For a new version of a board we use the 3,3 V capable AT90LS8535 device. However, if we want to program the device with the Kanda ISP programmer, there are errors. Some bits are erased when they aren't supposed to be erased. E.g. 0x60 is programmed as 0x40. Most of the bits are correctly programmed, some of them are wrong.We've built our own programmer,
> programmed our own code around it and it works great for 5V devices. However, we have the same problem as with the Kanda programmer on the 3,3 V device. Most bytes are correct, some of
> them are wrong.There's no particular order of the wrong bytes. They seem to be ad random. We've respected the timing specs (diiferent from 5V to 3,3V).Our programmer checks the returned MISO code and compares it with the sended MOSI code. The AVR seems to return the correct code (otherwise our programmer gives a fail-message), but when we reading the device (both with the Kanda and our own programmer!!! ) we see the same result when we compare with the original
Are there known issues with 3,3V devices?
We ran the Kanda programmer on Win 98 (on a Pentium 60 based PC). It gave many errors. Than we tried the programmer on a Win 95 PC (pentium 166 based PC). A few errors occured.
Our own programmer runs on a win 95 pc (a 486 dx/2). The AVR returns correctly the previous MOSI code. So the AVR gets the good code, with a right timing (otherwise it shouldn't return the code).
But some programmed bytes in flash are wrong.
We didn't had this problem ever before with the 5V device, it occurs now with the 3,3 V.
Does anybody has experience with this problem?