What are the dangers of non-isolated power supplies?

Thread Starter

yardleydobon

Joined Apr 19, 2009
19
Hi, I'd like to try out the LM675 dual voltage power supply talked about here:
http://forum.allaboutcircuits.com/showthread.php?p=84026

I have a 36 V, 3 A, AC/DC converter that I believe would be called non-isolated. It has a three prong AC plug, and two DC wires, one of which has a connection of about 1Ω to safety ground.

I've read that isolated power supplies are safer, but have never found an explanation why. Would some please explain why?

What problems might I have if I use a non-isolated supply? A noisy ground?

Also what is about the upper limit of current draw on an average electronics project using DC? I'm not sure what size fuse to use.
 

SgtWookie

Joined Jul 17, 2007
22,230
Hi, I'd like to try out the LM675 dual voltage power supply talked about here:
http://forum.allaboutcircuits.com/showthread.php?p=84026

I have a 36 V, 3 A, AC/DC converter that I believe would be called non-isolated. It has a three prong AC plug, and two DC wires, one of which has a connection of about 1Ω to safety ground.
It sounds like it's an isolated supply. We don't "like" people to use non-isolated supplies, because they can expose people to very high power levels that could kill them very quickly.

I've read that isolated power supplies are safer, but have never found an explanation why. Would some please explain why?
Transformers and switching supplies isolate mains power from the output power. The limits in these supplies keeps the power to a safe level.

What problems might I have if I use a non-isolated supply? A noisy ground?
Death. :eek:

Also what is about the upper limit of current draw on an average electronics project using DC? I'm not sure what size fuse to use.
If it's a commercially built supply, it may already have protection in the primary side. Since the secondary is rated for 3A, use a 3A fuse or less for that side. Fuses are there to not only protect humans, but also to protect equipment/circuits from being turned into smoking piles of rubble. :rolleyes:

Using simulations like SPICE prior to assembly of your circuit will help you to determine nominal current consumption; then you will have a good idea of what your current usage should be. Adding a fuse will be cheap insurance for your circuit in case anything should go horribly wrong. Mistakes are easy to make. Components can be expensive. Fuses are usually pretty cheap.
 

Thread Starter

yardleydobon

Joined Apr 19, 2009
19
Thanks for your reply. Thanks for the circuit schematic too! However, I'm now more confused. I was thinking floating ground is equivalent to isolated. Is this not right?

It's at least correct to say this particular supply does not have a floating ground, right? Don't most AC/DC converters have floating grounds?

Can you just give me a rough number for how much an average "tinkering on a breadboard" circuit may draw? Three amps is quite a lot. I want to limit the possible short circuit current down to something a bit safer with a resettable fuse. I'd like to go as low as possible. Is 250 mA enough for most circuits?
 

SgtWookie

Joined Jul 17, 2007
22,230
Thanks for your reply. Thanks for the circuit schematic too! However, I'm now more confused. I was thinking floating ground is equivalent to isolated. Is this not right?
Well, everything's relative. ;) The "simulated ground" in the LM675 supply is just an offset from actual earth ground. As long as the entire circuit being powered by the supply doesn't contact a "real" earth ground, you should be OK. However, if you try to use that particular supply design to power something that's connected to another circuit that DOES use "earth ground", you will experience problems.

It's at least correct to say this particular supply does not have a floating ground, right?
The simulated ground has a deliberate and adjustable offset from the real earth ground.

Don't most AC/DC converters have floating grounds?
Not necessarily. Usually, the chassis is connected to earth ground. Most frequently, the secondary side of the transformer has at least one connection to earth ground as well. These safety items protect the human user in the event that there is a fault within the transformer.

Can you just give me a rough number for how much an average "tinkering on a breadboard" circuit may draw? Three amps is quite a lot. I want to limit the possible short circuit current down to something a bit safer with a resettable fuse. I'd like to go as low as possible. Is 250 mA enough for most circuits?
It all depends upon what you're attempting to build.

If you're working only with 4000 series CMOS logic IC's, 250mA may be way too much.

If you're working with old TTL logic (54/74 series), 250mA won't get you very far.

If you're building circuits using 555 timers, a single 555 can source/sink up to 200mA, and has 200mA "glitches" when the output changes states.

There is no one size fuse that will work for all circuits. It's subjective. Look at datasheets to determine what the currents might be. Build the circuits in simulators first, and use virtual ammeters or low-value resistors to determine the current in the simulation.
 

Thread Starter

yardleydobon

Joined Apr 19, 2009
19
I should have made myself more clear. I understand how the LM675 circuit works. I plan on using it as a rail splitter so I can get 30, 15, and GND, and then use the 15 V rail as common. Then I'll have 15 and -15 relatively speaking.:) I just want to use it for breadboarding op amp circuits. I do understand your point about not thinking the 15 V rail is an honest ground.

My main concern is with the AC to DC converter I have. See the return leg on the DC side has a connection to safety ground. I don't know if this is a good thing or a bad thing for me. This means this supply is not floating right? Will this cause a problem with LM675 circuit?
 

SgtWookie

Joined Jul 17, 2007
22,230
I see. You should be OK with that, as long as you don't connect your circuit under test to another device/circuit that's connected to a real earth ground. If you will be using a line-powered oscilloscope, remember that the o-scope ground may be connected to earth ground.
 

Thread Starter

yardleydobon

Joined Apr 19, 2009
19
Thanks for all your help! Since I've got your attention, I'd like to ask one more question.

I'd like to use an L272 instead of an LM675. The datasheet for the L272 says it can output 1 A. Will the L272 be able to sink and source this much current, or just source it?
 

SgtWookie

Joined Jul 17, 2007
22,230
You must be reading the ST Microelectronics datasheet for the L272.

If you look at Fairchild's datasheet, it specifies 0.7A.

While you might get away with using the L272 for low currents, you'd be better off with the LM675; the package is much better able to dissipate heat (particularly when you use a heatsink) than the DIP of the L272.
 
Top