Noisy encoder signal on industrial robot.

[Gibbedy]

Hello All.
I'm after some feedback on noise I'm seeing on some encoder wiring.

I put together a video to communicate the issue to a previous electrician who worked on the equipment.

I thought there might be someone in this community that enjoys solving problems like this, and it does happen to involve a teensy 3.2.

If anyone feels like watching it's 10:41 long video with scope of an encoder clock and data lines of two seperate encoders. These encoders belong to an old robot hoist system and I was just getting some baseline readings to assist when we have issues with the encoders/drives involved in this system. I was expecting to get similar signal noise levels on all 4 axis encoders, but strangely 2 look very noisy.

Here is the link if anyone feels like viewing this sort of thing.

youtube.com/watch?v=jOwtJ9JmMpg&feature=youtu.be

 

[MarkT]

Can you swap the cables to see if the problem moves with the old cables?

Do the noisy channels have wiring that runs next to high power cabling?

Are the encoders differential? If so that level of noise isn't a concern unless its differential-mode...

 

[defragster]

swap the cables to see if the problem moves with the old cables?

was going to say that or ask if just (safely) swap the connected encoder to see if it follows.

but, if I thought of it - seemed like it might have been done?

 

[Gibbedy]

Hi MarkT

Can you swap the cables to see if the problem moves with the old cables?

Not easily, that would/has only be done when the machine won't run due to an encoder issue.

Even moving cables (and the encoder attached to the end) between inputs needs carefull consideration before doing.
The problem is these encoders need a knuckle boom to access, permits, hire etc. Not something I'd do for my R&D purposes.

Do the noisy channels have wiring that runs next to high power cabling?

Most likely they do. The two noisy encoders run in channel down one side of a printing press and the two cleaner signals run down the other. I'm guessing a 40m run with 25m of energy track.

It's a possibility that the noise can be accounted for by differnt cable route alone.

Are the encoders differential? If so that level of noise isn't a concern unless its differential-mode...

The encoders are SSI differential. There is some differential noise. But I'm confused as to why 2 of the four encoders show much more common mode noise than the other two.

One theory i have is that the sheilding is broken due to the thousands of movements of the enrgy track.
Or high resistance on the data lines.
I discuss this at around 8:20 in the video.

Thankyou MarkT.

 

[Gibbedy]

was going to say that or ask if just (safely) swap the connected encoder to see if it follows.

but, if I thought of it - seemed like it might have been done?

Unfortunately its very difficult to get to the encoders.
I could swap encoder and cable at the controller end to prove it is the cable and encoder, but that has risks i need to feel comfortable with. This hoist runs a master slave setup. Were if i were to swap encoders, and somehow the machine is reset, one axis will imediatly drive at the highest possible speed to match the position of the master.

I'm doing a bit of research while the machine while it is down in the hopes to speed up fault finding when it wont run.

 

[MarkT]

If they are differential you need to look at the differential mode noise, can you use math mode to plot the difference and see what the true differential noise is like?

 

[Gibbedy]

If they are differential you need to look at the differential mode noise, can you use math mode to plot the difference and see what the true differential noise is like?

I must admit I've never used any maths functions on my scopes.

I scope both an and b of my differential channel with reference to 0v (single ended) and I scope between an and b to get the differential signal with the common mode noise gone. (using differential probes)

Anything that Is jumping outside my nice square wave I take to be differential noise.

I will look into math mode as it would provide an easy way to document and communicate differential noise if it does what I think you are saying it does. That is give me a waveform of the noise only portion. The picture below tells me that there is very little differential noise.

Blue and Red are the single ended measurements of a and b for the data channel.
Brown is the differential measurement of a and b.

My goal was to get a baseline reading of what a reasonable SSI signal looks on this controller. What I have is one channel with excessive common mode noise, one channel with reasonable common mode noise and two channels that look perfectly clean.

I have no manuals to know at what point the bosh rho3 kinematic controller that these encoders go into will not read these signals correctly. That is there will be a point where the noise is too far outside it's range with respect to 0v.

Thanks for your input Mark.

If I get more real test, waveforms clearly showing levels etc I'll post the results here.
There is an identical controller/drive setup on another machine I have access to. I'm interested to see what readings I get there from all four channels that machine.

Gavin.

 

[MarkT]

Common mode noise is ignored by a differential receiver, that's why they are used for encoders in industrial environments. If the differential mode signal is clean there is no problem.