Freewheeling/Snubber Diode for High Temps/High Current DC Motor

Thread Starter

LowQCab

Joined Nov 6, 2012
4,023
Hello,
Please correct my terminology, it seems that 3 or 4 terms are used for the same Diode/Motor configuration.
Automotive application, i.e. ~200F under-hood temps.
Air conditioning blower motor, and/or, electric radiator fans.
Motors are crude, brute force, inefficient, brushed, permanent magnet, with a high inertia flywheel type load on start-up.
26 to 28 amps DC, 12 to 14 volts on high speed,
(heavy, forced-air-cooled, open wire-wound resistor, for speed control)...... OR
( series/parallel relay contraption for 2 speed operation,
(actually works quite well, as long as you have 2 motors working in tandem with each other ) ).
Filtered with 33,000 mfd capacitor, per-each motor,
(capacitors mounted inside where it's always cool, using way too much 14ga. wire).

IXYS Power Schottkey Rectifier DSS 60-0045B Data Sheet = http://ixapps.ixys.com/Datasheet/L127.pdf
IFAV = 60 A
VRRM = 45 V
VF = 0.57 V

The Problem....
I would like to cut electrical noise from the brushes, and
hopefully extend brush life and "maybe" get a slight power output gain,
from my heavy under-hood motors. (5 of them),
(the alternator is easily up to the task of providing
twice the required current while maintaining exactly 14 vdc under all conditions).
This pursuit started years ago when I spent
quite a bit of time trying to build PWM controllers for similar motors.
At which, I was a repeated, dismal, failure,
with loads of the evil blue smoke squirting from every semiconductor.
(Sometimes with a spectacular display)
And, I think that "one" of the stumbling blocks was finding the correct diodes to protect the MosFets.
I have since come up with several other, "not quite as satisfactory", methods of speed control.
I have recently dabbled at addressing this situation again,
and thought I would finally ask for some suggestions.

The above referenced Power Schottkey's seem to be well over-rated for the job,
but I'm obviously not taking some, or many, factors into consideration.
These diodes, when close connected to the motor(s),
produce no detectable temperature rise with the motors at full load, and
with repeated start/stop cycling, with NO heat-sink.
Even so, I mounted them to a small heat-sink made of 1/8th inch aluminum angle,
approx. 1"x 1"x 3" long, with proper heat-sink compound.
They consistently fail, in a shorted condition,
after around ~10 to ~15 hours of run time at ~160 to ~180F ambient temps.
( My engine control computer is factory mounted under the hood as well,
with zero failures in 15 years )
I could possibly fan cool them down to around ~120F.
( PWM alternator field winding noise ??? maybe ?? )

I would prefer to not have to mount them inside the cab of the truck along
with their attendant trash radiating wiring. (twisted 16ga. stranded wires)
This may be some-what silly, in that, I don't really have any noise related wiring problems.
But I'm still kinda stuck on making this a more elegantly engineered, and operating,
motor control system.
The commercially available solutions have many limitations, or exorbitant price tags.
I do not have much in the way of test equipment, or really even the time to
properly teach myself how to properly design such a system, so any
suggestions would be greatly appreciated.
I hope I am providing someone with a unique challenge that they find satisfying,
I know I've spent crazy hours helping others (in different fields) so
I feel I'm deserving of some help back. Thank You for your consideration .
Jim
 

be80be

Joined Jul 5, 2008
2,072
Maybe have a look here
https://en.wikipedia.org/wiki/Flyback_diode
In an ideal flyback diode selection, one would seek a diode which has very large peak forward current capacity (to handle voltage transients without burning out the diode), low forward voltage drop, and a reverse breakdown voltage suited to the inductor's power supply. Depending on the application and equipment involved, some voltage surges can be upwards of 10 times the voltage of the power source, so it is critical not to underestimate the energy contained within an energized inductor.
 

Thread Starter

LowQCab

Joined Nov 6, 2012
4,023
Thank You,
I'm not "that" unfamiliar with the principles in play, but I don't have the capabilities to accurately
measure inductance, or a nice scope to visualize it on.
My line of thinking goes something like this......
The motors are somewhat crude in design, in that, they are designed to
a "cost-point", rather than a "performance or efficiency point".
Meaning that..... they are "probably" designed with as little metal and wire as possible,
using inferior grades of metal in the armatures, with high losses and lots of heat generation.
Which "should" mean, that they are poor, lossy inductors, and have few turns of wire for
generating a high voltage spike.
But of course, I'm not a motor or inductor engineer, and I could have it completely backwards......
I was also thinking that the diode basically re-routes that voltage back into the
inductor, (or motor), before the voltage has the time to rise even to the level of the input voltage.
Basically "clamping" that spike to almost an insignificant level.
I thought, probably incorrectly,
that that was the whole reason for having a "fast" diode, like a Schottkey design,
that is to say, "normal" speed diodes do not respond quick enough to the reverse in current,
allowing the voltage spike to rise to undesirable levels before re-routing it back into the loop.
After all, isn't that "faster" re-routing what contributes to the efficiency increase that is
observed when using a "faster" diode ?
But of course, fast-ER, isn't an absolute, but isn't there a "measurable" "speed" at which
the voltage can increase which is based on the composition of the core metals, and would not
that speed be "reduced" to some degree by using inferior metallurgy in the core ??
Or have I got that backwards too.
I guess it could be possible that certain inferior metallurgies and configurations could
actually increase the rate of voltage rise and it's severity.
In that case, my ASSumptions could be just that.......
It also kinda blows my mind that there could be enough induction, and/or, resistance in a large
(33,000mfd) electrolytic cap that it could not absorb the spike(s), and
yet not be destroyed, while a 45volt rated diode is blown through and shorted.
( My big caps are undamaged with many hours on them, with only a 25v rating ).
(The capacitors also filter the 3-phase ripple from the alternator, the battery is not in this circuit ).
This brings up another interesting thought......
Does the "failure mode" of the diode tend to indicate the reason for failure ???
(too high current, too high voltage, too high temperature)
I mean they "could" fail "open", but they always fail shorted and blow a fuse.
 

