Broken prop?

Philldapill

I have no idea what I've done, but I think I may have killed my prop. It's been outside in an enclosure for over 4 months as a remote control gate opener and worked flawlessly. Today, I decided to add a few more remote codes to it so that a friend can use his remote to get in and out. I took it out of the enclosure, brought it inside to reprogram, and all seemed well as I plugged it in and programed it. I noticed it didn't work right and long story short, I can program it, but none of the pins will do anything. I tried another prop to make sure something was wrong, and the other works fine under the same program - just a simple flashing light. This one seems to have all the pins broken or something. Still programs fine. Any ideas???

Mike Green

There have been a few cases of Props that have somehow been stressed with static or too much voltage, nobody knows for sure, but the PLL seems to be the first structure to fail on the chip. Symptoms are that the chip will identify and you can download programs to it, but they won't execute. The chip will operate if the clock is RCSLOW or RCFAST, but none of the PLL clock modes will work.

Paul Baker

Sounds like you fried your PLL, to test this load a program with the clock mode set to RCFAST, if it runs, the PLL is gone.

We have determined the leading cause for this·are ground loops and inductive kickback.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Philldapill

Ouch. Thank you both. I did as you said and sure enough the LED is blinking. I never thought I'd be disappointed to have the propeller WORKING.

Next question - I guess with the RCFAST setting, the chip will be too slow to be able to process TV remote codes? All I'm using this one for is to continously watch an IR detector for TV remote pulses, and if it sees them, closes a relay for a second, and repeats.

Philldapill

Well, maybe some better news... I noticed you said none of the PLL modes will work, which they don't BUT, if I set the clockmode just to xtal1 and input freq to 5mhz, it works! Does this still mean it's using the internal RC timer, or is it simply using the 5mhz input, no PLL interaction at all?

Mike Green

RCFAST should be pretty good for your application. It provides a clock somewhere between 8MHz and 20MHz, typically 12MHz. It's a little temperature sensitive and supply voltage sensitive, but stable for short periods of time. It should be fine for your application.

If the clock mode is xtal1, the chip is running off the external crystal oscillator.

Philldapill

Hmmm, I'm glad I won't have to get out of my car and walk 10 feet to open the gate now, BUT... It's looking like my TV routines aren't working. Does the prop need a fast oscillator to generate TV or something? I'm trying just a barebones version to print text to the tv but no luck.

Mike Green

Neither the internal clock nor a slow external clock will work for TV or VGA video generation.

Serial I/O won't work well either, mostly because you don't know the real clock frequency
that's needed to set the Baud. That's why the Propeller bootloader uses a ratiometric
scheme to make it independent of the long-term clock timing.

Philldapill

So, since the chip is messed up, if I ordered a new surface mount chip for the protoboard, I should be able to just cut it out, and solder in a new one?(carefully)

sevs

Why not just get a faster external clock?

I would have thought that it would be cheaper than getting a new prop.· You just need to use a bit of maths to get the speed you want.

Philldapill

I need a speed of about 40mhz. The maximum clock input is well below this. Unless there is another way?

Ale

With an external oscillator you can feed it with 40 MHz.

Philldapill

Theoretically, then can't you feed it 80mhz, too? I was thinking there was some sort of input limit that you could feed, even if the total clock after PLL is still lower than 128mhz or something...

Leon

Philldapill said...
So, since the chip is messed up, if I ordered a new surface mount chip for the protoboard, I should be able to just cut it out, and solder in a new one?(carefully)

I use ChipQuik for replacing SMD chips, it works very well and doesn't damage the tracks.

Leon


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Rayman

I used a xacto knife to cut the leads and solder wick to clean up and it worked (except for one damaged pad that I had to replace with a wire. But, I only did this because I had a lot of stuff on the board... New proto boards are pretty cheap, so it's not really worth the risk if your board isn't stuffed.

Philldapill

Yeah, I would just buy a new board, but I do have some components and things on it that I'd rather not take off, and resolder to the new board. Instead, I did what you did, Rayman. I got a razor knife and cut the leads right at the base of the chip, and took the chip off with ease. I then got the solder tip at a low temperature, but just hot enough to melt solder, and wiped off the pin legs with the tip. I then cleaned the tip, and repeated this. The board looks great. This is my first time ever doing any solderwork with SMT and I have to say, either that solder mask is great, or I'm just good. haha. No, all the pads look very nice with just enough solder on them to stick to the chip when I get it.

Paul Baker

My preffered method of removing QFP type packages is using a pair of narrow-width flush wire cutters to snip the pins off right near the board (the chip is dead anyways) then once again flush with the chip so you can access the inner pins. Then after all the pins are removed, take out the package. Finally take a soldering iron and a dental pick and melt each pad while dragging the remainig bit of the pin off the pad. While it's not elegant, for dead chip removal it uses tools that are in nearly everyone's toolbox, and there is no chance for damaging the pads.

The creme de la creme is a hot air pencil with a QFP attachment the size of the chip your working on, it comes off in seconds and board and chip are still in good condition.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.