Capture 2 minutes SPI traffic and save it

[pilot5599]

Hello,

I have a Rigol MSO5204 oscilloscope. I'm trying to figure out if it's possible to capture a 2 minutes SPI trffic and save it to the USB drive. Or the capture can be saved to Windows?
I managed to connect the scope to my Windows via LAN. I'm able to do remote control from the web browser in my Windows. But I don't know if it's possible to save a relative long period of captures directly to Windows. I tried the USB drive. But it can save the current screen.

I understand I can decode those 4 analog channels. If I connect my device via the digial channels to the scope. Can I decode those digital channels?

Thanks in advance!

[thm_w]

Its much easier to do this with a $10 sigrok USB logic analyzer. Depending on the speed of your SPI bus.

Yes you can decode on either the analog or digital inputs.

[Someone]

I have a Rigol MSO5204 oscilloscope. I'm trying to figure out if it's possible to capture a 2 minutes SPI trffic and save it to the USB drive. Or the capture can be saved to Windows?
I managed to connect the scope to my Windows via LAN. I'm able to do remote control from the web browser in my Windows. But I don't know if it's possible to save a relative long period of captures directly to Windows. I tried the USB drive. But it can save the current screen.

Scope has 200Mpts of memory, can capture >2min at 1M sample per second.

Where is the problem?

[thm_w]

OP says they are only getting the current screen data, so maybe the file format or settings they use to export are incorrect.
Either way its a bit of a hassle, as mentioned here:
https://www.eevblog.com/forum/testgear/rigol-mso5000-unrecorded-memory-garbage/
https://github.com/ngscopeclient/scopehal-apps/issues/287

[dietert1]

I would try a STM32 nucleo board with 2 SPIs to sniff transactions on both MOSI and MISO and send it to a windows or linux host via USB continuously. Everybody uses what they are familiar with.

Regards, Dieter

[TizianoHV]

+1 for sigrok, it is easy to use and you can save directly logs to your pc and review them later.

[tooki]

Another useful tool, if one does SPI/I2C bus debugging a lot and can afford it, is a tool like the Aardvark, whose software is designed around bus debugging. In particular, compared to a pure logic analyzer, you can also write things to the bus as needed. But there are many, many ways to skin this cat, including DIY solutions like the Nucleo board dieter1 mentions (or any other fast MCU with enough buses for your situation).

[nctnico]

I have a Rigol MSO5204 oscilloscope. I'm trying to figure out if it's possible to capture a 2 minutes SPI trffic and save it to the USB drive. Or the capture can be saved to Windows?
I managed to connect the scope to my Windows via LAN. I'm able to do remote control from the web browser in my Windows. But I don't know if it's possible to save a relative long period of captures directly to Windows. I tried the USB drive. But it can save the current screen.

Scope has 200Mpts of memory, can capture >2min at 1M sample per second.

Where is the problem?

The problem is decoding only what is on screen instead of full memory. Then the question is how many SPI messages need to be captured; the number of decoded messages is typically limited as well. Last but not least, 1MS/s may not be enough. OTOH, if there is a lot of dead time between the SPI messages, segmented recording can be an option and requesting each (decoded) segment through a piece of software (Python script).

[Someone]

I have a Rigol MSO5204 oscilloscope. I'm trying to figure out if it's possible to capture a 2 minutes SPI trffic and save it to the USB drive. Or the capture can be saved to Windows?
I managed to connect the scope to my Windows via LAN. I'm able to do remote control from the web browser in my Windows. But I don't know if it's possible to save a relative long period of captures directly to Windows. I tried the USB drive. But it can save the current screen.

Scope has 200Mpts of memory, can capture >2min at 1M sample per second.

Where is the problem?

The problem is decoding only what is on screen instead of full memory. Then the question is how many SPI messages need to be captured; the number of decoded messages is typically limited as well. Last but not least, 1MS/s may not be enough. OTOH, if there is a lot of dead time between the SPI messages, segmented recording can be an option and requesting each (decoded) segment through a piece of software (Python script).

Perhaps, lots of guessing/imagination on your part. The OP needs to explain what limitations they are coming up against.

Can it record 2 minutes of data and off load that to usb or PC? yes. That was the question posed.

[kripton2035]

sigrok pulseview and this : https://www.aliexpress.com/item/1005005698409712.html

[dietert1]

The scope variant will be limited to 200 MSamples. In order to log 2 minutes that is about 1 MHz sampling, which isn't good enough for most SPI. Often SPI works at clock frequencies of several MHz up to 50 MHz or so.
That's why i thought about making a SPI sniffer, that can handle those higher clock rates without producing a vast overhead.
It could use some GPIOs as inputs for the nCS signals to tag recorded SPI transactions per bus device.

Regards, Dieter

[TomKatt]

The $149 DSlogic Plus LA can stream 3 channels at 100MHz (or 6 channels at 50MHz) sampling rate over usb to pc for ~160 seconds.  The $299 DSLogic U3Pro16 is USB3 and can stream 6 channels at 500MHz.  Their DSview software is based on Sigrok Pulseview and has basically the same feature set, including SPI decoding.

Depending on the task, I often switch between scope and LA for serial comms work.  The LA is great for decoding, but only the scope can identify signal issues (timing, missing or runt pulses, etc).

[pilot5599]

Thanks everyone!
I figured out the way to save the memory in csv format. I also found out how to configure the scope to decode the digital channels.
Thanks again!