timing in ni usb modules (6009 and 6501)

[kristos_b]

Hi Lads,

I have noticed something strange while developing some code for ni usb6009 (this same behavior on 6501). I want to load some digital patterns using digital i/o's. I dont really mind that samples are flushed out quite irregular (software timing) but there is one condition for which this module acts quite weird. It seems that when samples on any line are not changing their states, they are either ignored or flushed very fast (I dont know which is true). I hope example below will make it clearer.

I trigger the scope with P0.4->channel 1 on the scope.

This is what I am sending:

This is what I am getting on the scope (1->P0.4, 2->P0.3, 3->P0.2, 4->P0.1)

Than I push the trigger 4 clocks further:

And result is exactly this same as before.

Here I put trigger sooner when something is happening on the lines

And it works fine:

Did anybody came across problem like this?

Below is the code I am using to send the waveform. I have tried to replace waveform with 2d Boolean and send each bit in separate 'for loop' iteration but it didn't help.

Thanks for help,

Krystian

[LAVA 1.0 Content]

Hi Krystian,

Somehow I couldn't see the problem, could you make it a little bit more clear?

Ton

[kristos_b]

QUOTE (Ton @ Jul 7 2008, 07:23 PM)

Hi Krystian,

Somehow I couldn't see the problem, could you make it a little bit more clear?

Ton

Simple case is to write to one output line pattern i.e. 010101011111111111111111111010101 and on the scope you will see 010101011010101.

[LAVA 1.0 Content]

QUOTE (kristos_b @ Jul 7 2008, 09:00 PM)

Simple case is to write to one output line pattern i.e. 010101011111111111111111111010101 and on the scope you will see 010101011010101.

But I don't see that in the code/screenshots, maybe you should widen the scope frame?

Could you specify which code/image you mean?

In het second set I count 9 '1' on p0.2 while p0.3 is high, this is perfectly reflected by channel 2 and 3 on the scope image.

Wait, I see on the first set that the dip on channel 1 (p0.4) is too early.

But this is all with software timing?

I never trust those

Ton

[kristos_b]

QUOTE (Ton @ Jul 7 2008, 08:06 PM)

But I don't see that in the code/screenshots, maybe you should widen the scope frame?

Could you specify which code/image you mean?

In het second set I count 9 '1' on p0.2 while p0.3 is high, this is perfectly reflected by channel 2 and 3 on the scope image.

Wait, I see on the first set that the dip on channel 1 (p0.4) is too early.

But this is all with software timing?

I never trust those

Ton

I guess you can not expect much from software timing but still this looks pretty weird. Specially that the rest looks quite regular.

[LAVA 1.0 Content]

QUOTE (kristos_b @ Jul 7 2008, 09:25 PM)

I guess you can not expect much from software timing but still this looks pretty weird. Specially that the rest looks quite regular.

I believe that digital data uses RLE encoding (if nothing changes nothing is stored), maybe you see a side effect of that.

Ton