I don't have the skills to give input. The stuff that Saucy posted on filter performance when we were adding SincX filters went straight over my head.
I think it went through mine, but left little trace.
Saucy, can you refresh us?
I think the idea was that rather than dwell on a certain ratio, which creates a dead spot, of sorts, the integrator would always be meandering above and below the midpoint.
I should start testing the external bitstream feature ...
Hmm, grr, going back to Pnut is hurting my mental well-being. I like using Eric's #include "more-code.spin2" but I can't rely on it assembling correctly for the revB chips yet. I have all my helper subroutines in one file where any little fixes are naturally carried to all test programs that include that file.
I should start testing the external bitstream feature ...
Hmm, grr, going back to Pnut is hurting my mental well-being. I like using Eric's #include "more-code.spin2" but I can't rely on it assembling correctly for the revB chips yet. I have all my helper subroutines in one file where any little fixes are naturally carried to all test programs that include that file.
I should start testing the external bitstream feature ...
Hmm, grr, going back to Pnut is hurting my mental well-being. I like using Eric's #include "more-code.spin2" but I can't rely on it assembling correctly for the revB chips yet. I have all my helper subroutines in one file where any little fixes are naturally carried to all test programs that include that file.
If you don't mind being a guinea pig here's a version of fastspin that can compile for the new revB chips. Just pass "-2b" on the command line. To get the original silicon use "-2a". For now plain "-2" is the same as "-2a", but that will change once more of the new silicon is out in the wild.
I think the only instruction encoding changes are:
- rdlut/wrlut accept PTRA and PTRB special forms
- the PTRA/PTRB forms allow offset from -64 to +63 if there is no update specified
- new instructions setscp and getscp
AFAICS the other changes are compatible (e.g. the extra bits in drvh/drvl/etc.).
Oh, I'll be wanting the Linux binary of fastspin please.
It'll be in the "build" folder of your git repo after you do a "git pull; make" .
On the serious side I've pretty much assumed that most Linux users of fastspin are comfortable with (and perhaps even prefer) building it from source. If enough people tell me I'm wrong I can revisit that.
I should start testing the external bitstream feature ...
Hmm, grr, going back to Pnut is hurting my mental well-being. I like using Eric's #include "more-code.spin2" but I can't rely on it assembling correctly for the revB chips yet. I have all my helper subroutines in one file where any little fixes are naturally carried to all test programs that include that file.
If you don't mind being a guinea pig here's a version of fastspin that can compile for the new revB chips. Just pass "-2b" on the command line. To get the original silicon use "-2a". For now plain "-2" is the same as "-2a", but that will change once more of the new silicon is out in the wild.
I think the only instruction encoding changes are:
- rdlut/wrlut accept PTRA and PTRB special forms
- the PTRA/PTRB forms allow offset from -64 to +63 if there is no update specified
- new instructions setscp and getscp
AFAICS the other changes are compatible (e.g. the extra bits in drvh/drvl/etc.).
Did I miss anything?
Eric, the PTRA/B non-updating offset range is -32 to +31. Look at the latest Google doc for the exact encoding. Otherwise, your list of instruction encoding changes is correct.
Will this revision be available on an Eval board for us early adopters, or will we have to wait for the finished product? If so, when will it be available? (sorry if I missed the information somewhere in all threads)
Will this revision be available on an Eval board for us early adopters, or will we have to wait for the finished product? If so, when will it be available? (sorry if I missed the information somewhere in all threads)
/Johannes
Johannes, as soon as we get more chips, we will make new P2 Eval boards with them. We are expecting over 1,000 new chips. Waiting for ON Semi to find out the delivery date.
...as soon as we get more chips, we will make new P2 Eval boards with them. We are expecting over 1,000 new chips. Waiting for ON Semi to find out the delivery date.
...as soon as we get more chips, we will make new P2 Eval boards with them. We are expecting over 1,000 new chips. Waiting for ON Semi to find out the delivery date.
I really like the time is coming to start the next generation of work. In the mean time I like to watch others playing with words. So if iirc there are some links between the propeller 1 and Finland. So we have a finish propeller 1 and an almost finished propeller 2. So funny.
I really like the time is coming to start the next generation of work. In the mean time I like to watch others playing with words. So if iirc there are some links between the propeller 1 and Finland. So we have a finish propeller 1 and an almost finished propeller 2. So funny.
Oh sh.. does Ken knows that? He need to send you a P2 to finally get it finished!
Comments
I think it went through mine, but left little trace.
Saucy, can you refresh us?
I think the idea was that rather than dwell on a certain ratio, which creates a dead spot, of sorts, the integrator would always be meandering above and below the midpoint.
Thanks, Saucy.
Just search for "ADC Scope":
https://docs.google.com/document/d/1gn6oaT5Ib7CytvlZHacmrSbVBJsD9t_-kmvjd7nUR6o/edit?usp=sharing
Hmm, grr, going back to Pnut is hurting my mental well-being. I like using Eric's #include "more-code.spin2" but I can't rely on it assembling correctly for the revB chips yet. I have all my helper subroutines in one file where any little fixes are naturally carried to all test programs that include that file.
With the few tricks I've learnt over the past 6 months I find that I can save a few more longs in the SD and Monitor ROM
If you don't mind being a guinea pig here's a version of fastspin that can compile for the new revB chips. Just pass "-2b" on the command line. To get the original silicon use "-2a". For now plain "-2" is the same as "-2a", but that will change once more of the new silicon is out in the wild.
I think the only instruction encoding changes are:
- rdlut/wrlut accept PTRA and PTRB special forms
- the PTRA/PTRB forms allow offset from -64 to +63 if there is no update specified
- new instructions setscp and getscp
AFAICS the other changes are compatible (e.g. the extra bits in drvh/drvl/etc.).
Did I miss anything?
Only thing I can see that might be missing is GETCT WC
It'll be in the "build" folder of your git repo after you do a "git pull; make"
On the serious side I've pretty much assumed that most Linux users of fastspin are comfortable with (and perhaps even prefer) building it from source. If enough people tell me I'm wrong I can revisit that.
Eric, the PTRA/B non-updating offset range is -32 to +31. Look at the latest Google doc for the exact encoding. Otherwise, your list of instruction encoding changes is correct.
Will this revision be available on an Eval board for us early adopters, or will we have to wait for the finished product? If so, when will it be available? (sorry if I missed the information somewhere in all threads)
/Johannes
Johannes, as soon as we get more chips, we will make new P2 Eval boards with them. We are expecting over 1,000 new chips. Waiting for ON Semi to find out the delivery date.
If you want one, there will be plenty available.
Yes, we are supposed to receive at least 1,000 chips on October 18th. We are scheduled to make 192 P2 Eval boards then.
Yes.
They will pass ON's digital tests and Parallax' analog pin tests. No V-stress test will be applied.
Oh sh.. does Ken knows that? He need to send you a P2 to finally get it finished!
Enjoy!
Mike