Shop OBEX P1 Docs P2 Docs Learn Events
SX28 too hot to handle — Parallax Forums

SX28 too hot to handle

LoopyBytelooseLoopyByteloose Posts: 12,537
edited 2012-07-09 07:42 in General Discussion
I was running a series of test on the SX28 board to determine exactly how to use SLEEP with it. And suddenly the SX-KEY was not recognized.

The voltage regulator and the SX28 were extremely hot. This has happened before and I am wondering.

I let all cool down and everything seems back to normal. But I am wondering that if programming and reprogramming the chip causes excessive heat buildup or if I should be looking elsewhere.

I think I reprogrammed about 25 times in rapid succession and I was sometimes using the SX-KEY for a oscillator source, other times not.

Comments

  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-07-08 09:22
    If you run an SX at its maximum speed (above 50 MHz), it will get hot. In fact, the SX is the only micro I've ever used that I've had to heatsink.

    -Phil
  • LeonLeon Posts: 7,620
    edited 2012-07-08 09:45
    I've had an SX28 go into CMOS latch-up and get extremely hot. The only other device I've used that did that was the Inmos transputer - the package lid wasn't grounded and touching it whilst it was running sometimes induced latch-up.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-07-08 10:28
    Well, I am certainly NOT clocking it at 50Mhz and I understand that heat issue.

    It is being clocked at 20Mhz to emulate a BS2. I have a 3 byte count down delay toggling an LED and I have the WDT active. I am explo Lring behavior of CLR !WDT and SLEEP in various combinations and locations.

    It did it again and it may just be a CMOS latch up. But after it cools, the behavior seems to return to normal. That might explain the SX-KEY not being seen on its COM port.

    I suppose I should consider this particular SX28 damaged and use another one.

    It seems not one thinks it is caused by a rapid succession of loading program revisions. Is

    that right?

    The LM2940 regulator gets seriously hot as well. This is a Ubicom logoed SX28.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-07-08 10:46
    Loopy Byteloose wrote:
    It seems not one thinks it is caused by a rapid succession of loading program revisions.
    I can't imagine what the mechanism for that would be, and I've never had it be a precursor to problems. But then, I've never had an SX overheat -- except for running it at 75 and 100 MHz without a heatsink.

    Inadequate bypassing could cause latchup ...

    -Phil
  • LeonLeon Posts: 7,620
    edited 2012-07-08 10:58
    It can be caused by static (that was the case with the transputer) or a glitch on the supply. Earlier CMOS processes like that used for the Scenix chips were prone to it, it's very rare with modern devices.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-07-09 07:42
    I guess I will just mark the chip as damaged, swap it out and see what happens. It is really impossible to be sure of the cause, but since the SX-Key is okay and the board seems okay - I will try to forget about this one chip.

    Sadly, the saga continues as my Toshiba NB250 has suddenly had its video go haywire. And that has all the software in it. I suspect it is due to accumulated carbon dust from living in a steel mill town.

    Spent the afternoon on a first exploratory dis-assembly of the netbook for the possibility of cleaning with a spray can of tuner cleaner. And even re-assembled and got it running again. But now it is back in failure mode.

    The chip MAY HAVE been clocked at 50Mhz in the past, but it was so long ago that I cannot be sure.
Sign In or Register to comment.