Query on preventing reverse flow

[jinu_j]

Hi Folks

I need your help in a design problem. I have a power source that provide 3.3V, 5 V, +12V, and -12V. Each output is rated at 5A individually. Now i need to create a 4 way switch which will allow me to select the output I want. Unfortunately cannot find a 5 way rotary switch that can handle 5A. The ones i find are too big and used for main electricity board. Hence not suitable for my project.

So what i was thinking was to do the following
- Use normal 5 way rotary switch rated for 500mA.
- Connect the output of this to 4 channel 5V relay. So on turning the rotary switch the appropriate relay gets triggered
- The output of all the relays will be combined in to a single output.

My concern with this approach is the risk of reverse current flow to the non selected relay since all relay outputs are shorted. I am not sure if this is an issue before the non selected relays would be open. But just to be sure i thought of using a Schottky Diode rated a 10A at the output of each relay to ensure reverse flow does not happen. But the problem with Schottky Diode is the forward voltage (approx 0.7V) This is pretty high for the range of outputs i am operating on. For eg: for a 3.3V output the output would drop to 2.6V once is goes through the Schottky Diode.


So my queries
- Do i need Schottky Diode or will the assumption that the reverse flow will not happen because the other non selected relays are open hold
- If the reverse flow concern is true then is there another solution instead of using Schottky Diode, to ensure the output voltage does not drop ?

Thanks

[eccerr0r]

Re: Query on preventing reverse flow

Draw it to be sure, but the only issue is to make sure that the control signals do not overlap and select more than one relay at a time. Or you could design the switching tree such that it's impossible to select more than one output at a time, though this will require you to have some sort of binary input.

[R_J]

Re: Query on preventing reverse flow

.

[redwire]

Re: Query on preventing reverse flow

It's still a dangerous way to go. Imagine some switch or relay overlap on a 3.3V circuit getting -12V it's going to blow things up or even short the power supply rails.
If it takes 20msec for a relay contact to move, and the rotary switch is guaranteed a break-before-make type, it's still no good.
What's wrong with moving a wire on a banana plug?

[eccerr0r]

Re: Query on preventing reverse flow

Also, diodes won't protect against the -12 to +12 short case...

[jinu_j]

Re: Query on preventing reverse flow

Thanks folks for your responses. But does anyone have a solution for this?

One thing i can do is eliminate the -12V output.

But the other point is valid because for a fraction of a second when i switch between relays there is a possibility of 2 relays being live. The single output from the relays will go through a volt/Ammeter before going to the output banana plug. This will allow me to operate with a single volt/ammeter and output. Hence i dont want to use individual banana sockets for each output.

Any ideas on how to address this

[eccerr0r]

Re: Query on preventing reverse flow

Most standard SPDT relays are break before make, just make a selector tree (since you have 4 voltages, you'll need 3 relays, 2 if you use a DPDT for first level) and use appropriate selection circuitry, it should never short two outputs together. Technically speaking you might well just use a DPDT and a SPDT toggle switch, and depending on the position of the two switches you'll get each output.

Otherwise (or if you don't like binary selection and insist on a rotary switch) you'll need to be careful in selecting a circuit that guarantees break before make, either by selecting a switch that guarantees it or designing a circuit to do it...which probably complicates things immensely.