Why are these boards using two power supplies?

[elepo]

Disclaimer: I intend to make a product and sell it. I'm making this disclaimer because some people do not like to give free advice to people who intend to commercialize their product.

The product I'm intending to make is a door interlocking system. Basically electronic locks are installed on doors and the controller circuit will make sure that one of the door is always closed, i.e. both doors can never be open at the same time. Pretty simple to do with a microcontroller.
I already know how to make this but I'm looking at some similar commercial products to get hints and ideas since this is the first time I'm selling something I've built.

So I opened this commercial product and to my confusion it had two power supplies instead of one. I assumed perhaps they are at different voltages, but no they were at the same voltage (12v). I assumed perhaps they intended to isolate some parts of the circuit from other parts but then I realized that the grounds of the power supplies are actually connected. One of them is 12V-2A and the other one 12V-3A. I don't understand why they didn't use a 12V-5A instead which would be cheaper.
Here is the inside of the product:



Interestingly enough I have opened another similar product and it also uses two power supplies. This one a 12V-1.5A and a 5V-3A. I can't understand why they have chosen to do this. The 5v is being used for the relays and the MCU. And the 12v is being used for the electronic locks of the doors. Why aren't they using 12v relays and a 7805 for the MCU? That would save them a ton of money.
Picture of the product:



So I'm wondering if there's some reason behind this I don't know about?

[tooki]

My guess for the paralleled 12V supplies is redundancy: the circuit probably needs less than 2A, so that either supply alone can run it. If one fails, the other takes over. (As for one being 3A, maybe they ran out of 2A ones. Or they deliberately chose different ones to eliminate the chance of them being from the same batch, with the risk of a bad batch eliminated.)

As for 5V: simplicity and reliability? I don’t know how many relays they’re running, but assuming they need anywhere near the 3A they have available now, that’d be a ton of energy to dissipate in a linear regulator. (12V-5V)*3A=21W, waaaay above what a 7805 can dissipate. Even if it’s just, say, four relays at 50mA, that’d be (12V-5V)*200mA=1.4W, still more than a regular 7805.

As for why they don’t use 12V relays: maybe in their design, that’s easier to interface with the control logic.

[elepo]

My guess for the paralleled 12V supplies is redundancy: the circuit probably needs less than 2A, so that either supply alone can run it. If one fails, the other takes over. (As for one being 3A, maybe they ran out of 2A ones. Or they deliberately chose different ones to eliminate the chance of them being from the same batch, with the risk of a bad batch eliminated.)

It sounds overkill but that's the only thing I can think of too.

As for 5V: simplicity and reliability? I don’t know how many relays they’re running, but assuming they need anywhere near the 3A they have available now, that’d be a ton of energy to dissipate in a linear regulator. (12V-5V)*3A=21W, waaaay above what a 7805 can dissipate. Even if it’s just, say, four relays at 50mA, that’d be (12V-5V)*200mA=1.4W, still more than a regular 7805.

As for why they don’t use 12V relays: maybe in their design, that’s easier to interface with the control logic.

Oh I should have included the image of the board too. Here's the board:


The MCU is an ATmega16 and the IC is a ULN2003.
They could use 12V relays which use slightly less current too which is always good, and since they're using the ULN2003 they wouldn't have any problem driving them.
Can't be for redundancy either because if the 12V supply dies the doors won't work regardless of working of the MCU and the relays.
I don't want to simply assume that they've just been stupid but maybe that's the case?

A couple other disturbing things here is that I don't see a 0.1uF ceramic cap for the MCU, and no filtering of inputs. They will be connected through long wires (maybe 5+ meters) to the door sensors and push buttons.

And thanks for your answer.

[james_s]

Hilarious that they bothered to sand the numbers off those DIP ICs on the relay board. I bet it wouldn't take more than 5 minutes to reverse engineer enough to figure out what they all are.

[Benta]

Protection against noise, spikes, lightning, whatever.
One PSU supplies controller and associated relay coils.
The other supplies solenoids etc. on the doors, where there's a lot of wiring and cables that can pick up all sorts of nastiness.

Both PSU outputs are galvanically isolated from each other, ensuring that no disturbances are transmitted into the controller.

Good design IMO, keeping the controller and actuator parts separate makes sense.

[elepo]

Both PSU outputs are galvanically isolated from each other, ensuring that no disturbances are transmitted into the controller.

Are they isolated if their grounds are connected?

[Benta]

Both PSU outputs are galvanically isolated from each other, ensuring that no disturbances are transmitted into the controller.

Are they isolated if their grounds are connected?

From the pictures, they are grounded on the input side, which places the enclosures at zero potential. This does not stop the outputs from floating.

