Why encoders don't work?

[ellenwu]

My colleague met a very interesting problem recently,I checked the system together with him, but I don't know where is problem, I hope I can get some help here:
We install a production line, which is drived by 3x232kW DC motors, 2x37kW DC motors, 4x45kW ac motors and 4x1.5kW ac motors equipped with encoders.One transformer 1000kVA is allocated to feed this production line only. The problem occurs at these four 1.5kW motors and their encoders.
- First we use shield signal cables for these encoders, while the cables from VFD to motors are not shielded. we install all the cables together ( signals and power) in a cable tunnel. let me designate 1.5KW motors as motor 1, motor 2, motor 3 and motor 4. We only turn these four 1.5kw motors on while the others stopped. we find only encoder of motor 1 has signal, motor 2,3,4 no signal.
- Then we disconnet signal cable of motor 1, move this cable to encoders of motor2/3/4 one by one, then we find that encoder of motor 2/3/4 has signal.
- After that we remove all signal cables of encoders of motor1/2/3/4 to a seperate cable tray(in these cables tray still have some ac motor power cables). Tested, all encoders has signals.
- Finally we turned on 3x45kW ac motors,with other DC motors stopped, we find that encoders of motor2/3/4 doesn't have signals again, only encoder of motor 1 has signal.
- We worried about that harmonic of ac motors, then we remove all signal cables of encoders from the cable tray,directly put the cables on the steel platform(note the cable tray is under the steel platform.) still encoders of motor2/3/4 no signal with ac motors.
- Check the grounding inside the panel, signal grounding and power grounding is connected together.
Appreciated if you give me some comments.

 

[Skogsgurra]

Do the encoder signals really disappear? You checked with an oscilloskope? Or does "no signal" mean that yor drives says there is something wrong with the signals?

What signal levels do you use? TTL (5 V) or 12/15 or 24 V? What make are the drives? And the encoders?

Gunnar Englund

http://www.gke.org

 

[ellenwu]

No, we haven't use oscilloskope to check. No signals have two situations:
1. The drivers didn't recieve signals.
2. Sometime, the drivers got signals, but the signals is lower than the the true value.
The encoder we used is 5V type.

Thank you a lot, skogsgurra!

 

[Skogsgurra]

Hello again,

Using TTL in the plant is not a good idea. You should try the Leine&Linde or similar encoders. If that isn't possible, try to use pull-up resistors (1 - 2,2 kohms connected to +5 V) on incoming signals. I would try 1 kohm first. The lower the better - until the zero voltages comes creeping up.

I guess that your PPR is something like 1024 or so. That means around 30 kHz pulse frequency at 1500 RPM. Then you can also use a capacitor on the signal lines to filter out interference. A 1 nanofarad capacitor between signals and ground usually helps. That will give you an f0 around 160 kHz - assuming a very weak encoder output. A 160 kHz low-pass will let your pulses through and, at the same time, filter out interference from the inverters.

But, as I said first, use industrial grade encoders (L&L etc) if that is possible.

Good luck!

Gunnar Englund

http://www.gke.org

 

[jraef]

Sometimes spikes or severe noise on the encoder line also causes the drive to display "no signal" as an indication that the signal is unreadable, not necessarily missing.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[ellenwu]

Yes, My PPR is 1024, It will be diffcult to change the type of the encoders because the production line already puts into use. we will try to use capacitor as skogsgurra suggested.
Thanks.

 

[nbucska]

The best solution would be to put a debouncing logic
with the line drivers close to the encoder. Use
shielded or twisted pair cables and add additional filter
to the receiving end.

Basically after the signalk is debounced, you want to
amplify it , emphasize the high frequencies and on
the receiving end attenuate them.

If your application allows for it, a phase-lock loop is
very effective filter for noisy signals.






Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[EdDanzer]

It is possible that installing a ferrite bead over the encoder wires may help, also try disconnecting the shield at each end and both ends. I have seen micro lighting in free air cause malfunction of TTL controls.
Any DC wire should cross AC wires at 90°.

 

[ScottyUK]

"Micro lightning"? Could you explain a little more about this mysterious phenomena?

This has the feel of a grounding problem, particularly if the signal and power grounds are commoned. If possible ground the screen of the encoder signal at the controller only. Be very careful to avoid grounding at more than one location otherwise currents will flow in the screen.

Agree with Skogs over choice of encoder type for a noisy environment.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[EdDanzer]

My horror story of micro lightning caused problems. We have built a machine used in green house nurseries that grow trees for reforestation to push the small trees from their growing block one row at a time. There is a servo motor, encoder and Galil motion control card for positioning and I/O functions. We delivered a machine into northern California. Some time the machine would go wild for no reason, it worked at our shop, and in several other facilities. I went to debug the problem for the second time and decided to drive a ground rod next to the machine, this made the problem worse. As I’m standing next to the machine holding my Fluke 87 probes about 6” apart I see a 6 volt reading, and the machine takes off. As I look around I realize the building is aluminum framed with a poly cover and the sprinkler system just turned on 20 ft. away.
Several minutes of observation produced many 3 to 6 volts spikes on the multi meter. I clipped all shield drains back to the cover, clamped ferrite beads over the end of the shields, and removed the earth ground. The machine has worked fine ever since.
A few years ago we retrofitted a K&T horizontal machining center, things went ok until spring and we started having random errors. We finally moved the 120 VAC control power to cross at 90° and solved the random errors.
These may not solve your specific problems, but are simple and inexpensive to try.

 

[nbucska]

VFD generates a large amount of noise.
1.) keep the signal line far at 90 deg. from power lines.
2.) use buffering with as high VTG as practical
3.) use HF emphasis/deeemphasis
4.) Use debouncing logic and optical fiber -- costs
more but absolutelly eliminates noise pick-up
5.) On the far end use phase lock loop for further
noise rejection.



Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>