ifixit

Joined Nov 20, 2008
652
Most likely it is brush noise/arcing that is causing diode failures over time. Brush noise is extremely high frequency, up into the hundreds of megahertz. Semiconductor diodes that are used in your application are way too slow to even start to turn on so the voltage spikes appear across the junction and over time can cause a junction failure. These spikes come from the nano Henry wire inductances right next to the brushes. Since the energy in the spikes is usually low it is fairly easy to deal with them.

There is a second emf produced by the DC brushed motors winding inductance. This is in the order of tens of millihenries and is easily suppressed by the diodes built into the MOSFET drivers.

There is a third emf produced by DC brushed motors when they continue turning by inertia after power is removed. In other words they become generators. The voltage is normal, somewhat less than was being applied before interruption, the same polarity, and is proportional to motor RPM. You can short this voltage out to affect dynamic braking of the motor.

Some tips to suppressing brush noise:
  1. Place ceramic capacitor filtering at the brush terminals.
    1. 1nF to 10nF should work.
    2. Voltage rating of 4 times the highest DC motor voltage.
    3. Place across the brush terminals and/or from each terminal to motor case.
  2. Ground the motor case.
  3. Place the PWM electronics at least 2 feet of wire away from the motor. The wire inductance helps to prevent high frequency noise from getting into the electronics.
  4. Twist the wires together. This reduces the "antenna" effect, which can radiate interference to nearby devices causing malfunction and/or damage. Add an EMI filter core to the leads as well.
Good Luck,
Ifixit
 

Thread Starter

LowQCab

Joined Nov 6, 2012
4,023
WOW, great answer, Thank You.
I should have remembered several decades ago I looked into the very things you're referring to, (RF noise).
But it had completely slipped my mind for some reason.
I think it must be because I was more concerned with IC power supply bypassing with multiple sized
ceramic caps at the time, with motor controls on the back-burner.
Speaking of which, since you are referring to RF noise, which can cover a very large frequency range,
do you estimate that there would be any advantage to multiple, varying sizes of capacitors in
a brushed motor application ? Or is that getting into silly over-kill ?

I remember having some concerns about making the PWM Fets constantly pump a
capacitor strapped directly across the package leads, but I never got that far into it to find out.

Other things, just for clarity, (I'm guessing they are not critical).......
The internal connections of the brushes/windings are inaccessible.
The stamped sheet metal case is spot-welded together, and non-serviceable.
But, small bonus, the case is the ground connection, with the positive terminal
protruding through a drilled hole with a rubber grommet holding it in place.
There is also a possible disadvantage to this.... the PWM Fets will be switching the ground.
But I suppose the noise can be bypassed around the Fets anyway, without harm,
in fact that might be the exact recommended configuration,
both at the motor, and at the Fets.
I'm guessing I'll have to figure out, "how-much-is-too-much" capacitance for
the Fets to have to pump..... Possibly some resistance required at some level/frequency.
Or, it could be a moot point, and I'm just worrying too much, with not enough data or
the proper methods and considerations for actually "knowing" it will work properly on paper.
The gears are turning....
It occurs to me that the Fets have their own capacitance "built-in" that they routinely pump,
and would scoff at a measly extra 10nf load at the relatively low frequencies they will be switching.
I'll have to look at those specs.
Ferite Beads/Cores are something that never occurred to me,
and could block much of this RF noise right at the source, and with a 5 minute install time !!!
And/or, AC input filter modules could be just the ticket,
with a heavy-duty steel case with mounting ears and solder lugs !!

Thank You very much for renewing my interest in this pursuit !!
Any further comments/ideas would be very welcome.
........................Jim....
 

ifixit

Joined Nov 20, 2008
652
  • In your case just place the 10nF cap on the motor between the protruding positive post and the nearby case. The capacitor leads should be as short as possible. Only 1 cap, but you can experiment with higher values. You can use more than one cap in parallel so long as they are of the same type and value.
  • The wires mentioned in my list item 3 have series inductance and resistance. This also serves to isolate the 10nF capacitance load from the PWM drivers. They likely could easily drive this capacitance load anyway. You have to use an oscilloscope to check for ringing. An RC snubber can be used to deal with ringing if needed. Ringing is an undesired source of EM emissions.
  • Limiting EM emissions and susceptibility is a bit of an art. There are dozens of factors both physical and electrical to consider. Google can help you to create your own 'best practices' list for limiting EM. You can use a portable FM radio to probe around to see where EM is coming from and what effect your improvements have.
  • Have fun.
Ifixit
 
Top