Unique Electronics Switching Design Problem (Please Help)

[hshah8970]

Hello everybody!

I'm working on a project and am facing a problem I simply cannot solve. I'm unable to design the electronic circuit that accomplishes the following task:

I want to couple 2 DC inputs in such a way that when Input 2 reaches certain voltage level (from zero), both inputs become connected in series. Before this threshold voltage level is reached by Input 2, we simply get Input 1 at the output. Check out the attached image to see what I mean.

For voltage values of DC Input 2 higher than the threshold voltage, the series combination is maintained.

I would highly appreciate it if somebody helped me out with this! Thanks! And an explanation to go along the circuit would be nice too. Again, thanks a lot to anybody who responds!

 

[OlPhart]

Making the rash assumption that "voltages in series" is to add them...

I'm thinking a 2 input summing amp:
input a = Input 1, input b = output of a 2 input analog switch.
switch input a = ground, input b = Input 2, switch control = output of a comparator.
comparator input a = Input 2, input b = threshold voltage as set by potentiometer.

When Input 2 > threshold, comparator switches & summing amp output = Input 1 + Input 2.
Else: Input 2 <= threshold, summing amp output = Input 1 + 0V (ground)

It should work at small-medium signal levels.
Any kind of real current required will require amplifiers.
High voltages will require scaling then amplifying the result.
If this is a homework assignment, I've already said too much... (bad dog, no biscuit) <<<)))

 

[hshah8970]

Thanks for the reply!

No no, not a homework assignment. But I AM an early student, and this is the only thing that's keeping me from completing my project.

The voltage levels will be around 12-17V. And yes, the voltage in series was meant to imply addition.

And "OlPhart" I'd IMMENSELY appreciate a schematic/diagram. I'm a 'noob' at electronics and conceptualizing your description is proving to be a challenge.

 

[picbits]

As above though - how much current is required ?

If you are switching say <10ma then its probably going to require a totally different circuit than if you're switching a few amps.

Other things to consider are :

How fast do you need the switch over ?
Do you need hystersis on the rising / falling initial switch
Do you have available another power supply to supply the overall circuit or does it have to be powered from the two inputs ?
Are the grounds common or are both inputs isolated from each other ?

 

[hshah8970]

Hello "picbits" !

Half a second or less would be about right for my project.
No hysteresis required.
No other power supply available.
The inputs are isolated from one another.

Also, for the current rating: I'm not entirely sure yet. I'll be running a DC motor on it (a medium-sized one) with a rating of 12V. I've yet to go out and research and buy the motor.

 

[picbits]

Ok - here's how I'd do it.

I'd connect the - of DC2 to the + of DC1 and have a DPDT relay with the common connected to the motor. One side of the DPDT goes to the - of DC2 and the other side of DPDT goes to + of DC2. The other side of the motor goes to DC1 -

When the relay is not active you get DC1 and when active you get DC1+DC2

Triggering could be done by a comparator (powered via DC1) sensing the voltage on DC2 (probably through a divider as it will be above DC1)

 

[hshah8970]

Thanks for your help picbits!

But do you mean a SPDT relay? I haven't used relays before but a quick web search showed that SPDT relays have one common and match your solution. Please confirm https://upload.wikimedia.org/wikipedia/commons/6/67/Relay_symbols.svg

Also, I'd be grateful for an elaboration on the triggering. A schematic would help me greatly.

 

[OlPhart]

Since it's not classwork...

I'd look at the Exact voltages involved to see if available op amps, analog switches & comparators (oh my) might squeak by.

I'm not a fan of using a relay as the switch element. That's probably a bias on my part. I'd Much rather keep it all in lower volt level signal stuff, then amplify it back up to real world requirements. Analog switches (4066..etc) are easy to use, have predictable characteristics, and don't exhibit noise. They just have range limits... oh well.

(I'm breaking my arm here...) No chops for the concept?? (teehee) <<<)))

P.S. since I pay the bills by designing stuff, my limit is discussing & describing.
Also picbits isn't wrong, I just disagree; it's 2 viewpoints, neither bad, but one will prove out better. I'm probably as curious as he is...

 

[crutschow]

Here's a schematic of two relays to connect either one power supply (V1 with K1 energized) to the load or two power supplies (V1 + V2 with both K1 and K2 energized) in series.

 

[MrAl]

Hi,

Just a quick note...

When using relays to switch power supplies or voltage sources it's a good idea if not mandatory to use BREAK-BEFORE-MAKE type relay contacts. This is especially true if there is a the possibility of two sources becoming connected in parallel or a short across one source even for a short time period (present case included).

Even better is to design in some definite lag time between relay closures. In other words, open all relays, wait 100 milliseconds or more, then close the required relays. This prevents utter disaster in some cases.

You would not believe the number of expensive power supplies i've seen blow up simply because the relay logic didnt take into account the true relay operation (thousands of dollars worth of equipment).

 

[crutschow]

..........
When using relays to switch power supplies or voltage sources it's a good idea if not mandatory to use BREAK-BEFORE-MAKE type relay contacts.
.................

I believe all common double-throw relays are break-before-make. It would require a special design to do otherwise.

 

[MrAl]

I believe all common double-throw relays are break-before-make. It would require a special design to do otherwise.

Even so it is worth pointing out that even that isnt always enough. I definite time delay is the best policy especially when high power is involved. As i said before, i've seen a lot of expensive stuff blow up because of problems with the crossover relay switching.

 

[hshah8970]

@OlPhart: Thanks for all your help! I'd prefer to use a non-relay-involving switch as well, as I have never dealt with relays before. The voltage levels, just to let everybody know, can be considered the following: DC Input 1 will operate around 12V while the threshold voltage for DC Input 2 will be, say, 5V. Once the threshold is reached and a series connection is achieved, the voltage value of DC Input 2 will only increase - not border on 5V. Also, I appreciate you providing me with your expertise even though you're not getting paid for it.

@crutschow: Thank you for the schematic! I've viewed it and it will achieve exactly what I want. But don't you think it would be simpler if we went with the SPDT relay method suggested by picbits (see above)? My question now is how to employ comparators and how to power them through the given power supplies. Note that I will have no other power supplies available. I'd appreciate if somebody would use a schematic to elaborate on the last line posted by picbits "Triggering could be done by a comparator (powered via DC1) sensing the voltage on DC2 (probably through a divider as it will be above DC1)"

@MrAl: Thank you for your help. I will look into the type of relays you mentioned and keep the problem you described in mind while completing my circuit. Thanks for potentially saving my input supplies from disaster.

 

[picbits]

Something I forgot to ask but is DC1 a fairly constant voltage ? If not, then what is its range ?

I suggest you do some homework on comparators - its fairly simple stuff and if you do a little research yourself then you'll know how they work.

If you were to wire the circuit as per my suggestion then using a 75:25 voltage divider you could get away with a zener, a few resistors, an opamp, a transistor, diode and a relay.

 

[crutschow]

................
@crutschow: Thank you for the schematic! I've viewed it and it will achieve exactly what I want. But don't you think it would be simpler if we went with the SPDT relay method suggested by picbits (see above)? ..............

Yes, you could used two SPDT relays.

 

[mbarazeen]

only one relay is enough as suggested by picbits