Optical encoder nightmare!

[jimbo750]

Hi there!

I'm having a nightmare with a motor and an optical encoder,

When I test the setup with the motor I want to use (a planetary 12V DC, precious metal brushes), the optical encoder seems to generate additional false pulses whenever the motor is moving, so it creeps and looses it's position very quickly.

I have another, cheaper and dirtier 12V motor which is 50% more powerful, and it exhibits ZERO encoder errors/ creep when wired into the same circuit, even when moving very quickly. If I turn the encoder by hand, even very fast moves won't confuse it, the software side is working very well.

So, I assume that these false Encoder pulses generated when the DC motor is running are a result of EMI? This is my first serious encounter with EMI, I don't know the place to start looking in my circuit (see attached). Each IC is bypassed. I have tried with 1k and 10k pull-up resistors on the Encoder outputs, as it doesn't have them internally.

Any tips on where to start my EMI reduction journey, or have I missed something else entirely?

Jimbo

[Al_The_Man]

Did you try connecting both 12v & 5v supply commons together, if they are not already?
Al.

[jimbo750]

They are, unfortunately!

[jimbo750]

Here's an updated image- I've put in the regulator exactly how it is wired up. The whole circuit runs on the 12V and ground.

The motor draws ~0.8A at stall, but it seriously interferes with the encoder/uC communication even when rotating unloaded, where current draw will be <100ma I believe.

[H500]

Your bypass caps should be .01 to .1u ceramics. The 10u don't work well at high frequencies. Also, are you using proper grounding techniques? Separate the high and low current grounds and joint at at single point as close to the big power capacitor as possible.

[Al_The_Man]

You may want to ensure the motor frame is connected to the common earth ground as well as the P.S. commons.
Al.

[jimbo750]

Thanks for the quick reply!

Originally Posted by H500 Your bypass caps should be .01 to .1u ceramics. The 10u don't work well at high frequencies.

Aha, I didn't know this, I'll give it a go. Would there be any down-side to using 'both' the 10u and the ~0.1u? (I ask because I could easily add the ~0.1u to an existing board to give it a try)

Originally Posted by H500 Also, are you using proper grounding techniques? Separate the high and low current grounds and joint at at single point as close to the big power capacitor as possible.

I think this is fine-- they are separate, and join only at the Ground pin of the regulator, right next to the 10uF cap which bypasses the 12Vin/Gnd pins of the regulator, i.e. bypassing the board's 12V input. Does this make sense?

[jimbo750]

Originally Posted by Al_The_Man You may want to ensure the motor frame is connected to the common earth ground as well as the P.S. commons. Al.

Thanks- that never occurred to me, I'll give that a go too.

[H500]

Putting the caps in parallel would be fine. Be sure to do it as close to each chip as possible. I believe some of the AVR chips have several supplies which must be bypassed. Be sure to check yours.

If noise is a problem, try placing a RC filters inline with the encoders signal lines.

You might also try a much bigger cap >1000u across the 12v supply.

[bharbour]

You have not mentioned anything about cable routing and connector assignments in your setup. Try routing the encoder cabling as far from the motor power wiring as possible. If you are going to put the motor and encoder wiring into a connector, separate them as much as possible, and don't share grounds or shields. Better still would be to use separate connectors.

Put some 0.1 or 0.01 uF ceramic caps across the DC power inputs to the drive, as close to the drive as possible.

Just a thought, but you mentioned pullup resistors on the encoder: are you seeing good logic levels on the encoder inputs at the chip at operating speed for the motor? Open collector logic tends to not be very fast.

A last resort would be to use differential signalling between the encoder and drive. With good wire routing, good power supply layout and reasonably short wiring, that should not be necesary.

Good Luck,
BobH