Arrrg, electrical interference problem

[Brietech]

So here is the setup: I have an array of 12 tiny solenoid valves mounted over a 36" conveyor belt that is being driven by a 1/30 HP, 90V motor (and is located about 8 inches from the valves). I have a chip (actually an FPGA) that outputs 2.5V for hi and 0V for lo, and is driving 12 NPN transistors through 500 ohm resistors. The solenoids are hooked up across 5V and the transistor, with a protection diode in place, just like on page 23 of the "interfacing circuits" PDF. When the motor gets above a certain speed though, the valves all start randomly firing, and I can see that the outputs of my chip are no longer at 0V (jump around from -.6V to 2.5V somewhat randomly). The FPGA is powered from a 9V wall-wart type power supply with an onboard 2.5V and 5V regulator, and the valves are powered from a 5V lab PSU, while the motor is powered by a 90V rail-mount PSU. It's somehow interfering, but I don't understand how, so any advice would be appreciated!

p.s. There is a picaxe feeding data to this, so it's not completely off-topic!

 

[Ralpht]

You appear to be suffering from back EMF along with interference. Something that happens when motors turn - they induce a voltage onto the power supply and also transmit thr the air. The combined effect can screw up electronics.
I suggest putting approx 100uH cored inductors in series with the output rail of each voltage regulator right at the regulator as close as you can. As well 22uF to 100uF Caps across the supply lines and 0.1uF caps across each chip in the circuit (between the supply pins) as close to the chips as possible. Plus a handfull of both types of caps sprinkled around the board anyway.
That may help, unfortunatley 90V motors can put enough EMF onto supply lines to cause random resets of CPU's etc.
These cored inductors should be readily available at most electronics outlets - I found they work well.

Good Luck

 

 

[manuka]

Never mind 90V, even 0.9 V "solar motors" can upset a sensitive Picaxe! It's standard to use caps.(typically 100nF) across ALL likely EM interference sources, right at the brush/contact terminals themselves.

 

[Brietech]

thanks, i will try it out and get back

 

[Brietech]

Hmm, putting caps across the supply voltages and inductors in line with the supply rails has had virtually no effect. The pins of the FPGA just go haywire and randomly start jumping whenever the motor is switched on. I will try to put a cap across the motor and see what happens, though. Any other suggestions?

 

[Wrenow]

Brietech,

In R/C hobbies we virtually always cap the motors. One across the two terminals, one from each terminal to the motor can. In extreme cases, we also run the power leads through a ferrite core (sometimes helps).

Good luck,

Wreno

 

[Brietech]

I put a cap across the 90V motor contacts, and still no luck. Even when my FPGA is running off of batteries and is about 3 feet away from the motor, the outputs just randomly jump all over the place! Any more suggestions?

 

[Dippy]

Brietech,
can you put a schematic on Tinypic so that people can get a clearer picture?

 

[hippy]

Have you tried disconnecting the motors to make sure it is the motors causing a problem and not something else ?

 

[BeanieBots]

Brietech, if the suggestions have not helped, it is likely to be a layout issue. Like Dippy says, post a diagram but make sure that the diagram also gives some sort of indication about how/where each element is connected. In particular, pay attention to where 0v lines are connected.

Just to confirm that the motor is the culprit, try running everything but with the motor replaced by a suitable resistive load. I normally use an incandescent lamp but finding one at 90v might be tricky so a suitably rated resistor will have to do.

 

[Brietech]

I'll work on the schematic. The interference only occurs when the motor is turned on and moving. Additionally, the board is now completely isolated electrically (I ran the FPGA board and valve array off of a battery pack), and the interference is still there. The motor is just connected to a 90V industrial psu and a speed controller and is on a completely separate circuit.

 

[manuka]

Suspect that "industrial PSU" & speed thyristor (?)controller? How is the rest of the circuit powered- mains? Motor hash on shared mains connections maybe? Tried batteries?

 

[Brietech]

I tried running the entire circuit on batteries with the exact same results. The only thing "connecting" the two circuits was about 3 feet of air!

 

[Wrenow]

Brietech,

You put a cap across the motor contacts, good start. But, you should also put caps from each terminal to the motor housing (can). That is the standard RC configuration for brushed DC motors.

Good luck,

Wreno

 

[manuka]

Agreed- caps. right at the brushes. We mentioned that about 10 posts back of course. Thyristor based motor speed controllers can radiate RF quite some distance, even if battery powered. It's nothing to have them cover "airspace" - I've just been using a 12V cordless drill near an AM radio & had audible noise on the station even as the trigger press started "squealing", PRE motor spinning.

 

[boriz]

Do you have an oscilloscope?

I would start by probing various parts of the system to get a picture of what kind of ripples are on what lines. Vital information!

 

[Dippy]

No offense intended, but without a scheme of your set-up people can only give you a list of educated guesses. I love the way you've sneaked a PICAXE in there so people can sort out your industrial electrical problem ;-) My hat is doffed in respect.

For tricky stuff a 'scope is a necessity unless you get lucky.

Your interference will be air-borne and/or cable-borne (power and/or signal) - unless it is sub-space interference from the Annesummers Galaxy.

Maybe if we could see a schematic we could see possible routes for interference.
Is the noise getting back to the transistor bases or are they biased strangely? It could be a lot of things and you will have to check each stage at a time as suggested previously.

Have you tried screening (cables and/or circuit boards) ? Is your grounding commoned and/or good? Is it possible to replace problematic connections with opto-isolators?

 

[jodicalhon]

Wrenow: <i>You put a cap across the motor contacts, good start. But, you should also put caps from each terminal to the motor housing (can). That is the standard RC configuration for brushed DC motors. </i>

I often see the caps from terminals to motor housing on R/C cars, and have often seen it advised, but what does it actually do?

It appears at first glance to just parallel two more caps with the one directly across the terminals, unless the motor housing is grounded or something.

Genuinely interested in this. It seems to be something of a black art. Is inductance from the cap's leads worth considering?

 

[boriz]

At a guess, I would say that the metal motor housing can act like a screen IF it is connected to earth. But if the motor can be reversed then neither of the motor leads is suitable, so maybe the two caps idea is second best. Maybe?