Filtering motor spikes

[darkith]

Hi guys,

I've been working on a 900mhz micro TX/RX unit using Integrations IA4320/4220 chips, and run into a problem. I've reached the stage where I can drive actuators & motor, but the motor spikes are causing the IA4320 receiver chip to reset.

I've tried a handful of things, but no success in stopping this behavior. I've tried several parallel caps (470 uF elec + 4.7 uf SMT + 10000pf SMT + 100 pf SMT), I've tried moving the battery leads closer (on top of) the motor leads, I've tried moving the battery leads closer to the RX chip, etc. The large caps appear to be keeping the voltage high enough that the PIC doesn't brown-out and reset, but it seems like the IA4320 chip is more sensitive. On the scope, it's around a .5v spike. I'm driving the IRLM2502 FET with PFM, so it's a lot of frequent, short pulses & spikes, versus the longer pulse of PWM.

Things work better if I put a resistor inline in the motor lead, it reduces the spikes sufficiently that the RX chip doesn't reset, but of course the resistor is wasting crap loads of power.

Any suggestions? I'd like to avoid the voltage regulator route...

David

[Dan Baldwin]

Some isolation between the battery and the receiver chip might help. You might try a resistor in series with the sensitive supply, or possibly a schottky diode if you can tolerate the voltage drop. Make sure that the capacitors are still on the receiver side of the isolation resistor or diode.

Dan

[vintage1]

This is the full kit.

You don't say WHAT motor, and how its connected..a Cqt diag would be good.

[Comatose]

Just about every RX on earth uses a 3.3v LDO to power its internals for exactly this reason. I'm partial to LP2980 series regulators for this task. They're tiny, inexpensive and don't require a lot of bypass capacitance. They also work very well. Running your circuit directly from the battery is just asking for trouble in my opinion. The dropout is only 150mV.

http://www.national.com/ds/LP/LP2980.pdf

[Mr.RC-CAM]

Any suggestions? I'd like to avoid the voltage regulator route...Are you saying that your PIC and Rx IC are powered directly from the raw battery voltage?

[dburley]

If you are creating the PWM signals on the PIC it may be that the
PWM method permits the spikes to effect the signal. I had the same
problem on a very small Rx until I revised the PWM method. It is very
essential that all changes to the output pins be made at the same
instance. If this is across 2 ports then the 2nd output must follow
the 1st immediately so the latency is only 2 micro seconds.

Hope this helps !!

Dave

[darkith]

Motor: 6mm ZipZap, probably drawing a few hundred mA.

Yup, the PIC/RX are powered directly from a single cell LIPO.

An LDO is an option, but I've been trying to keep size/weight down (in the 0.5-0.75 g range). The other reason is that the IA4320 has a low battery detection capability, which I was hoping to harness to provide a low battery motor cut-off. Providing a regulated voltage supply prevents me from doing that.

D.

[Mr.RC-CAM]

If the resets are due to fast rise time induced noise, then this will be a tough problem to solve (even with a Vreg). However, if it is due to basic motor commutating ripple then a cheap solution would be to isolate the Vcc going just to the two chips. Feed the battery voltage through a low Vf schottky diode and place your .1uF and 470uF cap on the cathode (chip) side. This will offer some isolation from the short duration low voltage depressions. It will also help protect against reverse voltage when a user dumb thumbs the installation.

For sure, your motor should be given the standard 3-cap and schottky diode treatment.

[Terry S]

I had the same problem with a home made pan and tilt but I gave in and fitted resistors in the motor leads as it was the only thing that worked. I did have a reduction in interfierence by reducing the frequency of the speed controllers but not enough to cure it with out the resistors. No amount of chokes and capacitors had any real effect. I would be interested if you find a better answer.

Good Luck, Terry

[Zlatko]

Hiya darkith,

Good to see your success with the IA chipset :)

Check the capacitor across the motor that its still fine, perhaps change it just to be sure. You may also try a resistor before the filters for the PIC/IA , perhaps something like a 5 Ohm. At 5 Ohm the voltage drop should only be 0.1V ( assuming 20mA draw from PIC / IA chipset ) , this shouldn't affect the operation of the chips and just offset the IA low voltage detection by this amount.

Cheers

[green-boat]

Have you tried twisting the wires as in a twisted pair. This has worked for me when ferrite rings wouldn't.

[darkith]

Hrmmm...some good suggestions there.

Have to build a new one me thinks (somebody let the magic smoke out of this one...)

D.

[darkith]

Got it.

Ended up with a combination of isolating the actuator FET & motor supply from the PIC/RX supply via a 10 ohm resistor (with suitable filtering (47uF + 1uF + 100pf) on PIC/RX side), plus a larger capacitor (4.7uF) across the motor outputs.
A smaller resistor and larger cap would probably work better...but I don't have any in SMT on hand.

Debugging was aggrevated by a crappy power supply which was resetting everytime the 5v output was pulled too low, and a broken trace. :P

D.