Dual Power supply switch with 5-10 sec delay

[NuclearOnion]

Hey all. I'm new to the forum and hoping some of you with much more DC electronic experience than me would help me out.

• I'm working with a 12v dc system. 12a max.
• I have 2 independent power supplies and 1 end point I want to supply power to.
• Both power supplies are 12v DC [8-12a max]
  
• The end point is 12v DC [12a max]

1. I want power supply A to be the primary source of power to my end point.
2. However, when power supply B is introduced, I want all power disconnected completely from my end point for 5 seconds before being supplied with power from power supply B.
3. If power supply B is removed, I want to disconnect power completely from my end point for 5 seconds before reconnecting back to power supply A.

I have 2 SPDT relays and 2 NE555 timer relays. How would I set up the system I'm looking for above with these? Or is there a better / simpler way or product(s) to do this?

 

[ramussons]

A specific question - is the requirement A or B?
A) When Power at the end point is removed, there needs to be a 5 second break before power is supplied again - source does'nt matter.
B) When Source B is activated, Disconnect Source A and Connect B after 5 seconds and Vice Versa

 

[NuclearOnion]

Hey @ramussons Thanks for taking a look at this.

The requirement is B.

Power supply A is the default.
When power supply B is available, the system should disconnect power from the end point completely for 5 seconds then supply power from B...
...Until B is removed, then wait 5 seconds (no power to end point during this time) before applying the default power supply A back to the end point.

 

[CircutScoper]

Hey @ramussons Thanks for taking a look at this.

The requirement is B.

Power supply A is the default.
When power supply B is available, the system should disconnect power from the end point completely for 5 seconds then supply power from B...
...Until B is removed, then wait 5 seconds (no power to end point during this time) before applying the default power supply A back to the end point.

I think the circuit sketched below will do what you need.

1. When supply B is off, both relays are de-energized, connecting the load to supply A.
2. When B comes on, RLY 1 is energized, causing it to switch contacts. This disconnects A from the load and causes a brief (a millisecond or two) glitch that triggers the 555, causing it to energize RLY2, which isolates the load.
3. RLY2 remains energized, keeping the load unpowered, for the 5 second timeconstant of R1xC1 until the 555 times out.
4. RLY2 is de-energized reconnecting the load to now be powered by B.
5. When B turns back off, the resulting glitch triggers the 555 again, energizing RLY2 for another 5 seconds of load isolation while RLY1 switches back to A.

Please note that I've assumed the coil resistance of the relays is at least 60 Ohms. Otherwise the 555 will be overloaded by RLY2.

 

[crutschow]

What are the relay coils voltage and current rating?

2. When B comes on, RLY 1 is energized, causing it to switch contacts. This disconnects A from the load and causes a brief (a millisecond or two) glitch that triggers the 555, causing it to discharge C1 and energize RLY2, which isolates the load.

That will connect B to the load immediately after A is disconnected, for the time it takes for RLY 2 to energize (likely several tens of milliseconds).
Will that be a problem for this application?

And how do you know that the load voltage will drop sufficiently to trigger the 555 during Rly 1's short crossover time?
If the load has any significant capacitive filtering, it may not.

Also RLY 2 needs spike protection across its coil to avoid zapping the 555.

 

[CircutScoper]

...
Also RLY 2 needs spike protection across its coil to avoid zapping the 555.

If the inductive kick of the relay tries to drag the output pin negative, the 555's internal output darlington will turn on and safely absorb it. No external protection is necessary.

Meanwhile, good point about wondering if the output will drop low enough (i.e., < 4V) during the switch-over to trigger the 555. RLY1 may need to be a DPDT to provide a separate pair of contacts for the trigger function with a pull-down resistor to ground at pin 2.

 

[crutschow]

Okay, here's my take on using one 555 timer to give an overlap off delay when the supplies switch over (LTspice simulation below):

Supply A is on continuously and supply B is shown turning on and off (green trace).
The 555 is level triggered by the output from relay U2 (signal "T") which generates a 5s on delay and a 5s off delay (yellow trace), as determined by the values of R2 and C1 (blue trace).
I set the B voltage to 11v so you can tell which voltage is coming from the output of relay U3 (red trace).

 

[CircutScoper]

Okay, here's my take on using one 555 timer to give an overlap off delay when the supplies switch over (LTspice simulation below):

Supply A is on continuously and supply B is shown turning on and off (green trace).
The 555 is level triggered by the output from relay U2 (signal "T") which generates a 5s on delay and a 5s off delay (yellow trace), as determined by the values of R2 and C1 (blue trace).
I set the B voltage to 11v so you can tell which voltage is coming from the output of relay U3 (red trace).

View attachment 55046

My compliments to the chef, chief!

Meanwhile, I gotta' say it -- what a weird way to spend a Saturday.

 

[crutschow]

what a weird way to spend a Saturday

When you are retired, Saturday is generally no different than any other.

 

[CircutScoper]

When you are retired, Saturday is generally no different than any other.

Speaking of "other days," I imagine a simpler time when grandpaw would have just replaced RLY2's coil with a bi-metallic strip and 5-10 second thermal timeconstant heater connected to B, and thus dispensed with all this silly silicon nonsense.

