Probing causing issues to device under test.

[yyzkevin]

Hello!

I made an extender board for a project I am working on, with the intention of being able to  measure all the signals on a pcmcia card under test.  We are not talking cardbus here, so this is ISA like speeds.

Setup is an old Agilent 1680AD Logic Analyzer with E5339A high density probes, and a custom PCMCIA extender board with the 38pin MIKTOR connectors.  I am operating in state acquisition mode with the read/write signals serving as the clocks and card enables as the clock qualifiers.

If I plug the extender board into the PC along with a card and no probes attached at all everything is fine.
If I plug probes into the MIKTOR connectors but do not attach them to the  logic analyzer, also there is no problem
If I plug the probes into the logic analyzer though,    it will cause corruption but not consistently, for lack of a better explanation only when there is a lot of activity it will cause the system to lock up.

Looking at the signals with an oscilloscope it is not obvious to me that anything is different when the probes are plugged in.  I can plug and unplug the probe and I see no change in the waveform.

So wondering where I have gone wrong here.  I am a hobbyist,  I thought using the MIKTOR and high-density probes would result in a cleaner workbench and cleaner signals vs  flying leads.

[yyzkevin]

When using the E5346A probes I am no longer having any issue, previously I was having issues with the eye finder / thresholds as I have mixed 3.3v and 5.0v logic on the pcmcia bus,     using the E5339A probes I did not have that issue and was able to  properly identify high/low signals for both the 3.3v and 5.0v,  but it came with the other mysterious issue of crashing the setup, which still happens if I  use them.

So I am unsure why the E5346A probes work now and did not work before, with the same threshold and position values I had used previously.  I had tried all different arrangements of pods and cables so I am fairly sure it is not  a broken ground connections etc. 

[alm]

I made an extender board for a project I am working on, with the intention of being able to  measure all the signals on a pcmcia card under test.  We are not talking cardbus here, so this is ISA like speeds.

What are the edge rates (as measured by a scope), however? If the laptop also supports Cardbus speeds, then the edge rates will likely be higher than the older ISA bus. If you haven't, read the HPAK appnote Probing Solutions for Logic Analyzers, in particular the part about the 40-pin Mictor probes starting at page 23. Is the stub length between PCMCIA bus and Mictor connector below the limit they specify?

Setup is an old Agilent 1680AD Logic Analyzer with E5339A high density probes, and a custom PCMCIA extender board with the 38pin MIKTOR connectors.  I am operating in state acquisition mode with the read/write signals serving as the clocks and card enables as the clock qualifiers.

The 5339A is specified to have a dynamic range of +/- 5 V around its threshold voltage. I wonder if there's diode clamping going on beyond that voltage that's upsetting the bus, although it seems unlikely unless you have an odd trigger setting like -3 V? I guess you could try feeding a variable DC voltage into the 5339A with a DMM measuring current in between, and see if the current goes up substantially at the voltages you observe on the bus. Just don't exceed the maximum input voltage (+/- 20 V).

If I plug the extender board into the PC along with a card and no probes attached at all everything is fine.
If I plug probes into the MIKTOR connectors but do not attach them to the  logic analyzer, also there is no problem
If I plug the probes into the logic analyzer though,    it will cause corruption but not consistently, for lack of a better explanation only when there is a lot of activity it will cause the system to lock up.

This is puzzling to me. Connecting the logic analyzer should barely change the capacitance but mostly add a parallel resistance of 100 kOhm (E5346A) or 50 kOhm (E5339A), both of which I can't imagine would make much of a difference on a digital bus.

So wondering where I have gone wrong here.  I am a hobbyist,  I thought using the MIKTOR and high-density probes would result in a cleaner workbench and cleaner signals vs  flying leads.

Flying leads are definitely not much fun for parallel buses.

When using the E5346A probes I am no longer having any issue, previously I was having issues with the eye finder / thresholds as I have mixed 3.3v and 5.0v logic on the pcmcia bus,     using the E5339A probes I did not have that issue and was able to  properly identify high/low signals for both the 3.3v and 5.0v,  but it came with the other mysterious issue of crashing the setup, which still happens if I  use them.

The E5346A is specified for signals down to 500 mV p-p, while the E5339A is specified down to 250 mV p-p. So it's odd that the E5346A has problems with 3.3 V logic. Are you setting the probe correctly on the logic analyzer (E5339A or general purpose)? Is there any termination network between the Mictor connector and the PCMCIA bus as you would use for the E5351A, as described on page 28 of the Probing Solutions for Logic Analyzers document?

So I am unsure why the E5346A probes work now and did not work before, with the same threshold and position values I had used previously.  I had tried all different arrangements of pods and cables so I am fairly sure it is not  a broken ground connections etc.

Did the probe setting (5339A / general purpose) change on the logic analyzer? It will not automatically detect this, you have to change it manually, and it's easy to forget if you switch back and forth.