Sinking feeling

Joe McClure

I have a bad feeling that my probstick is broken.
The IDE recognizes it fine. It programs just fine. It·loads, verifies etc. both RAM and EEPROM, but it acts like it has nothing programmed into it, it·does nothing.
I was programming away tonight and then it suddenly at next upload, it·stopped doing what I had programmed it to do.
I even gave it the simplest command to do, which·was to toggle pin0.
I put the logic probe on pin0 and nothing. None of the iopins seems active now.
I have power (9v bat, like it always has had) like·I said it programs without error but it's like it didn't take.
It looks bad and I am really down in the dumps over it. It is only a couple of weeks old (just a baby) and I have been so careful with it.
Any ideas? Seems hopeless.

Sad, sad, sad.

RobotWorkshop

Have you put a voltmeter on the power to make sure you haven't drained that 9v battery? I've seen several posts across all the forums where the problem ended up being a weak battery. The battery is fine when it's done but during all the work on the bench while your working with the Propeller you may want to use a wall wart power supply or some sort of steady power. Rip one out of and old PC or some other electronic gadget that people seem to throw out these days. A 5v supply should not be too hard to find.

Good luck,

Robert

Joe McClure

Thanks Robert,
Yeah, first thing I checked. I tried fresh bats, and both 5v and 9v power supplies.
Looks grim I'm afraid.

cgracey

If you can download to it, it's working. At least, it's working off the internal 12MHz oscillator.

Can you verify that your crystal is plugged in okay? Can you also try running your test program off the internal 12MHz oscillator:

CON··· _CLKMODE = RCFAST

It sounds like the external oscillator has quit running for some reason.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.

Dennis Ferron

It may not be as bad as it appears. AFAIK, when the Prop is downloading a program from the PC, the Propeller must run a program out of it's own ROM to do this download. So you see if you can successfully download code, then the Propeller must be at least halfway alive in order to do even that.

My guess is that maybe your Propeller is being held in reset mode. Sometimes the drivers for USB serial ports don't play nice and send the wrong signals, causing the Propeller to reset at odd times. You should be able to check the voltage on the reset line.

Another thing to check is maybe the PLL (clock frequency) directive. The first Propeller I had in my Prop Stick had a bad PLL block on the chip; it would run in RCFAST or RCSLOW but any kind of PLL clock setting with the crystal would make it go completely dead. You tried a simple program to toggle the LED, but have you tried removing the clock setting directives from that LED toggle program?

Come to think of it, the Propeller I had with the broken PLL exhibited EXACTLY the same behavior as you describe. I mailed it back to Parallax and they kindly replaced it; I put the new one in my Propeller and it has been working fine for a while now.

T Chap

Can you post the test code you are using?

Joe McClure

Thanks guys,
I did what Chip and Dennis suggested and I DO get activity with RCFAST and RCSLOW, yay! at least with the toggling test.
The other code that uses the FullDuplexSerial code doesn't work though. Dang. Does that have to with not using the crystal now? What could cause this to happen? It is a propstickUSB BTW (if it matters at all).
The crystal looks like it's in there snuggly and looks ok. It looks like the only detachable component.
Do I need a new crystal?

Mike Green

FullDuplexSerial really requires a crystal to function. The built-in RC oscillator is just not accurate enough to do standard serial I/O. The bootloader gets around this by using the serial bitstream as its own timing reference.

You could also try your test program with _clkmode = xtal1 (without the pll16x term). This would use the external crystal and oscillator, but without the phase locked loop circuitry. If it works (although only as fast as 5MHz), that tells us that the crystal and oscillator are working.

Joe McClure

Thanks Mike,
Not working with _clkmode = xtal1. And upon closer inspection (my eyes are tired) but the rcslow doesn't work now, maybe it didn't earlier.
So at this point my propstick only works (limps) with _clkmode = rcfast. But without the serial io I'm sunk.
If i just need a new xtal that wouldn't be so bad. But at $79 for at new propstick, it will be awhile before i can save up for a new one. crud.

Mike Green

If RCFAST works, RCSLOW ought to work (but very very much more slowly - 1:1000 ratio). In any event, Parallax will replace a defective unit. E-mail or call their support department.

Joe McClure

Mike, I see about rcslow being really slow.
Thanks everyone for all your help on this.
I really appreciate it. You guys are the best.

Joe

rokicki

Crystals have been known to break, by the way, so that's certainly something to try. On the other hand,
they normally break from being dropped rather than just sitting on a test bench.

I hope you find it's something simple like just reseating the crystal!

(Be glad it's only the Prop, though; my wife's Sony VAIO just gave up the ghost. While the Parallax toys
aren't super cheap, so far they are a *bargain* in the fun department.)

LoopyByteloose

Umm... Are you providing 3.3volts power to the Propellor. [noparse][[/noparse]You mentioned 5 volts, not 3.3 volts].

It is just possible that you have put it on a slow destruct at 5 volts and the external oscillator circuit was the first to die.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"If you want more fiber, eat the package.· Not enough?· Eat the manual."········

···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

Joe McClure

kramer,
it's a propstick (built-in voltage regulator etc..)

rokicki,
thanks and hey I been meaning to thank you for doing all that work on the sd code.
really nice, i will be using it (when i get my propstick fixed, that is)

Joe

squidx

I'm having a similar problem in the sense of partial failure.

I'm using the PropKit, and suddenly the TV out has given up the ghost. One minute I'm getting graphics and text, the next, nothing. However, I can send sound to the speaker, no problem. But I gave it a quick try on an LCD panel and it looks like the VGA isn't working either. However, I never observed that working, so I might be doing something wrong there.

I have two kits in my lab, so I have switched over to the other one, and it's fine, I can run TV_Text_Demo or Graphics_Demo just fine.

I figure I'll have to send it back, but if anyone has any thoughts on reviving it on my own, that'd be my preference.

Thanks,

Alex

rokicki

Sounds like we need a simple "pintest" program that just iterates through the pins, sets them high then low ten times, then sets them back to an input and
we can scope them (they should all show squiggles with a duty cycle of a couple percent). Something like

pub start : i
   repeat
      repeat i from 0 to 39
         dira[noparse][[/noparse] i] := 1
         repeat 10
            outa[noparse][[/noparse] i] := 1
            outa[noparse][[/noparse] i] := 0
         dira[noparse][[/noparse] i] := 0

This assumes you have a scope. If you don't have a scope, something
more complicated is called for to be able to test the pins with just a
multimeter or LED.

cgracey

squidx said...
I'm having a similar problem in the sense of partial failure.

I figure I'll have to send it back, but if anyone has any thoughts on reviving it on my own, that'd be my preference.

Could you and Joe please email me so that I can send you out replacements·today? I would like to get back the failed units, too, so that we can try to determine what went wrong. We can work out the details through email. Thanks.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Chip Gracey
Parallax, Inc.