 

[NuclearOnion]

Wow. You guys quite impressively went over my head. I was once MECP certified for dc car electronics back in the Circuit City Roadshop days. I have a baseline understanding and can interpret the diagram with a google search or two to jog my memory. I'll take a look through your responses and see if I can make some sense of it. I didn't specify that I am using the NE555 circuit HERE
Also I assume by your answers that there's no pre-existing product you know of that does this full function on it's own. Essentially what you suggest only on Amazon. A guy can hope can't he?

If not, I'll get to work and see if I can make what you laid out. Thanks again gentlemen and I think it's a perfectly decent way to spend a Saturday.

 

[crutschow]

Also I assume by your answers that there's no pre-existing product you know of that does this full function on it's own.

Yes.
Your requirements are unique enough, that I doubt there's an off-the-shelf device to do that.
Edit: You might be able to do something similar to my circuit if you can find a module that will both delay on and delay off at the same time (along with an added relay).
The problem is, the description for many delay modules on Amazon aren't clear about whether they can do that.

I cleaned up the schematic a little, perhaps making it easier to follow (below).

 

[CircutScoper]

Yes.
Your requirements are unique enough, that I doubt there's an off-the-shelf device to do that.

I cleaned up the schematic a little, perhaps making it easier to follow (below).

View attachment 55051

Question, Maestro. Why did you return the C1 timing cap to +12 instead of ground? In a non-critical timing app' like this, it won't make any practical difference, but more generally I think it would tend to maximize the negative effect of jitter, noise, and ripple on the supply upon timing consistency.

Was your intent to force the circuit to power up with supply A selected? In which case, C2 should probably be returned to +12 too.

 

[crutschow]

Was your intent to force the circuit to power up with supply A selected? In which case, C2 should probably be returned to +12 too

Yes and No.
Yes, the C1 connection is to have the circuit come up with A selected (555 OUT low) when the A power comes up.
With C1 to ground, the OUT will be momentarily high upon power application.

The circuit should, of course, have a decoupling capacitor from power to ground near the 555 to minimize the affects of any noise on the power bus, which I neglected to show.

But since CV is connected internally to the TRIG and THRES trigger point resistor string for noise suppression, you want C2 connected to ground to help keep the trigger points low, since you want the THRES voltage to stay above the THRES trigger point during power up, insuring the 555 OUT stays low.

 

[CircutScoper]

But since CV is connected internally to the TRIG and THRES trigger point resistor string for noise suppression, you want C2 connected to ground to help keep the trigger points low, since you want the THRES voltage to stay above the THRES trigger point during power up, insuring the 555 OUT stays low.

Actually, that's not entirely true. For best noise rejection, ideally you want the trigger points to track the timing capacitor. Since the timing capacitor is usually referenced to ground, the CV bypass capacitor usually should be too. But in this case, with the timing cap' tied to V+, doing the same with C2 is better than tying it to ground.

I don't know if your simulator is realistic enough to verify that, but it's true regardless. You might want to try it.

 

[crutschow]

Actually, that's not entirely true. For best noise rejection, ideally you want the trigger points to track the timing capacitor.

True.
But for this non-critical timing application I'm not that concerned about power bus noise.
Typical power bus noise would likely only affect the timing by small fraction of a second.

I don't know if your simulator is realistic enough to verify that, but it's true regardless

It likely will if I want to go the trouble of simulating noise on the power bus, but I'm not that interested.

 

[CircutScoper]

Actually, that's not entirely true. For best noise rejection, ideally you want the trigger points to track the timing capacitor. Since the timing capacitor is usually referenced to ground, the CV bypass capacitor usually should be too. But in this case, with the timing cap' tied to V+, doing the same with C2 is better than tying it to ground.

I don't know if your simulator is realistic enough to verify that, but it's true regardless. You might want to try it.

C2 tied to V+ would make Vthreshold immune to supply noise. By contrast, C2 tied to ground is worse than no C2 at all.

 

[crutschow]

C2 tied to V+ would make Vthreshold immune to supply noise. By contrast, C2 tied to ground is worse than no C2 at all.

I agree.
Below is the modified circuit:

 

[CircutScoper]

I agree.
Below is the modified circuit:

View attachment 55056

Thank you. In a different thread you said that an unusual 555 application circuit shown in a datasheet was wrong. That particular case was correct, but we've just demonstrated that the reflexive connection to ground of every CV capacitor, that's shown in every 555 datasheet, sometimes IS wrong.

Conclusion: This is a pretty good forum! You learn useful tricks here that you won't see in the datasheets.

 

[crutschow]

Yes, this and a couple other electronic forums are a good place to learn the basics along with some tricks.

For example, when I first started here I knew little about the 555, but there were many requests for timing circuits where the 555 was a good choice, so I learned more about it.
Once you understand that it is just a FF latch with level-sensitive set (TRIG) and reset (THRS) inputs, along with a pin (DIS) to discharge the timing capacitor when the reset input is triggered, it's then fairly easy to understand and design various types of timing circuits, such as the delay-on / delay-off circuit I used here.