switching multiple supply voltages on same low-side darlington array IC

Question

Say I have an MCU switching a couple SSRs using the ubiquitous ULN2803 at 5V on the "common" pin, but I'd like to switch a 12V mechanical relay with the same chip as well. Checking the equivalent circuit for the IC, it is basically a low-side switch so I can switch whatever voltage within specs right?

The "common" pin only act as a return path for the flywheel diodes integrated within each driver "channel". The 12V relay will have its own flywheel diode hooked to the 12V supply.

The only pitfall I see is that maybe the integral flywheel diode would switch on before the external diode and thus dump the back-EMF energy from the 12V relay coil to the 5V rail... But then is it any different than the back EMF from a 5V relay anyways? Most 5V relays operate on higher current than their 12V equivalents, back EMF might even be higher at 5V, if I memory serves back EMF in a coil is dI/dt not dV/dt....

But if it is a problem I can use a schottky diode for the relay and it will turn on first due to its lower forward voltage. Am I talking non-sense here?

Any insight appreciated.

Answer 1

it is basically a low-side switch so I can switch whatever voltage within specs right?

Yes, as far as the switches go. No, as far as the free-wheeling diodes go.

The only pitfall I see is that maybe the integral flywheel diode would switch on before the external diode and thus dump the back-EMF energy from the 12V relay coil to the 5V rail

If you connect the "COM" pin to the 5V rail, then whenever the transistor for the 12V channel is off, the voltage across the switch will rise toward 12 V, but once it reaches, say 5.7V, current will flow through that channel's freewheel diode to the 5V rail. If you are going to have a channel that would regularly have 12V across it's switch in the absence of the freewheel diode, then you cannot connect COM to the 5V rail, unless you can tolerate a constant 6.3V across it's load, even when it is supposed to be "off". Whether that is acceptable in any given case will obviously depend on the details of the case.

I might mention that rapid "springing" is important for maintaining the integrity of mechanical relay contacts. Even if a 12V relay "releases" at 6.3 V (it may not), it may release significantly slower than it is designed to release. This may result contact failure.

An alternative, however, is to connect the "COM" pin to the 12V rail if your other circuits can tolerate 12V of inductive kick.

This schematic may help to clarify the problem.

^{simulate this circuit – Schematic created using CircuitLab}

[When Q2 is off in this circuit, current will flow through L2 and D1 to the 5V rail]