SNAP-SCM-SSI Module and Novotechnik TP1


#1

Ok, I just successfully connected and got working an SSI module and Novotechnik magnostrictive linear sensor. I am bringing this up because I had to struggle a bit but once you know what to look for it is fairly straightforward. Also I wanted to point out just how accurate this thing is and how fast it is for the money.
I purchased this Novotechnik 24" long sensor with SSI interface for around $300 and and reads in absolute postion to to 1 micrometer (0.000039"), and is accurate to around 13 micormeters +/- 0.0005". On top of all that, it is capable of update reads at up to 16khz. If you set it to only 200khz clock speed (in the Opto module) it still reads at over 5800 times per second or every 172 microseconds…

I found that due to the usual issue in serial and analog technologies that + is not always + and - is not always -, the main problem was the numbers comming from the sensor were inverted. Therefore, if the data polarity is wrong (whether it is a cable mismarked or whatever) the number will always be in the millions at either end of the scale. Just simply swap the data wires and the number will then read correctly.

I also found that since I was in a hurry to get this running, I used a “Wall Wart” plugin transformer that produced enough milliamps at 24 vdc. Well apparently (as you might expect) the noise level of the WW was too much and the sensor would never read or change the reading. I then connected a Power-One 24 vdc linear (very quiet) and it began reading numbers (although they were inverted) that were changing with the sensor target movement.

Remember that on these sensors even though they will update at extremely high speeds, the probelms with the cabling will increase dramatically withg speed. I set mine to 200khz which is fine for short runs (10’) that do not have twisted pairs. The cable that came with it is not a twisted pair, is shielded and 5 meters long. I tried numerous other higher clock speeds but with varing degrees of reliability. I suspect if one needed to make a long run, then I suggest buying an M12 connector and installing cat5 shielded cable and keep the clock freq to equal to or less than 200khz.

One more thing, I found that in the process of figuring all this out, all the changes and cable upplugs, etc. can eventually cause either the module or the sensor to lock up. I think the system is probably quite reliable when properly installed, however when it locks up while you are trying to get it configured as in this case, it is hard to tell whether it is the configuration or locked. I suggest doing a hard boot for the brain/rack and power to the sensor.


#2

Hey Barrett,

Thanks so much for reporting back on your experience on this sucker. This sort of thing is just what the forums are here for. Sharing.

How did you go with it strategy wise?

Ben.


#3

As far as reading in a strategy, I set this up in a test strategy and set the loop time for reading it to 250ms on an S1 talking to an R1 and it reads perfect. Then after reading this I decided to se the chart loop time to 2 ms and then looked at the Pac Control loop comm time while this was running and it was at just over 1 ms comm loop time for Pac Control. So yes, it will read it at 2 ms, no problem, especially considering that the module itself is reading it at 5800+ times per sec and the R1 is reading it a better than 1ms, so yah, it is possible to get 1000 updates per second.
You can breathe on it and see it move by a few micro meters…