Long SPI acquisition

Post any questions you may have about our current range of oscilloscopes
Post Reply
TeeEff
Newbie
Posts: 0
Joined: Mon Nov 19, 2018 10:32 am

Long SPI acquisition

Post by TeeEff » Mon Nov 19, 2018 11:11 am

Hello,

I would like to spy a long SPI transaction. The SPI clock is 60ns (16.6MHs).
If I set the picoscope for a 20ms/div acquisition the sampling rate is 41.7 MS/s, sampling time is 24 ns and it's enough to decode SPI but I only get a 200 ms acquisition and I need more.

If I switch to 50ms/div the sampling time reach 56 ns that is not enough to decode.

I have an 3406D MSO PicoScope that should have a 1GS/s sampling rate and a 512 MS memory.

Even if I set the sampling number to 2GS and trig on Single, I'm still limited to 8.9MS.

How I could use the 512 MS memory with the 24 ns sampling time ?

Regards

Martyn
Site Admin
Site Admin
Posts: 3469
Joined: Fri Jun 10, 2011 8:15 am
Location: St. Neots

Re: Long SPI acquisition

Post by Martyn » Mon Nov 19, 2018 11:42 am

To get the maximum use of memory you would need to select a "Single Trigger". All other modes will segment the on board memory into two so that there is one segment holding collected data and one buffer collecting new data.
Martyn
Technical Support Manager

TeeEff
Newbie
Posts: 0
Joined: Mon Nov 19, 2018 10:32 am

Re: Long SPI acquisition

Post by TeeEff » Mon Nov 19, 2018 1:27 pm

Thanks for your prompt reply Martyn but if I set trigger to single or repeat I get the same result.

When does the number of sample should be set ? Because it return to 10 MS when I start the acquisition.

TeeEff
Newbie
Posts: 0
Joined: Mon Nov 19, 2018 10:32 am

Re: Long SPI acquisition

Post by TeeEff » Mon Nov 19, 2018 1:41 pm

Hello,

My picoScope interface was in French and I switch it to English in order to have the same name for the configuration as you requested.
And now it's OK. When I change the number of samples to 2GS it keep the value and I could set a long (10s) acquisition with a 24 ns sample interval. (With a Slow Sampling Transition set to 100 s/div).

So the French interface is not good. It's seems to me that we already have this king of problem.

Martyn
Site Admin
Site Admin
Posts: 3469
Joined: Fri Jun 10, 2011 8:15 am
Location: St. Neots

Re: Long SPI acquisition

Post by Martyn » Mon Nov 19, 2018 2:43 pm

It was possible that you had channel filtering or resolution enhancement enabled, which will limit you to 10MS, and these were turned off when you changed language.
Martyn
Technical Support Manager

TeeEff
Newbie
Posts: 0
Joined: Mon Nov 19, 2018 10:32 am

Re: Long SPI acquisition

Post by TeeEff » Mon Nov 26, 2018 2:07 pm

Hello,

I have got a new problem now. On Friday I only take care of the Sample interval but today when I would use the SPI Serial Decoding it doesn't work correctly with a time resolution upper than 50 ms/div.

With a time resolution of 50 ms/div, I have tested different values of Number of Sample. Down to 20MS the SPI trace could be decoded. This correspond to a Sample interval of 32ns while my SPI clock is 60ns (16.66 MHz).
But if I increase the time resolution to 100 ms/div, the SPI trace is not correctly decoded even I chose the lowest sample interval (250 MS/s -> 4ns).

I spy the SPI line with Digital inputs (CLK, CS, MTSR and MRST) which look correct but the decoding is wrong. Picoscope always see only 1 byte even when I receive 4.

Code: Select all

PicoScope® 6 - PC Oscilloscope software version: 6.13.11.3952
Copyright © 1995-2018, Pico Technology Ltd.

Model: PicoScope 3406D MSO
Serial Number: CZ765/060
USB Version: 3.0
Calibration Date: mercredi 8 avril 2015
Hardware Version: 1
Driver Version: 2.1.4.789
Firmware Version: 1.7.5.0 / 1.0.66.0

Post Reply