Fround error

scotta

Can someone confirm this:

Fround(-22.8) = 21
Fround( 22.8) = 23

It should be :

Fround(-22.8) = -23
Fround( 22.8) = 23

Scott

cgracey

Scott,

Was this in the FloatMath.spin object, or some other?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

scotta

It is in "FloatMath.spin", and I don't think I've fiddled with it.

Have not checked the "Float32FullDynamic" routines, but I have
a 600 watt motor going 80 inches a second with these routines. They
are good.


Scott

cgracey

Okay, I tested this out and it seems to work fine. Either your FloatMath.spin has been modified or something's going on elsewhere that is not apparent yet. Attached is·an archive of a demo which contains the real FloatMath.spin. The demo runs on the Propeller board and outputs your numbers in both float and round form to·a TV. It produces -23 and 23, as expected.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

scotta

Chip,

I just wasted some of your time..... It was another function I have
layered in.

Scott

cgracey

No problem, Scott. I just wish I had a simple answer for your other posted question.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

scotta

There are half a dozen people on the list that can do this
in 5 minutes, it will probably take me another few days
to come up with a workable solution. But there is a good
market for embedded controllers in my area of the country.

Thanks again !

Hows your new chip going ?
Would 12 bit addressing be too much to ask ?

Scott

cgracey

scotta said...
There are half a dozen people on the list that can do this
in 5 minutes, it will probably take me another few days
to come up with a workable solution. But there is a good
market for embedded controllers in my area of the country.

Thanks again !

Hows your new chip going ?
Would 12 bit addressing be too much to ask ?

Scott

You will benefit immensely from getting this figured out. There's no substitute for really understanding something. It will make your product much better.
The new chip is coming along great. Just working on the analog/digital I/O pads now.
Yes, 12 bits is not going to happen. Something better is in the works, though: eight-long transfers per hub access, enabling large-model code to run at half the clock speed with a full 16 cogs. Bill Henning, are you out there? This blows·out the video bottleneck, too.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

Harley

Thank you Scotta for asking, and to Chip for his replies.

There probably are many of us wondering about what Prop gen#2 will provide.
Sounds like very good news is in store for us. Keep the hints 'acoming.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Harley Shanko
h.a.s. designn

scotta

So true.

I'll send you a video of your chip cutting glass soon. The assembler
PID loop is fantastic (140 us for PID position loop, 90us for PI in the
velocity loop).

Analog/Digital pads ?

We could use my C# objects to replace quite a few NI projects.

AnalogIO myanalog = propeller.NewAnalog();
AnalogInput myinput = myanalog.NewChannel(ANALOG.IN,3);

double temp=myinput.Read();
double avg=myinput.Average(100);

Scott

scotta

Harley,

I was born to bother people, ask my girl.

Litefire

hey scott, where'd you get the PID code? i'd LOVE to have a look at it if you don't mind...

~~Brian