Re: Programmer troubles with a new twist.

Cameron Rector wrote:
> First of all; I have to say my troubles are looking more and more self induced.
> 
> I talked to some people on the techtool fourm and got a tip to watch A11 with a logic probe while programming. Well intresting enough; A11 goes high right at 800h and stays high though the rest of the programming (that is what it shoud do). So this got me thinking of performing a few more tests.
> 
> 1. first I loaded a blank eprom into ram, then I viewed the ram and saw it was filled with FF (which is correct). Then I performed a verify and it passes (and it should). I next loaded my binary file (I am assuming its binary) and re-ran verify and it fails (and it should). Then I program the ram (containing the file) into the eprom and the post program verify fails (and it should not). So I see it fails at 800h (as it always does) and it shows ram = 00 and the device = A2 (or something simular, I forgot the exact number). So at this point I see from 800h on things don't match up. I next load ram again with FF just to make sure it is clear and then download the device into ram. I then view the ram at 800h and it says 800h = 00. I again run verify and again it fails at 800h saying the same thing ram = 00 and the device = A2.
> 
> Now how the heck can that happen? I just downloaded the device into ram and verified it against its own contents and it fails showing it has a different number than what it downloaded.
> 
> 2. Just to test the hardware of the programmer. I verified a new blank eprom and it passed all FF. I then loaded a constant 00 into ram and programmed the eprom and it passes. This tells me that I can take all highs at all pins and change them to all lows at all pins. This also tell me that the programmer must be doing the hardware thing correctly.
> 
> 3. I programmed another blank eprom with an abitrary constant CC and it passes. CC is loaded in all locations.
> 
> I hope you can follow all of this.
> Thank you for all the help folks!
> Cameron

Definitely a programmer problem.  As an experiment, try adding a 4.7 K (or 4K7 if you're European) resistor from the A11 line to 5 volts at the EPROM and see if the problem goes away.

 

Yes, the programmer is taking the A11 line high, BUT HOW FAST.  If the driver is bad (or missing a pullup resistor, if used) you could be seeing the contents of 0000 instead of 0800 momentarily.  To have the problems you have this would have to be a VERY borderline timing issue.

 

(I don't know if this answer will get through or not, I'm sending it from work (that's sometimes a crap-shoot).  Have a MAJOR internet crash at home---my ISP has (had?) a contract with MCI.  They got bought by Verizon.  Verizon decided to tear up the contract with almost no notice.  It will still be a while until everything returns to normal.)