Bug with One Shot blocks?

Hi Christian,

I got exactly the same results on my 1701 mini board -- the one-shots do interact with each other in a bizarre manner. Substituting additional pulse counters for the one-shots works around this problem (see the attached project).

Best regards,

Bob

Hi Bob,

it works in general now but unfortunately this has a little drawback.

I use the Start1 for starting the bit pattern by setting it to 1. Start1 stays at 1 until the next time I need the bit pattern. Then I set it to 0 and back to 1.

This means if I use pulse counters instead of one shot blocks and the Start1 stays at 1 the pulse counter output is reset to 0 but immediately steps to 1 again (Start pin is 1 and pulse input (=Start1) is 1) if the reset gets back to 0.

Regards

Christian

Hi Christian,

Sorry I didn't think of that. Anyhow, ADI needs to investigate and fix the strange behavior of one-shots with the Sigma100. It does appear that when you use two one-shots in a project, they interact exactly the way you pointed out.

Bob

Hi,

I have a different project with 2 one shot blocks that works as expected.

My assumption is that a combination of one shot and feedback blocks causes this behaviour.

Hope that somebody from ADI reads that...

Christian

Hi Christian:

We are looking into this problem.

We'll get back to you when we have a solution.

Miguel

Hi Jun,

so it will be in the 3.10 (Beta) Release?

Is there a planned date for the next release version or Beta?

Regards

Christian

Hi Christian,

Thanks your feedback. It is do a bug which caused by a continuous data memory definition with input pin name. The good news is that it is fixed now. And you can get it in next release version.

Jun

Hi Christian,

It could be included in the 3.10.4 Beta version.

But I don't know if there a plan recently. Maybe Miguel know.

Thanks,

Jun

Hi Christian,

One more thing should remember is that you need to drop again the newer "OneShot Rise, Reset" block in the schematic for the changes to take effect once you get a newer version of Sigma Studio.

Thanks,

Jun

Hello again,

I replaced now all one shot blocks with other blocks but still have a problem.

I extended my project with a sine going to DACs, reading data on the audio ADCs and running a bandpass filter.

The filter generates an output even with no input signal (ADCs mute or power down).

I need a solution / workaround for that issue as we have a tough project plan for that device.

Is it really the one shot block that causes a problem or do I have to replace also other blocks?

Thanks

Christian

Hello Christian,

What is the nature and level of the noise you're experiencing? If for example it's a strong white noise, I had that problem with a large project that used feedback extensively (as you're doing). Adding a filter not even in the audio path caused all DACs in use to output an awful hiss. The fix was to change the filter to a different one with a similar response, see https://ez.analog.com/message/63536#63536

If on the other hand, the noise is very low level, be sure you're using double-precision filters (once in a project, single and double precision filters appear the same). See https://ez.analog.com/message/131118#131118

Hopefully the promised revision of SigmaStudio may also help as this will allow you to return to your original design using one-shots.

Best regards,

Bob

Hello Bob,

I get a full-scale noise signal on the DACs. I always use double precision filters.

In the meantime I did a redesign using a "brute force" approach by generating the bit pattern with 39 "Tolerance Analyzer" blocks - which is not very funny. I've reduced the number of feedback blocks to a minimum (2 with very short feedback path) and it works as far as I can see up to now.

I'm also hoping the release will fix that issues.

Regards

Christian

Hello Christian,

I will send you the dlls that you need shortly. Please let me know if those dlls helped you.

Thanks,

Wilfrido.