Any suggestions for overcoming static discharge interference during multichannel data acquisition for Operating Deflection Shape? Several components on a size press are to be analyzed. Sixteen channels of data cables spanning the machine. Paper coming out of dryer is <3% moisture so there is lots of static. Thanks.

Does anyone have an opinion or experience that would suggest that the use of charge type accelerometers with charge amplifiers and microdot cables would work any better in this application than the ICP accelerometers?

I have the same problem taking readings on the dryer felts due to this problem. Have you attempted to have them ground the bearings? This condition is so bad that it has shutdown my analyzer on overload. The only way I can get any kind of data is to use a stinger. But this is not the greatest option.

Charge mode accels can be used in endwindings of large generators with surprisingly good results, not that there is so much thunder and lightning but the magnetic field is fine. They are also completely isolated. I have seen places in papermachines where there was more voltage btw. the outgoing shaft of the drive gbx. and ground than the ingoing from the motor and the outgoing bearing kept going bad. So what is the main problem, are the ICP accels dying due to the flash or is the lightnings destroying the readings? Anyway I would suggest making sure the accels are isolated from the frame, we have something we call like "fiberbakelite" that also still transfer vibration pretty good, then a double shielded cable where the outer shield is a actual screen and not a part of the signal path as in a standard RG58 coax. If nothing else works, try a modern velocity probe with balanced signal transmission. Olov

When using a data collector with an ICP accel I can get good readings because I'm using a short cable and a single channel. There are often several good readings to be had between "zaps", but it does sometimes cause the data collector to lock up and need rebooting. My problem is that now I want to do an ODS with at least 16 channels being collected simultaneously. When I get 50-60 feet of cable for each channel strung out in this area with about half of them spanning the web, there are no gaps betwen the "zaps". They are just erratic swings in the waveform data. I am not sure if the signal gets screwed up in the sensor or in the cables. I am using shielded, twisted pair cables, not coax, and I have tried several different combinations of shield grounding, with no improvement.

The sensors are not dying. I have used them in other locations and they work fine.

Olov, I thought about your velocity sensors, too. I know where I can rent the charge mode accelerometers here in the US, but I don't know where to rent your sensors and required associated signal conditioning devices.

Skip Hartman

Skip,

In other troublesome noisy places last resort is to have a double shielded cable where one shiled goes from one end and the other from the other it does not give a double shield effect but increased.
Vendor, Sam P. could a velocity kit for you to test. Olov

Skip,

I guess the question is whether the static discharge is causing an induced vibration on the accelerometer or simply a voltage spike in the wire?

I don't know if this would work:

1) Isolate accelerometer case (including mounting stud) from metal surface
2) Use two-wire twisted shielded wire
3) Ground wire shield near analyzer
4) Try grounding the vinyl cable jacket near the analyzer with metal foil and a ground wire incase of surface conductance

How about a intrinsic safe barrier or opto-isolator to prevent an electrical discharge path to analyzer. What do all of the accelerometer vendors have to say about this?

Walt

I have been performing ODS tests on paper machines for 15 years. I always use case isolated accels with coax cable and BNC connections whenever possible.
To date, I have never experienced this problem.
Since the static discharge is unlikely to be synchronous to the shaft, I would check the acquisition method.

Walt,
I am hearing from the accelerometer manufacturers that probably charge mode accelerometers would be worse. I am also hearing, both from the accelerometer guys and the instrumentation manufacturer that I should try almost to the letter the configuration you described with the accel case attached to the cable shield through the connector. I already use shielded twisted pair, but all my connectors leave the shield floating at the accelerometer connector. I just need to come up with a way to test the configuration!

Skip

Tom,

I have also been doing ODS on paper machines for a very long time. I have never had to do it with lightening bolts shooting all around, however. The frequency of interest in this test is not synchronous with any of the shaft speeds, either.

Of course, static discharge is not synchronous to a shaft speed. Static discharge generates RF. The RF induces erroneous data (that looks like saturated sensor preamps) in the sensor cables if the correct shielding arrangement isn't configured. The correct type of shielding configuration is somewhat dependent on the type of input configuration the data acquisition system has. Mine has single ended inputs.

Skip

IEPE or ICP accels are by design mostly single ended so what you connect it to makes less difference, it may be worth a try if it is possible to run the collection thing on battery. Only place I have seen differential connected accels are in monitoring aeroengines in the aeroplane and testbed both charge mode and IEPE(ICP). Guess they also have some electrical noise and some flash once in a while in those. But they are expensive. Olov

Can you run your cables inside a metal tube and earth the tube to the machine. This should give you some basic Faraday shielding as long as you insulate the cables from the tube - say wrapping the bundle in PTFE tape.

Best Regards,
Tom Murphy

Skip,

This may be extreme, but I would ground the accelerometer case to the cable shield and ground the cable shield near the measurement instrument. This Could provide a metallic Faraday shield without adding an additional metal shield as Tom indicated.

To test all of these theories; get a small kite and attach accelerometer and then connect a key to end of cable and then place key in glass jar. Let kite fly over the paper machine and hold on tight. Well, it worked for Ben Franklin!

Walt

Well there is this big fan in the roof right above the size press, so I think the kite will fly.....

Skip