12v Step regulator cuts out


I am attempting to utilise 2 U3V50F12 to power a small computer and it appears either the regulators are dropping out, or the computer is not accepting the power?

I have configured both regulators in parallel to share the load of the computer as a single unit would have been to close to the power requirements of the computer. Hopefully this is acceptable and I understand that they may not share equally, but did not think this would cause me huge issues.

Total computer requirements are about 12Watts - or 1 amp at 12 volts.
The aim was to run this from a 3.7 volt lithium battery - so my expectations on load was 12watts/3.7volts/0.7 = 4.6amps (assume worst case efficiency of 70%)
The battery is rated 3000 MAh and has a 2C discharge rating, this should allow 6amps continuous, higher for momentary.

This all appears to be within the specs of the power supply, definitely within 2 units.

With this setup the computer will not start. So to diagnose I connected the regulators to a external power supply. When I set this to 6 volts the computer would finally start up and draw about 2 amps. As expected this is 12watts of power, so with losses the computer is drawing slightly less.

As I turn the voltage down on the power supply, eventually the computer will entirely shut-down at 5v and will not start back up again. at this time it is drawing about 2.5 amps. Also I watch the output power on the 12v side and it stays steady at 12v with not apparent drop in voltage as the primary voltage is reduced. After the computer shuts down the regulators continue to show 12v, does this mean it was not the regulators shutting down? I also read that as the regulator maximum input current is reached they would reduce the output voltage rather then shut-down.

The concern I have is that the computer is not shutting down because of lack of power but for some other reason. Is it possible that at lower supply voltage’s the output of the regulators produce noise that the computer does not like? If this was the case what could I do about it?

Thanks for your help,


Hello, Matthew.

I am not sure how well paralleling the U3V50F12 boost regulator would work. In general, we do not recommend connecting regulators in parallel since it is difficult to ensure their voltages match and that the load is distributed evenly across them. It is much better to use a single regulator that can provide enough current. For your application, it sounds like a single U3V50F12 boost regulator should work.

When you were testing your setup with a power supply, how were you measuring the voltage and current of the regulator? If you were using a multimeter, it might not have been fast enough to see the regulator turning off momentarily when it was near its current limit. I suspect that the momentary shutoff might be responsible for turning your computer off. If you have access to an oscilloscope, you might try measuring the voltage and current to see if the regulator is turning off. Did you try powering your computer with just a single regulator and your battery pack? Have you consider using a larger battery pack (e.g., 2s or 3s LiPo)?

- Jeremy

Thanks for the reply,

it turns out the issue is with the max current that the battery can draw. I located a larger 3.6 cell - its a 6000 mAh at 2C so can do about 12 amps, the system will work of this battery.

I have discovered that the power requirements appear to be a lot higher then I expected. My thoughts are that it must be momentary load causing the issue? I have found that even with this battery it can still shut-down depending on what is happening on the computer. (If there is high processing it will not cope and shut-down).

the two power regulators appear to be working ok in parallel at the moment, I don’t think this is causing an issue right now, but maybe under higher loads it might be more of a problem. Can I ask the following question.
1.If they were un-balanced in terms of their output, It would mean that a single regulator would take the load.
2.This would continue until this regulators output voltage drops, as it drops the second regulator would start to contribute
3.The first regulator would stay at max load until the second regulator finally also starts to drop in power

Would this be how they might work if they were not balanced? Or is there other bad things that might happen?


As Jeremy suggested, it is a bad idea to connect voltage regulators in parallel. They are never matched exactly and since they are nonidentical, active devices (changing their characteristics as the load changes) it is nearly impossible to predict what will happen when they are connected in parallel.

so - what does that mean?

Does this mean I am wrong in my prediction of what will happen? If so why am I wrong.

Saying something is a “bad idea” without any detailed explaination just makes want to ask more questions - why?


