*Need some help from EEs or serious electrical/electronic hobbyists.
A friend needs 36VDC @ ~1.3 amps for a project he’s working on.
I have a couple of 19VDC 2.6 amp laptop chargers.
Can I wire the outputs of the laptop chargers in series to provide the desired voltage, without destroying the chargers? Or is there likely to be something in the charger output circuitry that would not be at all pleased if I did that?
No way of knowing with out trying, but you may find that you need something to “slow down” the load. Either a capacitor or PI filter on each one at the end further from the switcher might help. I’d do it with two separate filters, rather than one that spans both.
Shouldn’t be a problem, assuming you keep the polarity straight.
I agree with Don. If the polarity is correct, it should work fine.
Adding 2 of the same power supplies together in series will double the voltage.
Adding them in parallel will double the amperage.
I’m far from an expert on these things, but isn’t one of the legs referenced to ground or neutral or something? I didn’t think most garden variety switchmode power supplies were isolated. So if they’re not isolated, the common on supply A is going to be at a similar potential to common on B, with some sort of electrical path between them. So connecting + on A to - on B will be similar to a short. Or that’s how it works in my head for non-isolated power supplies. I’ll agree that you can treat an isolated power supply like a huge battery for voltage generating purposes.
I suppose you could check by checking voltage between - and ground and - and neutral on both supplies. If you get some sort of steady, not-floating value, then you’ll want to be careful when you hook the two supplies together.
Ether, it can’t hurt to ohm out the output wires to the input wires. Anything other than several kOhms or more means ‘come back and let us know so we can figure this out for you’.
I assume these are switching supplies, and these run at several dozens of kHz, thus using a tiny high-frequency transformer, which provides isolation. I suppose one could build one with a shunting configuration, but there are common mode issues (noise) that often nixes that idea.
OK here are the results:
The 2-prong wall plug (hot & neutral) of Charger A has no connectivity between either output and either input.
The 3-prong wall plug (hot, neutral, ground) of Charger B has no connectivity between either output and the hot or neutral inputs, but it does have connectivity between either output and the ground input.
With both chargers plugged in, there is no measurable voltage between either output of Charger A and either output of Charger B.
Do you mean you didn’t measure output voltage of any kind from either charger, or they both output the same measured voltage?
I ask because some supplies may not engage their output until they’ve been connected to a device, through the use of a sense pin that identifies the device and charger to each other. This is the reason some laptops cannot (easily) use off the shelf third party chargers. Could you elaborate on what brand of chargers they are?
Also, I am not what your friend’s budget (if any) is, but perhaps I could recommend something like this?
http://www.digikey.com/product-detail/en/ETMA360166UD-P5P-IC/T1214-P5P-ND/2920448
Matt
Neither.
What I intended to convey was this:
Chose either output wire (plus or minus) from Charger A, and chose either output wire (plus or minus) from Charger B. Then measure the voltage between the two chosen outputs. The voltage reading is zero regardless of choice.
It’s an interesting grammar question. Had I intended to convey what you said, I’d have written this:
There is no measurable voltage between the outputs of Charger A or the outputs of Charger B.
I don’t know if that’s the correct grammar or not.
Anyway, the purpose of the test was to determine if there is a leakage path of some sort which might cause a problem when the outputs of the two chargers are wired in series.
Also, I am not what your friend’s budget (if any) is, but perhaps I could recommend something like this?
ETMA360166UD-P5P-IC CUI Inc. | Power Supplies - External/Internal (Off-Board) | DigiKey
Thanks for the link. I’m pretty sure his budget is “whatever I have laying around”.
Sounds like the outputs are floating/isolated relative to each other, so you should be fine connecting them in series to get your 38VDC that you’re looking for.
For the record, I suspect two of your 3-prong output models might have been a problem, however.
Yeah.
so you should be fine connecting them in series to get your 38VDC that you’re looking for.
Even though they’re isolated (so there’s no leakage path), I still wonder:
These chargers are designed to source current, not sink it.
They’re probably switching technology.
If one’s switched “on” while the other’s switched “off”, then the “off” one sees the “on” one trying to push current through it. That’s where I think Eric’s earlier comment may come into play. Wish I knew more about this kind of stuff.
For the record, I suspect two of your 3-prong output models might have been a problem, however.
Hmm. I bet I’ve got another 3-prong one hiding somewhere. I’ll see if I can find it and try it.
Isolated switching power supplies (which A/C supplied switching power supplies usually are) can easily be and are regularly hooked in series. The supply will never be sinking current, current will always be going out the (+) terminal and in the (-) terminal. To be safe, reverse biased diodes should be connected to each output so that if one power supply comes on more quickly, it will not apply reverse voltage to the other. See here: http://www.acopian.com/acopianPowerSupplies/entry.aspx?nsId=17
While this is true for linear power supplies, it is not true for switching power supplies unless they are specifically designed for it. Because the output voltage of a switching power supply is actively controlled, any minute differences in output voltage will result in the supplies fighting each other. It will work for supplies that are well matched, but because they will use some current fighting each other, the available current will be less than combined rating and should only be done temporarily when there are no other options.
