Skip to main content ↑ ↓
Topic: Selftest result OLS (Read 3000 times) previous topic - next topic

Selftest result OLS

Since the testrelease topic is a bit clogged up, I opened a new topic.  In this topic we discuss the selftestresults. Selftest result is adding in bitstream 2.04 by Jack.gasset. He also updated the SUMP client to use this mode.

To use the new testmode upgrade to the new SPI firmware (I used 0.3) and the new 2.04 bitstream. Upgrade directions are in http://dangerousprototypes.com/forum/in ... opic=576.0 and on the blog http://dangerousprototypes.com/2010/05/ ... rs-needed/

I've running bitstream 2.04 and FW 0.3.

The simplest way to use the new client is to download and extract the BFP_Logic_Analyzer_1.0.zip from the gadgetfactory.com and copy the new analyzer.jar into this extracted directory (in my case  D:BFP_Logic_Analyzer_1.0 ). [s:]I also renamed the analyser.jar to analyzer2.jar.[/s:] this seems to be a bad idea see the post made sdixon belowI run the updated client with this command from a cmd-box:

Code: [Select]
D:BFP_Logic_Analyzer_1.0>jre1.6.0_14-b08binjava -jar analyzer.jar

Linux/mac/bsd/* should be simular ( s/\/ //g ).

Select the 'test mode' in 'numberscheme' (see ols_capture.gif) connect the unbuffered inputs to the buffered inputs (inside 0-15 to inside 16-31) and press capture. Jack.gasset provided a know good capture (see test_waveform.png).

I ended up with a slightly different one (see ols_spikes.gif and ols_spikes.sla.text). I think it is caused by long lead and having the second connector ('Wingslot') not soldered. I don't have a proper header, so need to get me one and try again. When I have a proper header I'll repost my results.


[s:]BTW: the .sla extension is not allowed, and trying to post nevertheless will sedn your post to /dev/null and you need to start all over!! Grrr. (so I present you post 2.0 ;))[/s:]

Re: Selftest result OLS

Reply #1 –
Hi,
just ran the the Selftest feature (wiring done with 17 breadboard patch-wires) .
I had similar spikes on one a channel initially, but they went away after adjusting the patch-cable on the header.

So SelfTest works up to 4k samples.
I have my doubt that it works at 8k memory usage even though the board returns the data without errors:
I did my first tests with the board with a slightly more complex signal that is produced by an arduino:
The signal produces an increasing number of spikes (1-9 then back to one) intermitted by a fixed length delay.
This is shown in the 4ksignal attachment.
You see the number spikes  increasing from 1 to 5 (this would go on up to 9 spikes, then back to 1 and the signal repeats.)

The 8ksignal capture looks good at first, but look at the number of spikes starting with the trigger we have
9-1-2-3-3-9-1-2-3
I did some more captures at 8k and it looks like the data returned by the board is a repetition of the first 4k of data.
I also verified this with a capture of I²C data.
It is a bit hard to spot with the testmode signal since it repeats so regulary.
Is there a chance to create a more complex testmode signal on the board?

Eberhard

Re: Selftest result OLS

Reply #2 –
Quote
BTW: the .sla extension is not allowed, and trying to post nevertheless will sedn your post to /dev/null and you need to start all over!! Grrr. (so I present you post 2.0 ;))

I hate that! I hope they fix it in the next release. I recently added .brd, .sch, .c, and .h to allowable. I will add .sla too.
Got a question? Please ask in the forum for the fastest answers.

Re: Selftest result OLS

Reply #3 –
It is completely my fault: I should read that only a couple of extensions are allowed.

Could you select a blacklist of extensions instead of allowable extensions? e.g. only deny .exe .vbs. .js .dll?

Re: Selftest result OLS

Reply #4 –
[quote author="Sjaak"]


The simplest way to use the new client is to download and extract the BFP_Logic_Analyzer_1.0.zip from the gadgetfactory.com and copy the new analyzer.jar into this extracted directory (in my case  D:BFP_Logic_Analyzer_1.0 ). I also renamed the analyser.jar to analyzer2.jar. I run the updated client with this command from a cmd-box:

Code: [Select]
D:BFP_Logic_Analyzer_1.0>jre1.6.0_14-b08binjava -jar analyzer2.jar

Linux/mac/bsd/* should be simular ( s/\/ //g ).


[/quote]

In looking at the Java code for the SUMP client, I see that there are two places where the specific filename "analyzer.jar" is opened during the process which constructs the menus for the GUI.  Basically, the code is going through the Java class files in the jar file and looking for supported devices and also tools (for the Tools menu).  So it is not a good idea to rename analyzer.jar unless you also change those two places in the code to correspond.  If you don't do this, one of two things will happen:  Either SUMP will fail to run correctly and give an exception (java.util.zip.ZipException: error in opening zip file) in which case the menus will fail to work correctly.  Or else it will work sort of accidentally when you happen to still have a file named analyzer.jar in your working directory.  In that case,  you will get some class files from your new jar file and some from the old analyzer.jar file.  Needless to say, that is unlikely to be a great idea.
So probably the best advice is to always keep the name analyzer.jar constant unless you are willing to modify the Java code.  Use separate working directories if you want to have multiple versions around rather than using different filenames for different versions.

Re: Selftest result OLS

Reply #5 –
Ooops, didn't know that. I seems to work and I got new menu item so I assumed it is ok.

I'll correct the startpost to copy the new analyzer.jar instead of renaming it.

Thanks for pointing it out.