It’s great for you to want to ask more questions; we just don’t have the answers. Regarding your speculation of how it might go, you might be right. You have presented basically the best-case scenario, so let’s consider something closer to a worst-case scenario: the lower-voltage regulator might actively try to lower the voltage of the higher-voltage regulator, making the higher-voltage unit have to work harder than if the other regulator weren’t even there. If the regulator does not gradually reduce the output voltage but instead tries harder and harder to achieve its set point, it will keep trying until it breaks or some protection feature shuts it off, at which point the whole load is transferred to the second regulator.

Since your setup is apparently working at least to some extent, it might not be that bad. But, there are many variables, like input voltage, load, ambient temperature, and so on, and for a proper design (as opposed to a bad idea), you should have confidence that your system will work across all combinations of those variables. The regulator chip designers and we as the regulator module manufacturers do many tests to have confidence the regulators will work, and if you do a similar amount of testing (e.g. sweeping across voltages and currents, monitoring output details and component temperatures), you could probably have confidence that your setup, at least for the two particular units you tested with, will keep working. Doing that is generally going to be much more expensive than just getting a higher-power module made for a higher load. Plus, it’s also quite possible that in doing many tests, you could end up finding an operating point that destroys your regulators and proves they don’t work that way–in which case you’d be out your testing time and still have to go with something else.

It just occurred to me that if you really want to have the two regulators work together, for the boost scenario, you might be best off cascading them, with one regulator boosting from your battery to some intermediate voltage, and then the second one from that intermediate one to the final voltage. It might be difficult to balance them right since the second regulator would always have a predictable input voltage, while the first one would have to deal with the fluctuations from the battery. And your your efficiency would not be great: two cascaded 80% units would give you 64% efficiency overall. But, that might still be better than what you get with them in parallel, and at least you wouldn’t be shorting two low-impedance outputs together.

- Jan

On further thought, that cascading suggestion was pretty dumb: if you just need a certain amount of current at that input voltage for the desired output to be attainable, the first regulator is always going to be a bottleneck. The multi-stage approach might be applicable if you have two very different regulators and needed to do a more substantial boost, like 3V to 100V.

- Jan

Thanks for the reply, understanding the issue in more detail is what I needed, I can keep an eye out for that scenario now,

Thanks for the more detailed response, so far it is working well and not having any adverse impact,


Another item I forgot to mention is that I placed diode on the output of the relays, essentially making is so that if one had higher voltage it could not feed back into the other. I needed this the computer can take anything between 11v and 19v and with this I can also connect a external 19v supply to the system and prevent it from feeding back through the regulators.

Would this prevent the types of issues you have referred to?

The diodes should help for that particular concern. As yet another example of how complicated this can be, though, diode drops typically decrease with temperature, so if your regulators are not balanced and one side has most of the current coming from it, that diode will heat up more and make the imbalance even worse.

- Jan

Yes, the voltage drop is not ideal, although the computer will run from 11v it is better at 12v. With standard diodes I am getting about .7v drop. I have 2 in parallel on each regulator to spread load.

With the imbalance, would this mean that we would be back to the best scenario where one regulator may take most of the load until the voltage drops and then the other would then start to take the load?

Thanks for your feedback

All I said was that your speculation of how things could go was plausible, not that it was actually going to be the case, so be careful about this “back to the best scenario” thinking: you were not necessarily ever there in real life.

Also, it looks like you keep assuming that the voltage will gradually go down as the load goes up and that the primary regulator would keep operating as the second one kicked in. It’s quite possible that the only thing making the output voltage go down would be failure of the regulator or some kind of failure protection turning the regulator off. In the better of those two scenarios, the two regulators would just take turns running until they overheated, which does not sound like a very good design.

By the way, if you can afford the additional voltage drop, you can get this behavior you keep expecting to happen (voltage drops as current goes up) by putting small resistors in line with your regulator outputs. If your wires between regulators happen to be long (which would make me nervous about other potential problems), the resistance of the wires might already be giving you that effect.

- Jan