Thanks for the information Kai and sorry for the VERY late reply. Been busy getting a design for an AT89LP4052 Single Board Computer done. Should have PCB's in hand today. This new little chip kicks butt! Have some projects in mind for the tssop package.

On the 8253, I decided to limit the clock speed to under 16 MHz and to use the X2 mode to get to the higher performance speeds. At 16 MHz and under, you can use the correct loading caps for the crystal (30pF for 20pF CL Crystal, 26pF for 18pF Crystal)and still have over 2 Vpp on the clock. This is acceptable to me for now. I hope they fix it later, but my guess is that with the new LP products on the horizon, they won't be putting a lot of effort into fixing it. You are absolutely correct on the chart in the data sheet. Thanks for pointing me to that. It clearly shows how the higher speed crystals won't work. Too bad they couldn't actually state that. The AT89S8253 seems to work well other than this one issue for me. No problems with startup or running.

BTW Erik, you DO have to configure ports as inputs or output on the new LP devices ;-)

AT89 ISP Software - Just for those that need ISP software for the new LP devices, I have the Atmel AT89 ISP software running 100% reliable with ALL devices now. I was having some problems, and finally traked it down to the SCK (Clock) signal line coming out of the parallel port. It was extremely rounded. I had the LPT1 set as ECP in BIOS, and this turned out to be the major problem. After switching to EPP the signal was nice and square. The rounded signal was screwing up the timing. It now works reliable with my notebook and a 25' LPT cable. If you have an AT89 ISP programmer, it works well as is with all of the new devices, but had a problem with the 8252. After reducing the value of the 100 Ohm resistors to 25 Ohms (except for the reset line) that go to the processor (not the LPT port resistors) it works great for the 8252 as well. I also layed out a PCB and added a switch for the enable lines on the 74VHC244 so that I could release the SPI pins in run mode. If not, the SPI pins are locked at whatever state the programmer is in. This didn't work for me as I use several SPI periferals. I have a schematic if anyone needs it. Other than that I left it stock. I left in the LM393 as ATMEL may use this feature (VCC Voltage sence 1=5V 0=3V) in the future. Just FYI...

Trev