Install new ecosystems

[HenrikMay_IPM]

Hey everyone,

I am new to the whole redpitaya topic and also quite a beginner in linux. I didn't succeed in making the redpitaya run on the latest OS but i had an SD card with 0.94 and made this one run. I was able to connect to the pitaya via LAN and SSH and played around a little with Matlab and die SCPI-server. So far so good.
My goal is to acquire a fast continuous stream of data as described here. Therefore I need to run an older ecosystem on the redpitaya. The question is now: How do I install it if it is not uploaded as an .img file (preferably with windows?). Any ideas why so few OSs work for me? Is it because I use windows (respectively WIN32DiskImager) to load the SD card?

Thanks in advance!
Henrik

[Nils Roos]

Hi Henrik,

installing the older ecosystem version is fairly easy. Insert the card into your Windows computer, open the only drive (partition) that it contains (when viewed in Windows), delete everything and then unzip the ecosystem zipfile in there. Safely remove the card and that's it.

Any ideas why so few OSs work for me?

What exactly is not working? Are you unable to establish a network connection after they have started or do they not start?
Have you tried to connect your Red Pitaya to your PC with the USB console? If not, try it and take a look at the output - eg. with PuTTY - or post the output here.

My goal is to acquire a fast continuous stream of data [...]. Therefore I need to run an older ecosystem on the redpitaya.

Not neccessarily, see here.

[HenrikMay_IPM]

Hey Nils,

thanks a lot for the quick reply and the patience. As I said I am starting from the very beginning here.
I got a big step further in the last couple of days: The reason I couldn't communicate with the rp was because windows messed up my network configuration. Having fixed that, I was able to:
1. Put the old ecosystem (0.92-81) on the SD card
2. Start and use the command rp_remote_acquire mentioned in the linked thread above.
3. Transmit, Receive and plot the signal with Matlab.

What I'd like to do now is to add some sort of clock, trigger, timestamp or increasing number to each sample (or one every certain number of samples) in order to distinguish which output sample belongs to which input and to spot missing packages (using UDP for the transmission). The goal is to sample at a decimation factor of 8.
Do you have any good ideas or suggestions?
I know this is very related to the old thread linked above so if you want, I can move to that one after your answer.

[Nils Roos]

You could either prepend a counter to the samples in the routines in transfer.c before they are handed over to the network stack. Or you change the fpga logic to insert a counter in regular intervals when they are written to RAM.

Changing the C code is certainly the simplest way, but I am unsure whether breaking up the page alignment of the buffers has negative performance impacts. You'd have to find out yourself, I guess. For starters maybe just add a send() call with the counter before the send_buffer() in transfer_buf_mmap() and see if it can still maintain the throughput on decimation setting 8.

If you think this could develop into a longer discussion that might be of interest to others, feel free to open a new topic about it.