How can input voltage fluctuations impact the output of a linear power supply?

[Hatsuwr]

Some background - I was charging some 18650s with a cheap linear power supply. Its displayed voltage isn't the most accurate, so I set the voltage limit with my multimeter. The output was steady to 1 mV. During the charging there was a brief power outage, no more than a few seconds. Looks like some nearby houses might still be without power though. When checking on the power supply, I noticed it was slightly above the limit I set.

Checking the output voltage with my multimeter and the line voltage coming in to my house showed fluctuations on both that seem to be correlated.

Hooked up my oscilloscope to the power supply to see this. Sorry for the picture of the screen.



My basic understanding of how these regulate voltage is that there is a voltage reference of some sort that is modulated by a potentiometer, and the output feeds into an op amp that drives a transistor to provide the resistance needed to achieve the desired output voltage. I can't figure out what in this chain would be sensitive to some minor fluctuations in the incoming line voltage.

[aliarifat794]

Use your oscilloscope to monitor the AC input voltage to see the extent of the fluctuations. If the fluctuations are significant, you may need  an uninterruptible power supply (UPS) to stabilize the input voltage.

[mikerj]

Regulators have two important performance characteristics, line regulation and load regulation.  Line regulation is how much the output voltage changes when the input voltage changes, load regulation is how much the output voltage changes with changes in output current.  Many things could cause poor line regulation, an obvious one would be a mediocre voltage reference such as a basic zener being driven by a resistor from the unregulated voltage rail.

[tooki]

Some background - I was charging some 18650s with a cheap linear power supply.

Don’t do this. It’s unsafe, and it’s also easy to fry your power supply.

As for the rest: please provide more info on the kind of power supply (model, etc) it is and how you have it set up.

[Hatsuwr]

Regulators have two important performance characteristics, line regulation and load regulation.  Line regulation is how much the output voltage changes when the input voltage changes, load regulation is how much the output voltage changes with changes in output current.  Many things could cause poor line regulation, an obvious one would be a mediocre voltage reference such as a basic zener being driven by a resistor from the unregulated voltage rail.

My initial thoughts were toward the voltage reference as well. Been working on trying to trace things out, but this thing is confusing me... There is a quad op amp with two of its outputs connected to the transistor driving the BJTs that drop the output voltage. I'm guessing those two outputs are one for the voltage limit and one for the current limit, but still trying to figure it out. Excuse the terrible tracing, made it before thinking about sharing it:

Don’t do this. It’s unsafe, and it’s also easy to fry your power supply.
As for the rest: please provide more info on the kind of power supply (model, etc) it is and how you have it set up.

I'm only doing 500 mA and 4.1 V, setup with each of those as limits and terminating at 50 mA. What are the safety and equipment concerns?
The power supply is a Lavolta BPS-305, which is a rebranded Longwei PS-303D-M that is sold under various other names as far as I can tell.

Use your oscilloscope to monitor the AC input voltage to see the extent of the fluctuations. If the fluctuations are significant, you may need  an uninterruptible power supply (UPS) to stabilize the input voltage.

Ok, I have a couple screenshots here. The upper one is for about 20 minutes and captures at least one very noisy event that happens occasionally. I haven't been able to determine a pattern to these, or if they are at all consistent with one another. The lower is <2 minutes and doesn't have one of these events.

RMS variation seems to be around 1.6 V. I'm not sure if this is the right way to calculate it, but 1.6 V * 4.1 V / 123 V = 53 mV, which is just a bit higher than the output voltage variation I observed when I had the supply set to 4.1 V.

[Hatsuwr]

Alright, the power outages nearby have been resolved and line voltage seems more normal. I got a plot of voltages going into and coming out of the BJTs... and there doesn't seem to be much relation between the fluctuations. At this point, I don't think the issue is related to the incoming voltage, at least not primarily. I haven't exactly been monitoring this sort of thing closely, but given the timing, I wonder if a surge related to the outage damaged the power supply. One of my Shelly devices died right around the same time, which adds some weight to the theory.

Here's the plot. Blue is voltage into the BJTs, orange is coming out (after passing through the 180 mOhm shunt resistors).

First one is with no load on the output, second is with about 250 mA load. All are 20 mV/division.

[MathWizard]

For a higher line voltage in, you will have a higher voltage after transformers/rectifiers, so then any type of resistor divider off of that, or with common zeners, they will all pass more current, and have higher voltage drops. And then any transistors that have higher Vce from the higher supply voltage, that will effect their effective beta. And all the slightly higher base voltage/currents, move levels around even more. All that goes on inside op-amps and stuff too.

Some effects will cancel out, and some designs must work better at that than others too, same with temperature variations.

[Hatsuwr]

Well, I think I've mostly ruled varying input voltage out as a (primary) cause, just coincidence it seems.

However, I do think the power outage (and likely associated surge) damaged the power supply, now I'm just trying to figure out exactly where. The output was stable earlier that day.

I've traced things out enough to see that the control circuitry is powered by +15V and -15V coming from a L7815CV and L7915CV regulator. The output from these is consistent to within a mV it seems, but somewhere between them and the voltage to the base of the BJTs, a lot of fluctuations are introduced. There are periods of relative stability in the output (shown below), but they are usually less than a minute long.

I don't really know much about the failure modes of different components, especially which would fail partially and cause inconsistent behavior. I'm leaning toward just checking all of the capacitors first.

[tooki]

Don’t do this. It’s unsafe, and it’s also easy to fry your power supply.
As for the rest: please provide more info on the kind of power supply (model, etc) it is and how you have it set up.

I'm only doing 500 mA and 4.1 V, setup with each of those as limits and terminating at 50 mA. What are the safety and equipment concerns?
The power supply is a Lavolta BPS-305, which is a rebranded Longwei PS-303D-M that is sold under various other names as far as I can tell.

A LiIon battery whose charge current is 500mA is not an insignificant battery. If anything goes wrong, you could end up with a fire.

Your power supply has no charge management. It has no cutoff, and it may not be precise enough anyway. And you have no thermal protection, which is something normally done with LiIon batteries.

Many people have fried their power supplies by turning off the power supply with the battery still connected, backfeeding into the PSU and frying its reverse-protection diode. (LiIon batteries can provide enormous short-circuit currents.) So at minimum, you need to add a diode in series to prevent backfeeding.

But as I said, I wouldn't do it. There's a reason proper LiIon chargers exist. Use one.

[MathWizard]

Ok so it's a Lithium battery, yeah don't they also need a certain charging profile for voltage and or current ? Yeah and be careful especially indoors around flammable stuff.

I've burnt out a few 78xx type regulators over the years.  Most of them probably overheated. A lot of times I will add a reverse diode across them, if there's any chance I could power them up backwards.

When things like op-amp's and BJT get damaged, their voltages can jump around a lot. If it was my PSU, I would remove the electrolytics (if I was going to check them), and make a map of the circuit while they are out. And make sure all the resistors and diodes are ok. Then do some calculations/simulations, and see what the BJT's and chips are doing.