PCI 5105 spikes on not connected channels

[vansoest]

Hi,

We use a PCI-5105 and recently we have spikes on our signals, connected en unconnected channels. Sinds unconnected channels are influenced by a connected channel, we think there is something wrong with de board. That there is something broken in the channel isolation.

The attachted image shows a triggered signal and when the red signal is about to peak, some of the other channels show spikes.

I have connected the scope to a signal generator and when measuring a pulse with a high dv/dt, the channels before and after show small (below 10mV) signals, but no spikes.

How can I test my board properly?

[balexand]

The test that you performed with the signal generator, and the resulting TestSignalGenerator.png file that you uploaded does seem to indicate some crosstalk between the channels.

Here's an idea:

To get more information on the repeatability, can you acquire data from all 8 channels, and see if the crosstalk appears to be better or worse depending on which channel is being driven by the signal generator input?

That is, connect the signal generator input to channel 0, acquire a waveform on all 8 channels and determine which channels show the crosstalk.  Then repeat with the signal generator input on channel 1, then 2, and so on...

[vansoest]

This is what I did after my post. There seems to be a small amount of crosstalk on the neighbour channels of the channel driven by the signal generator. But it is very small and the crosstalk has the characteristics of a capacitor; fast charge on high dv/dt and followed by a standard capacitor discharge curve. 

I could not reproduce the spikes from the TestSignal.png. Can the spikes be caused by a ground issue? I would think not, because all channels should have the same spikes, am I right?