Skip to main content
Topic: [SOLVED] Cannot for the life of me make this work (Read 3978 times) previous topic - next topic

[SOLVED] Cannot for the life of me make this work

First, I just want to say that I'm really psyched about the potential this device represents to hobbyists, for debugging complex MCU-based devices.  And also that as soon as I heard SeeedStudio was taking preorders, I jumped right on it. (Even though at the time it meant eating ramen for the rest of the week, hehe.)

Second, it is entirely possible I have no earthly idea what I am doing, I'll be the first to admit that. 8)

Now, when my OLS first arrived, my cables were still about a day or two from leaving HK, so I had to make do with F/M breadboard jumpers and other thrown-together cabling, making it hard to determine if the OLS was in fact working or if there was an issue with the collection of desk detritus I used to connect it...  I was testing it on communications between an Arduino and a couple ShiftBrites, since I at least have a pretty good idea what the results SHOULD look like.  But everything I was getting back from the OLS was strange-- either all low, or seemingly random signals --and then it started giving me 'device not found' when I tried to capture.  So I gave up for a couple of days.

Today, my cables arrived, so I gave it another shot.  It still wasn't working, so I re-burned the firmware with the tools included with the driver package.  Then it was recognized again, but wasn't capturing anything,  (All signals low.)  I updated the firmware again, to the new test version, and used the accompanying version of the client.  It stopped giving me communication errors, but it still wasn't capturing anything, although occasionally signals would appear on channels that weren't connected.  (Much to my dismay.)

The short of it is, I have no idea if I did something wrong, am still doing something wrong, or if I got a flaky unit.  (Though, my money is on me being dumb in some way.  As that's always a good bet.)

Some help would be greatly appreciated.  Bonus points for a quick primer on how I should be using it!  (I know I'm dumb, I just want to make sure I'm not really really dumb.)

Re: Cannot for the life of me make this work

Reply #1
Hello,

Well, here are a couple pointers.

-The default bitstream that ships with the OLS is only 4k deep. So if you are selecting more than 4k as the sample depth then you will get strange results. If you want more depth and less channels then you need to load another bitstream to the FPGA-ROM. We are working on an update to make it integrated with the client.
-It helps to make sure that pins adjacent to the pins you are sampling are grounded, if the pins float then they will sometimes pick up the signal next to them and it becomes confusing.
-I just released a test build that includes a test mode. I'm attaching it to this message, if you update the FPGA_ROM with the attached one and use the attached analyzer.jar then you should be able to use test mode. Connect a ribbon cable between the Wing header and the Buffered header and you will see a very distinct test pattern on 16 channels. Please keep in mind that this test mode is not well tested!

Jack.

Re: Cannot for the life of me make this work

Reply #2
Okay, I see now that setting the sample size too large definitely causes it to start crapping all sorts of madness. :)

And this is what the test mode gave me:

Re: Cannot for the life of me make this work

Reply #3
Yep, thats what you should see!

Jack.

Re: Cannot for the life of me make this work

Reply #4
[quote author="jack.gassett"]
Yep, thats what you should see!
[/quote]

Well, that's encouraging at least...

Still gives me all sorts of weirdness when it's hooked up to anything else though.  I'll have to try messing with it more.

Re: Cannot for the life of me make this work

Reply #5
Try sampling with nothing connected except a single line pulled high and see if you get the expected result.

Re: Cannot for the life of me make this work

Reply #6
Okay, it definitely seems to work.  I just had to play around with the sampling frequency and stuff.  And turning on RLE seems to make things go all explodey...I don't know what that's about.

It also seems my SeeedStudio cables are really noisy and full of crosstalk...  I'll have to make my own and re-use the test clips.  I like the idea of individual probes anyway.

It's times like this I'm thankful for owning a nice ratcheting pin crimper. :)

Re: Cannot for the life of me make this work

Reply #7
If you see crosstalk from the probes, try to
a) get familiar with the OLS by testing things with rather low frequency signals first
b) connect all probes you don't use to GND - i.e. put some wire loop on one of the GND pins on th OLS and clip all unused probes to it - don't let them "dangle in the air".

Re: Cannot for the life of me make this work

Reply #8
[quote author="Bakamoichigei"]
And turning on RLE seems to make things go all explodey...I don't know what that's about.
[/quote]

Confirmed - turning on RLE just generates noise for me, even on pins that are tied to ground.

Re: Cannot for the life of me make this work

Reply #9
[quote author="alanbur"]
Confirmed - turning on RLE just generates noise for me, even on pins that are tied to ground.
[/quote]

Yep.  That definitely sounds like what it's giving me with RLE on.  Just some horrible mess, heh.

Anyway, I've got this thing working now, just not with the SeeedStudio cables.  So I'mma mark the thread [Solved].

Thanks guys. :)

Re: [SOLVED] Cannot for the life of me make this work

Reply #10
I see just random transitions even with the pins grounded on rle mode.  It appears to transfer, but the data is bogus.

Re: [SOLVED] Cannot for the life of me make this work

Reply #11
There is definitely something wrong with the RLE option. I suggest not to use it for now. The priority is on making the communication between the Java client and the SUMP engine inside the FPGA more stable (and performant) before fixing the RLE option.

Re: [SOLVED] Cannot for the life of me make this work

Reply #12
Understood, just want to make sure it was understood that rle was broken for me, I did not see it in Ian's status post.  It is clearly secondary.