are too powerful for D1 or perhaps
a snubber circuit is necessary across the motor. Does anyone have a link
to an article or book that would help me resolve this problem.

TIA,

Fred


D1 is doing nothing in the way of protection. It only fails and then
takes Q1 out with it. Get rid of D1.

The contacts in a DPDT relay do not always open and close like you think
they should. Sometimes one pair can switch before the other. When that
happens, you can get a short directly from the 24 Volt supply to the drain of Q1
with nothing to limit the current. The relay isn't failing because Q1 acts like
a fuse to protect it. Use the micro to insert a delay interval into the signal
to the gate of Q1 every time the direction command changes to give the relay
time to complete its switchover. Turn Q1 off for an interval before you change
direction and keep it turned off for a while after you change direction.

Jim


Hi James,

Thanks for the reply. Always nice to bounce ideas off a second party.

I think your right. D1 will be removed.

There is delay of about 120 ms (worst case) between relay actions. I'm
lenthening that to
1/4 sec and in the case of a relay direction change that happened in the
very near term, I'll
throw in an extra 1/4 sec.

The troubling part is that the fet sometimes goes out all on it's own.
It could be overheating that causes
this, say if the relay sticks as you mention below, but I have the
nagging suspicion that a transient is getting in
to thedrain of the fet. Maybe if I remove D1 and replace with a snubber
to cut the rate of dv/dt at the motor
cable entry to the PCB, that would give some added protection.

I'm doing a search for snubber design literature this AM.

Thanks again.

Fred


----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= East/West-Coast Server Farms - Total Privacy via Encryption =---

 

[James Meyer]

On Thu, 28 Oct 2004 11:01:52 +1300, Terry Given <my_name@ieee.org> wroth:

not only does D1 not protect anything, when the relay is energised D1
"shorts out" the DC motor - its anode is +24V, the cathode is connected
to FET Q1.

Dang! I missed that one. The motor is supposed to run in *BOTH*
directions which means that the polarity of the applied voltage reverses too.
Any sort of unidirectional diode for D1 is entirely wrong. Perfectly OK when
the motor runs in one direction and a disaster when the direction changes.

Good catch, Terry.

Jim

 

[Terry Given]

James Meyer wrote:

On Wed, 27 Oct 2004 10:36:21 -0500, Fred <fps@pmr-iowa.com> wroth:



There is delay of about 120 ms (worst case) between relay actions. I'm
lenthening that to
1/4 sec and in the case of a relay direction change that happened in the
very near term, I'll
throw in an extra 1/4 sec.

The troubling part is that the fet sometimes goes out all on it's own.
It could be overheating that causes
this, say if the relay sticks as you mention below, but I have the
nagging suspicion that a transient is getting in
to thedrain of the fet. Maybe if I remove D1 and replace with a snubber
to cut the rate of dv/dt at the motor
cable entry to the PCB, that would give some added protection.



I'd try a snubber from Q1's drain to ground. A 0.1uF ceramic cap would
shunt transients to ground. Of course, Q1 would then be switching the normal 24
volt charge on the snubber cap to ground when it turned on, but that shouldn't
be a problem unless you were switching Q1 on and off rapidly as in some sort of
PWM modulation of the motor speed.

Jim

a series R = Zo helps reduce the peak current flowing, but for low
Fswitch, yeah. be careful what ceramic cap you pick though - Z5U and Y5V
(if its not spec'd it'll be one of those) are nasty with applied voltage
(down 70%) and temperature (down 80%) so at 50C your 100nF might look
more like 4nF. Most engineers know about the shit tempco, but a
surprising number are unaware of the voltage coefficient. X5R/X7R/XB are
great.

the peak energy in the snubber cap is 0.5*Csnub*Vpeak^2 where Vpeak is
the peak voltage across the cap, so the power loss at Fswitch is simply
Pswitch = 0.5*Csnub*(Vpeak^2)*Fswitch

the location of the snubber should be such as to ensure the ringing
current flows in the smallest possible physical loop.

Cheers
Terry

 

[Fred]

Terry Given wrote:

James Meyer wrote:

On Wed, 27 Oct 2004 10:36:21 -0500, Fred <fps@pmr-iowa.com> wroth:



There is delay of about 120 ms (worst case) between relay actions.
I'm lenthening that to
1/4 sec and in the case of a relay direction change that happened in
the very near term, I'll
throw in an extra 1/4 sec.

The troubling part is that the fet sometimes goes out all on it's
own. It could be overheating that causes
this, say if the relay sticks as you mention below, but I have the
nagging suspicion that a transient is getting in
to thedrain of the fet. Maybe if I remove D1 and replace with a
snubber to cut the rate of dv/dt at the motor
cable entry to the PCB, that would give some added protection.



I'd try a snubber from Q1's drain to ground. A 0.1uF ceramic cap
would
shunt transients to ground. Of course, Q1 would then be switching the
normal 24
volt charge on the snubber cap to ground when it turned on, but that
shouldn't
be a problem unless you were switching Q1 on and off rapidly as in
some sort of
PWM modulation of the motor speed.

Jim


a series R = Zo helps reduce the peak current flowing, but for low
Fswitch, yeah. be careful what ceramic cap you pick though - Z5U and Y5V
(if its not spec'd it'll be one of those) are nasty with applied voltage
(down 70%) and temperature (down 80%) so at 50C your 100nF might look
more like 4nF. Most engineers know about the shit tempco, but a
surprising number are unaware of the voltage coefficient. X5R/X7R/XB are
great.

the peak energy in the snubber cap is 0.5*Csnub*Vpeak^2 where Vpeak is
the peak voltage across the cap, so the power loss at Fswitch is simply
Pswitch = 0.5*Csnub*(Vpeak^2)*Fswitch

the location of the snubber should be such as to ensure the ringing
current flows in the smallest possible physical loop.

Cheers
Terry

Thanks to all how have contributed to my understanding of the problem.
D1 is a bidirectional diode. I couldn't find the proper symbol for it so
grabbed the single sided one. Then I forgot to mention that in the text.

I'm using D2 to do just as suggested above; to snub a transient if the
bulk cap. is insufficient to allow D3 to get the job done. I'm
superstious as well!

Regards,

Fred


----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= East/West-Coast Server Farms - Total Privacy via Encryption =---