High Side darlington array

deperkin · Joined: 04 Feb 2009 Posts: 83 Location: PA

i would say (but difficult to verify at this point) that I will be around 50mA. For sure I am under 100mA.

But to get an exact number requires alot of research into this since there are many I/O's and each have different functions.

I will look into this however.

My question now is: the UDN2891 shows only test condition voltages.
I have tested these with a 24V input... it seems to work as expected and they are not getting warm. I believe at this point the quickest (not best) solution is to created a total of ~50 voltage dividers here for the UDN's.
Is there a better solution for this? for easier wiring (i.e. 8-channel buck regulator).

It states for the 2981 a 15 VIN (i am assuming for now this is max rating).

deperkin · Joined: 04 Feb 2009 Posts: 83 Location: PA

Mike Walne,

no I mean 24v to 5v.

For testing I am using a 24V supply...

I will update the schematic tonight to show you exactly.

I have this:

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

If you need to convert a 0 to 24V switch signal to a PIC input the obvious choice is a resistive voltage divider, preferably with a cap on the output.

But if ground noise may be a problem a good solution is a PNP common base level shifter. Connect a small PNP transistor with the base to the logic rail, 5V or 3V whatever the PIC runs from. Run the input through a current limiting resistor (10k) to the emitter. Connect the collector to the PIC with a pull-down resistor and probably a cap to ground. All these parts are available in multiple packs (MMPQ3906).

The advantage of the common base is that the switching point is the logic rail plus 0.7V which is a safe distance from ground to avoid ground noise problems. It is also non-inverting.
_________________
The search for better is endless. Instead simply find very good and get the job done.

Mike Walne · Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK

OK

Now you're telling us you need two level shifters.

24V to 5V for inputs to the UDNs.
And 5V to 24V for MCP to UDN's.

Use common base devices for both.

Mike

temtronic · Posted: Fri Jan 18, 2013 4:22 pm

Overall it sounds like a "PLC" system with 64 inputs and 64 outputs ?

deperkin · Joined: 04 Feb 2009 Posts: 83 Location: PA

yes pretty much this type of system, but not exactly since i have slightly more outputs than inputs.

No i do not need to convert 5V to 24V... that is what the UDN's do.

I just need for the inputs to convert 24 down to at least 15... in total there are 43 inputs here.

two of the MCP chips will be all outputs on both channels... Ill try to update the schematic tonight.

temtronic · Posted: Fri Jan 18, 2013 8:26 pm

Sounds like you want an 'IDC5' type device. Heavily used in industrial systems,robust,near bullit proof! These are 'solid state modules' which are esentially optoisolators and 'signal conditioning' in an epoxy package.
The input (LED half) is floating and isolated from the PIC or PLC side (the transistor)
You might get by with just a simple 4N26 type device and a couple resistors. One to limit LED current, the other for the input to the PIC.
If your product can share a common signal ground, then a simple resistor-zener diode may work or an LM78L05 will do as well.
I've used them all and like the 'L05 solution for automotive interfacing.

options...you've got lots...

hth
jay

deperkin · Joined: 04 Feb 2009 Posts: 83 Location: PA

yes this is the type i need, but because of the amount of I/Os i thought I would do something that comes in 8-channel IC's...
I plan on having two of these test benchs for a single test once i get everything working well.

Total i really have over 300 I/Os

Mike Walne · Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK

OK. We've got there, having been round the houses.

Yes. Your inputs are latchng, because that's how you've wired them up.

What you've got from UDN output to MCP input is this.