PCI-MIO-16E-1 external scan clock does not function as digital. What up?

[<Mike Ross> mross@smpcorp.com]

The context is noise on the external clock input to the E-Series DAQ board PCI-MIO-16E-1.

I am unable to clock an acquisition because of noise picked up by the PCI-MIO-16E-1.
The clock is connnected to PFI17/Startscan at pin 38 of the SC-2345.

I have assumed (first mistake) the this would be a digital input. To my feeble ME mind this means that a "high" signal is accepted when the signal rises above 3V, and the "low" signal is accepted when the signal drops below 0.8V. This hysterisis would be the whole point of digital signals to begin with.

Can someone confirm that this is the tradition for TTL?

It would appear that the E Series board cannot be counted on to respond this way. I was told that above 0.8V all bets are off. Indeed that is how my board seems to work. If I get noise above 0.8V the damn thing will clock.

Is this a deficiency in the board design?
Am I getting my leg pulled by an uninformed app engineer?
Is my board screwed up?

Regarding the actual error, the problem manifiests itself this way: If I spin up the appparatus, and look at the signal generated by my encoder with the scope, it is pretty clean (no drive noise is present at all). There is a little cross talk between encoder signals that shows up as little ringing blips when the other signals slew from low to high. If I run this through a schmitt trigger that is removed. I have a very nice looking clock pulse at this point. But, whan I put the schmitt output onto the Startscan pin, and make a ground connection betweent the SC-2345 and the ground of the encoder system, the signal now has radiated noise from the motor drive on it.

This noise is roughly the same. Bursts of sine wave; six or seven waves at about 4MHz. Amplitude varies, but is often over the 0.8V that ruins the acquisition.

It really bothers me that the DAQ board can't clock from this signal. If this is how it was designed then it is pretty weak.

Enough of griping. Any ideas?
Can any one recommend an analog notch filter circuit?
I could go to an air motor or give up on variable speed and just hook the motor up without the drive. These are guaranteed, but very unsatisfactory solutions.

Thanks for any help.

Michael E. Ross
Senior Design Engineer

Standard Motor Products, Inc
2717 Commerce Road
Wilson, NC 27893

Office: (252)234-5821 (voice mail)
Lab: (252)234-2000 x4054 (no voice mail)
Fax: (252)234-1900
mross@smpcorp.com

[Sastry V.]

Hi Mike,

My first suggestion would be check the actual clock signal in an oscilloscope for possible spikes.

Just in "clock config", you specify it where are you getting the scan clock from. I assume you have done all this.

And regarding the ground connections between SC-2345 and the encoder, there might be a chance that they are not at same ground which causes noise.

You may want to take a look at the "Connecting Signals" in the user manual for your card.

There you will see that you cannot just connect grounds of two different sources. Try not connecting those two grounds and see what happens.

Let me know how it goes.

Sincerely,
Sastry V.
Applications Engineer
National Instruments

[<Mike Ross> mross@smpcorp.com]

A question for clarity:

Are you saying that I connect only the two outputs from the encoder (Z for start trigger and A for clock pulse) to the SC-2345? Do not make a connection between the grounds?

They are at some point commonly grounded. The PC grounds out to the wall receptacle and the power supply of the encoder is likewise plugged into the wall.

I have also tried using the power supply internal to the SC-2345 in both external power and E Series power modes (PWR01 module) no change, noise the same.

The "connecting signals" refers to analog hookups through normal channels. I cannot remember seeing any references on the triggers and external clocks and how they should be conditioned.

The noise is not seen on the i
ncoming power lines, it is not present on the encoder. It seems to be radiated and received by the PC/DAQ.