Jason,
While this is true, a simple fix for parallel switchers is 'OR’ing diodes. This will cause the 2 supplies to droop share.
Ether,
Traditional switching supplies are isolated so running them in series is usually fine. Adding a steering diode should give you protection from reverse biasing the supply. I’d also suggest powering the downstream supply first (for the same reverse bias reason … although it should not matter).
In other words, the diodes are there to prevent Supply A from trying to push forward current through Supply B (via the load), when Supply B is not ready.
That’s essentially what I said in my earlier post, except I expressed it in terms of Supply B being in the “off” switching state, rather than not having powered up yet.
Can someone explain what aspect of the supply’s output circuitry does not like having a voltage applied across it… said voltage being of opposite polarity so that said voltage is attempting to push current through the supply in the forward direction (as would be the case we are discussing with the two supplies in series)?
Am I correct in thinking you checked for connectivity with a meter set to measure Ohms? If so, what was the resistance measured? If the resistance is lower than say 100 Ohms, it is Ground referenced.
It is possible that the three-pronged charger may be referenced to ground. If the charger contains a regular 60 Hz transformer, it may be a type that provides no input-to-output isolation. If it contains a switched mode power supply, it is much more likely to provide isolation from the input ground.
Obtain a resistor, anything from 10 kOhm to 47 kOhm, 1/4 W or higher. Connect one end of the resistor to the positive of one of the chargers and the other end to a known Ground point (third prong on outlet). Use your meter to measure the voltage of the negative output relative to Ground. Then connect one end of the resistor to the negative of one of the chargers and the other end to a known Ground point (third prong on outlet). Measure the voltage of the positive output relative to Ground. if you get about 19 V in both cases, the output is one that floats relative to ground.
If just one of them is floating, it should be Okay to connect them in series. If one of them is Ground referenced, you may have to consider the effect on the equipment that you are powering with this supply. If both are floating, then you can treat them like a battery.
Paralleling diodes only prevent one supply with lower output voltage from dragging down the other. The supply with the lower output voltage will not be able to push out much output current since it’s paralleling diode will not be forward-biased as much or may be reverse-biased in the extreme.
True parallel systems typically have an isolating impedance (resistor or inductor) to allow each supply to put out slightly different voltages. Otherwise, they would all need to be controlled from a single controller that forces the paralleled supplies to all have the same output current. The inverters I work on at my day-job uses sharing inductors for all the higher powered models (518 to 1500 kVA).
There may be more, but the first thing that comes to mind is the electrolytic capacitors. They will fail even under small reverse voltages, and if they aren’t properly vented, they can fail quite catastrophically: http://www.youtube.com/watch?v=D-RZ5RTAdSg.
In addition to what Jason said, there is a voltage feedback to the switching controller that may not like a voltage on it’s feedback input being lower than it’s ground.
Mostly right. When current is pulled from 2 supplies with outputs connected by 'OR’ing (matched) diodes, the current will only be pulled from the supply with higher voltage until that supply cannot keep up. *At that point its output voltage will start to droop. When it droops to the level of the second supply that supply will then take the rest of the load.
*
It’s not the best way to parallel supplies, but it does work when 1 supply wont source enough current.
Guys,
I would really like to get a handle on this discussion. It is OK to add power supplies in series. It is not OK to add battery chargers in series. Luckily I believe we are talking power supplies. Current flow will will be supplied by both power modules and since they are in series there is no need to add anything to smooth or share current flow. It will occur naturally. Ether, I am going to guess you did not measure low resistance from the third pin to both outputs. Likely what you measured were output protection devices like MOVs or some other device meant to keep ESD spikes to a minimum. As long as the resistance was a couple of kohms you are fine. Good practice would suggest that the three pin supply provide ground to the circuit you are powering and the two pin supply provide the higher voltage. Since the two pin supply has no continuity to primary wiring, connecting it as the higher supply should be just fine. As additional protection, a fuse in one leg of each supply might be a good idea in case something should happen with one supply, it would isolate the fault from damaging the other supply.
I’m not comfortable with those blanket statements. We really don’t know what is inside those converters.
Turning on one before the other applies a reversed polarity voltage across the slacker. That could become a problem. The classic and cheap transformer rectifier linear regulator setup has two components that possibly could have a problem. The output capacitor could be an electrolytic (as jason said) or a tantalum. A majority of linear regulators can’t handle reversed polarity. As for switchers, depending on the topology the commutation cell could become an unintended reverse path and pop like the linear.
My comment about output filtering assumed a pair of identical supplies. Stacking switchers that operate at the same frequency isn’t a good idea, as their control loops can become confused. The quick’n’dirty fix is to pi filter the accidental control connection away.
