Zero not being zero....

[Norse]

Hi,
I have a consistent and puzzling discrepancy between my ISOPOD v1 and v2 .
------------------------------------

V1:

SCRUB
IsoMax V0.6
OK
FVARIABLE A OK
0.0E0 A F! OK
A F? 0.0000E-32768 OK

V2:

SCRUB
IsoMax V0.6
OK
FVARIABLE A OK
0.0E0 A F! OK
A F? 0.0000 OK

Any thoughts about this?

Cheers,

Per.

[nmitech]

Here is my result of IsoPod V1,

COLD
IsoMax V0.6
FVARIABLE A OK
0.0E A F! OK
A F? 0.0000 OK

Let me email you my srecord & the CFG file for you to reflash.

[Norse]

That was fast.... :-)
I got your zip file and will reflash the pod later today.

Hopefully that will correct the error.

Cheers,

Per.

[Norse]

Hi Again....
When I run the isopod batch-files I get a Jtagflash.exe error about "The Exeption Priviliged Instruction. etc..... " and a "click to terminate program. I run WIN2000pro

I saved the provided 6.1 files in a separate folder from the 6.0. However all the needed files were included in the folder.

As as a lark, I went back to the 6.0 folder and reflashed the pods with the 805.cfg as provided a while back. No problem, it flashed and verified both pods.....

The Isopod v1 still shows the rather odd Zero value. It can however save a "zero" if I store a 0.000001E.....

Any thoughts?

Per.

[RMDumse]

I have a vague memory of a zero problem. Thought it was much older than V.6 though. If I can recall correctly, it was a problem with the print formating routine rather than the actual math. Think I satisfied my self by doing a PI FDUP F- and then test to see if that was zero, which of course it must be.

Well, until memory serves better, I wanted to mention that in case it helps any.

[Norse]

Interesting results..... :-)

Isopod V1:

IsoMax V0.6
PI FDUP F- OK
F. 0.0000E-32768 OK


Isopod v2:

IsoMax V0.6
PI FDUP F- OK
F. 0.0000 OK

Something to ponder over......

Cheers,

Per.

[RMDumse]

Now I have the straight scoop from the language programmer. From his notes:

v0.6 release, 6 jun 03
----------------------
Incorporated fixes for find_exp (used by E.) and UM/MOD.

v0.60, 28 may 03 (tentative)
----------------
...

It's possible that the v1 pod was shipped with the 28 may prerelease, and the v2 pod was shipped with the find_exp fix (which I believe would affect F. as well as E.). Upgrading the v1 pod to 0.61 should solve this, but note a later fix in version 0.63:

[Norse]

Hi again.

I did receive a zip file with 0.61 but I could not flash the pods with it as I got an error message. I went back to the 0.60 version files just to see if there was something wrong with my JTAG cable etc. The 0.60 flashed just fine. I am registered for updates. The old isopodv..bat first had to be changed to reflect the flash ini file supplied 0.61.
If I get a set of properly configured files then I can reflash to 0.61. or newer version. Hopefully that will correct the problem... :-)

Cheers,

Per

[nmitech]

Did you use the same flash80x.CFG file as the working one? Anyway, I will resend you the files again. Check your email.

[Norse]

ISOMAX 0.61 is now on both Isopods... :-)
However, Zero is still not Zero on V1..

COLD
IsoMax V0.61
PI FDUP F- OK
F. 0.0000E-32768 OK

It is just fne on V2

COLD
IsoMax V0.61
PI FDUP F- OK
F. 0.0000 OK

I may rewire my A/D inputs and shift all of it and the floating point stuff over to the V2 in order to avoid potential math problems.

Cheers,

Per

[nmitech]

:confused: My IsoPod V1 does not have any problem with V0.6 and V0.61. See my results below,

COLD
IsoMax V0.61
PI FDUP F- F. 0.0000 OK
0.0E F. 0.0000 OK
FVARIABLE ZERO OK
0.0E ZERO F! OK
ZERO F? 0.0000 OK
ZERO F@ F. 0.0000 OK

[Norse]

I guess we are both stumped...
It should obviously work just fine and I can't see it being caused by any processor harware flaw.
Oh well, I may give the V1 unit to the kids and get a Servopod a bit later on.
Any plans for a Tinypod with two or three A/D's? It may be useful for some ideas I am working on....

Keep up the good work!!!

Cheers
Per.

[Norse]

Good times are here again..... :-)

Chris, the last zip file you just emailed me fixed the problem!!!

PI FDUP F- OK
F. 0.0000 OK

Thanks for your great support!

Cheers,

Per.

[ngkdc]

Looks like my ServoPod has the same problem:

COLD
IsoMax V0.61
PI FDUP F- F. 0.0000E-32768 OK
0.0E F. 0.0000E-32768 OK
FVARIABLE ZERO OK
0.0E ZERO F! OK
ZERO F? 0.0000E-32768 OK
ZERO F@ F. 0.0000E-32768 OK


Right now this is causing me severe grief when any floating point value is sent to the host computer ... host expects fixed width data.

I also have to work with negative numbers, so converting to a double-word and using <# ... #> doesn't help me.

Do you have a fix for the ServoPod as well?????

BTW, I had to reflash my ServoPod the other day when I flipped a value and an equals sign ... and blew up the kernel. My application still ran, but I couldn't scrub, couldn't list words ... nothing. The kernel image disk only states "ISOMAX" with no rev, the kernel file is: v061-7.s, 114 KB, 8/27/2003 10:07 AM.

HELP!

Rick

[nmitech]

Please check your email. Thnx!

[ngkdc]

Wow ... that was fast!

Here's my results:


PI FDUP F- F. 0.0000 OK
0.0E F. 0.0000 OK
FVARIABLE ZERO OK
0.0E ZERO F! OK
ZERO F? 0.0000 OK
ZERO F@ F. 0.0000 OK


Thanks for the VERY fast turn-around.


Regards,


Rick