PIC to PIC noise problem

I have some instrumentation (40 units in the field) using a 16F628 to control a linear stepper motor and supply logic-level control to an automated rotary valve. The '628 communicates with an '877 master controller which accepts a "ready" high from the '628 and issues a "go" high to the '628. All units have been performing flawlessly for three years under extremely harsh field conditions. UNTIL.....

In order to make room for new pneumatics, it was necessary to replace the venerable, but large, linear power supply with a switcher. (I lost that arguement...let's design a smaller linear). Now, one out of three units misbehaves due to what we have isolated as noise on the "ready" line giving a false "high" to the '877. The "ready" line is between porta.1 of the '628 and portd.3 of the '877.

To anticipate some obvious questions:
- The ONLY electronic change was the power supply.
- ports are initialized and TRIS'd properly.
- there are no passive steering components on the line but board is
thru-hole so there is room to "glue" things on if need be. However,
getting to the board is not an easy task.
- all units have been operating properly for 3 years with linear PS.
- '877 PBP with MCSP and MC Loader
- '628 has to removed for programming.

It would be difficult to post all the code here but all observed failures require a high on the "ready" line. The '877 sits in a loop waiting for that signal, often for several seconds so noise has a good chance of triggering a response. Any suggestions on how to deal with this will be greatly appreciated.

I think it's time to retire (again).

Steve Kerrin

Steve,

I design power supplies and instrumentation systems. The switching power supply noise couples to your circuit through conducted and radiated energy. The 1st choice is to eliminate the noise at the source. If you have an oscilloscope you measure the amount of noise at the power supply output terminals. Use a 60 to 100 mhz scope and set the scope input sensitivity to
50 or 100mV and AC coupling, 5 uS sweep would be a good place to start.
You can also use the probe as a near field probe by connecting the scope probe ground to the probe tip. The 3" loop will show your radiated sources.

You probably will find the biggest improvement using very low esr capacitors, tantalum capacitors work well, connected to the power supply output terminals. 1kv ceramic disc capacitors connected from the output terminals to chassis gnd often helps. You need a scope to measure your improvements.

It is tough to cover everything but I hope this helps get you started.

Mark