[elepo]

From the pictures, they are grounded on the input side, which places the enclosures at zero potential. This does not stop the outputs from floating.

I'm sorry I'm not following.
You said the reason for separate power supplies might be to provide isolation.
But the ground (negative/common) of the power supplies are connected together through the circuit. Doesn't that mean they are no longer isolated?

[ajb]

I think the confusion is from people missing your statement that the DC negatives of the power supplies are tied together in the circuit and are assuming that the DC sides of the two power supplies may be galvanically isolated.  Even if the DC grounds are connected, if the DC positives are NOT connected, then there is some 'isolation' in the sense that a fault on one will not directly affect the other.  For example if the load on one power supply is shorted out, that supply can shut down while the other one, and its load, continue to operate.  This is a form of fault isolation, which is different from complete electrical isolation.  Even if the other devices powered from the failed supply don't work maybe there's value in the rest of the system being able to still operate, like to assert a fault signal or something. 

Another possibility reason for using two power supplies of the same voltage is that the product may be available with different IO voltages for connecting to peripheral devices (like your door locks) designed for different voltages.  If they use the same 12V/2A supply in every single unit, but then need 12V/3A in some and 24V/1.5A in others it may be more economical to do it this way.  Or maybe some versions only need the one 12V/2A supply. 

As for the other system that has a 12V and a 5V power supply, well clearly they needed both voltages, and all of the above factors apply as well.  If they need anything close to 3A@5V, that's a bit much for a 7805.  Sure they could use a switching regulator, but maybe that hardware evolved from a legacy design, or they just didn't want to go to the trouble of designing one.  The electronics that are visible in your photo are super low cost construction and low complexity (being single sided phenolic boards they have to be!) so fitting a switching DC-DC converter on there would probably mean stepping up the complexity and cost to a larger degree than if the system were already higher complexity with a 2+ layer PCB etc. 

Suggesting that "it would be cheaper if..." on someone else's product is always going out on a limb, because there are a lot of factors that go into manufacturing costs that are only really visible to the manufacturer themselves.  Supporting different product variants as mentioned above, or sharing parts between products can have a big effect on product cost as compared to designing one product in isolation.  Even if it would have been cheaper to use a single 12V/5A supply, maybe there was a supply chain issue and they chose to pay more for two power supplies than to have to delay selling/shipping product.  Or maybe the two power supplies shown here actually WERE cheaper--smaller power supplies are probably made in much larger quantities so maybe better price per watt, or maybe their distributer had a good deal.  It's impossible to really know without a lot more insight into who made these when and why and what else they make.

[Benta]

From the pictures, they are grounded on the input side, which places the enclosures at zero potential. This does not stop the outputs from floating.

I'm sorry I'm not following.
You said the reason for separate power supplies might be to provide isolation.
But the ground (negative/common) of the power supplies are connected together through the circuit. Doesn't that mean they are no longer isolated?

"Ground" or "Earth" is not the same as negative. Your imprecise phrasing leads to confusion.

[elepo]

@ajb Thank you very much that was very insightful.

"Ground" or "Earth" is not the same as negative. Your imprecise phrasing leads to confusion.

As far as I know "ground" usually refers to the circuit common point (GND or power supply negative), and "earth" refers to the earth wire that is connected to metal enclosures to protect from shocks. Which is why I used the word "ground" here. But sorry if it was misleading.

[Benta]

"Ground" can be anything you want it to be, there's no fixed definition. If you call the positive output of a 12 V power supply "ground", then the negative output is -12 V.

"Earth" or "Protective Earth" is a different matter.

[james_s]

@ajb Thank you very much that was very insightful.

"Ground" or "Earth" is not the same as negative. Your imprecise phrasing leads to confusion.

As far as I know "ground" usually refers to the circuit common point (GND or power supply negative), and "earth" refers to the earth wire that is connected to metal enclosures to protect from shocks. Which is why I used the word "ground" here. But sorry if it was misleading.

That depends. In the USA the earth connection is referred to as "ground" or "equipment grounding conductor". The term "earth" makes more sense in my opinion but it's not really used here.

[David Hess]

Protection against noise, spikes, lightning, whatever.
One PSU supplies controller and associated relay coils.
The other supplies solenoids etc. on the doors, where there's a lot of wiring and cables that can pick up all sorts of nastiness.

Both PSU outputs are galvanically isolated from each other, ensuring that no disturbances are transmitted into the controller.

Good design IMO, keeping the controller and actuator parts separate makes sense.

That is exactly it and I have done the same thing when needed.  With separate power supplies, a fault on the solenoid side cannot cause the control circuits to misbehave.  This is important in safety and security systems.