Pin isolation/protection ?

Hello everyone,

I’m working on a project with a Motorola HC11-E9 micro controller. I will be connected to 4 of the AD channels and sampling a slow system (temperature, humidity, etc in a room).

I’ve read through a lot of the reference manual for this chip but most of what I could take away was that for my A/D pins I should just have a series resistor to limit the current into the pin under the worst conditions that could be found on the board. Which would probably be a short to 12V. I believe they said a 1K resistor should work to limit the current (<25mA) without causing leakage from going to high with the input resistance. They also showed a bypass cap for noise filtration which would reduce the available sampling time but with my system sampling once every couple seconds would be fast enough.

So to rather than continue to make a short story long....

Could somebody suggest a good method or components to protect/isolate my pins that won’t adversely affect the accuracy of my readings?

I would also like to protect my output pins and since I’m just using on/off control all I need is a pulse and I should be able to use an optical isolator if needed.

Could somebody suggest an isolating IC perhaps with 8 channels or so for this portion of my system.. if such devices exist.

Thanks in advance!

It's sometimes necessary to clamp input voltages to the rails using a couple of fast diodes in series joined with a resistor at the port to limit damaging currents.
Quite often, a low value decoupling cap is used to filter noise at the input as well, but the added parasitic capacitance added by the diodes is quite often sufficient on it's own. It really depends on what sort of signal levels you are trying to sample.
An octal driver chip suggestion might be the ULN2803A. Assuming that you can live with open collector outputs. I suggest this device because it can handle reasonable current, gives you a way to level translate if needed and has inbuilt freewheeling diode protection so can be used to drive small inductive loads directly. If you really need optical isolation, then just use a couple of general purpose quad optocouplers, although the current they can handle is somewhat lower than the darlington driver I mentioned.

rgds

I picked up a few ULN2803A, I was planning on using this to drive the coils of my mechanical relays. I was told by my instructor, who will be evaluating my designs, that the darlington array won't provide any isolation and that I should consider using optical isolators..... that could just be an opinon

The statement below is from the reference manual and is what I am going on with the use of 1k series resistors alone. Diodes are also mentioned:

"Since there are no P-channel devices directly connected to the A/D input
or reference pins, voltages above VDD do not pose a latchup threat. If an
A/D input rises above the threshold of the protection device, an input
protection device avalanches, and current into this device should be
limited. Because of an inherent diode to VSS, A/D inputs must not go
below VSS or the input can be permanently damaged. A series resistor
of 1 kΩ will prevent damage, but avoid a series resistor of more than
10 kΩ because input leakage acting through this impedance will
degrade A/D accuracy. External clamping diodes on A/D inputs should
be avoided because the leakage through these devices is greater than
the input pin leakage current and could significantly degrade accuracy if
significant resistance exists in series with the analog source."


I'm not really sure why I'm being told I need isolation like optical isolation. I'm just working with varying DC signals in a non industrial enviroment.

As a simplest design would it be possible to just directly hook up the outputs of my micro to the ULN2803A and the relay coils directly to the collector? I'll probably throw a schottky diode on the coil just because I have many of them and I don't think it could hurt.

I will likely be using some op-amps to modify my voltage signal to match the 0-5V range to get the best resolution out of my samples. If I'm only applying 5V to my op-amp and I use a 1k series resistor would that be a safe design to apply to my A/D ports?

Thanks again!

I would think you are on the right track....I see no real need for optical isolation, buffering of the micro for sure, but not isolation. He may be suggesting it because of possible noise or pulses being generated on the supply rails with the inductive loads. Adding a couple of quad optical isolators won't hurt as such, but you then require isolated supplies etc. Personally I think it's overkill.

The freewheeling diodes in the ULN2803A will safely clamp the back emf, no need for additional clamping diodes on your relay coils.
I would say though, that your supply ground and signal ground need particular care, make sure that you have clean rails or the micro will get upset by seeing emf pulses going below or pulling the rail voltage...

rgds