Main bus undervolt

Forums › Gadgets, robotics, home automation, electronics (including wearables) › Main bus undervolt

neb

11294 posts

Uber Geek

Trusted

Lifetime subscriber

#253075 24-Jul-2019 17:21

I've got a bunch of 12V equipment powered off a 12V 4A Delta power supply, i.e. a good-quality one. The chain is: Delta SMPS -> 12V inline power meter -> 4-way splitter -> equipment. I recently added a 12V UPS before the main router, on one of the outputs from the splitter, and it started alarming about a too-low input voltage as soon as I plugged it in. While the Delta outputs 12.2V, by the time it got to the UPS it was 11.3V (measured with a second inline power meter). I've spent several hours checking all the cables and connectors and they all have a resistance of fractional ohms, the voltage (with a high-impedance DMM, i.e. no-load) is at the expected level, I've swapped out every cable and connector, and there's still a drop of close to 1V by the time the UPS sees the power.

My suspicion is that the inline power meter, an Aliexpress one, is pulling down the output voltage when under load, which is why it's not measurable with a DMM. Or at least that's the only explanation I can still think of for why close to 1V is vanishing between the power supply and the device being powered.

Does anyone else have any ideas?

1 | 2

11294 posts

Uber Geek

Trusted

Lifetime subscriber

#2282986 24-Jul-2019 17:51

As further info, the inline power meters are the QP2320 from Jaycar, a.k.a. the PZEM-031, but off Aliexpress at a fraction of the price.

richms

28168 posts

Uber Geek

Trusted

Lifetime subscriber

#2283015 24-Jul-2019 18:55

Measure its shunt resistance and see what it is. It will probably be in the negative which is why those meters are pretty much useless for interconnected equipment.

Richard rich.ms

mclean

581 posts

Ultimate Geek

Subscriber

#2283309 25-Jul-2019 11:17

Can't you simply measure the voltage at each point in the circuit while it's under load using the DMM? That should tell you where the voltage drop is.

wellygary

8312 posts

Uber Geek

#2283319 25-Jul-2019 11:39

First up you need one of these... sorry couldn't resist with a title like that..

neb

11294 posts

Uber Geek

Trusted

Lifetime subscriber

#2283386 25-Jul-2019 12:28

mclean:
Can't you simply measure the voltage at each point in the circuit while it's under load using the DMM? That should tell you where the voltage drop is.

To do that I'd have to scrape insulation off wires at each point (or some equivalent like putting a tap on each connection), which is why I was using the inline power meters.

Zeon

3916 posts

Uber Geek

Trusted

#2283527 25-Jul-2019 15:12

Sorry I'm not an expert here but couldn't you use a clamp meter for current and check voltage by running cables in parallel from the terminals? This way there should be no effect on your voltage by the metering at all?

Speedtest 2019-10-14

richms

28168 posts

Uber Geek

Trusted

Lifetime subscriber

#2283581 25-Jul-2019 16:12

Zeon:

Sorry I'm not an expert here but couldn't you use a clamp meter for current and check voltage by running cables in parallel from the terminals? This way there should be no effect on your voltage by the metering at all?

DC clamp meters are hopeless at low currents because its not much more than the magnetic fields that are around the place anyway.

Richard rich.ms

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

elpenguino

3419 posts

Uber Geek

#2285424 29-Jul-2019 22:47

neb: I've got a bunch of 12V equipment powered off a 12V 4A Delta power supply, i.e. a good-quality one. The chain is: Delta SMPS -> 12V inline power meter -> 4-way splitter -> equipment. I recently added a 12V UPS before the main router, on one of the outputs from the splitter, and it started alarming about a too-low input voltage as soon as I plugged it in. While the Delta outputs 12.2V, by the time it got to the UPS it was 11.3V (measured with a second inline power meter). I've spent several hours checking all the cables and connectors and they all have a resistance of fractional ohms, the voltage (with a high-impedance DMM, i.e. no-load) is at the expected level, I've swapped out every cable and connector, and there's still a drop of close to 1V by the time the UPS sees the power. My suspicion is that the inline power meter, an Aliexpress one, is pulling down the output voltage when under load, which is why it's not measurable with a DMM. Or at least that's the only explanation I can still think of for why close to 1V is vanishing between the power supply and the device being powered. Does anyone else have any ideas?

The power meter should be able to handle ~20A so let's accept the shunt resistance is low.

It sounds like you have impedance between your supply and your load. This impedance might seem low when tested with a meter but a DMM uses micro amps so it doesn't act as dynamically on the circuit as the load.

An easy thing to try is to connect your load more directly to the supply, by passing everything. If it works, then connect the load sequentially further away , you'll isolate the problem no probs.

Remember things like UPS often have high start up current, much higher than the average rating.

Most of the posters in this thread are just like chimpanzees on MDMA, full of feelings of bonhomie, joy, and optimism. Fred99 8/4/21

DonH

230 posts

Master Geek

Lifetime subscriber

#2285639 30-Jul-2019 12:15

richms:

DC clamp meters are hopeless at low currents because its not much more than the magnetic fields that are around the place anyway.

DC clamp meters such as the Uni-T UT211B measure down to milliamps quite accurately. They have a zero function which cancels the effect of stray magnetic fields.

People hear what they see. - Doris Day

tripper1000

1617 posts

Uber Geek

#2285756 30-Jul-2019 14:29

Your power supply voltage is a bit on the low side providing very little tolerance for voltage drops. Most 12v gear runs happily up to 13.8 or 14 volts. Bumping it up to ~13v would probably help a lot.

Trouble-shooting-wise - forget measuring current - you're chasing a voltage issue after all, and forget measuring resistance - at those low currents the resistances you are chasing are so low they are hard to measure, and besides, many faults show good resistance under no load (when you have the circuit off in order to use your ohm-meter) and bad resistance under load (when you can't measure with an ohm meter), so you simply can't detect some faults probing around with an ohm meter.

Use your DMM volt meter. First, measure the voltage from one point on the negative rail to another point on the negative rail then compare to the equivalent points in the positive rail. In a perfect world all readings should of course be zero volts - but we know your circuit is presently imperfect, so you will see 0.5v to 1v at some point(s) in the circuit. If you get equal voltage drops along the neg and pos rails then your wiring is probably a bit light weight or something is impacting both rails equally. If one rail drops more voltage (ie measure more point to point) than the other, then test each segment of the rail until you find the segment with the highest voltage drop. It could be a loose wire, bad solder joint, oxidised wiring/terminals, thin wiring etc.

Edit: It is possible the UPS load (to charge the battery) is simply too much for the wiring/power supply and pulling it down. Additionally, the battery will need approx 13.8v to 14.4 volts to charge it up. If the UPS does not have a boost power supply in it (to go from 12.2v up to 14.4v) then it will require a minimum of ~14-15 volts from you power supply, and hence the low voltage alarm.

hairy1

3332 posts

Uber Geek

ID Verified

Trusted

Lifetime subscriber

#2285805 30-Jul-2019 15:26

I came to this thread really hoping it was an Apollo 13 thread.... 😪

My views (except when I am looking out their windows) are not those of my employer.

neb

11294 posts

Uber Geek

Trusted

Lifetime subscriber

#2291783 7-Aug-2019 16:27

So the resolution of the issue is: The PZEM-031 meter is junk, it pulls down the supply voltage based on the load being applied. Here's a shot with the lowest-level load I could dig up, 150mA:

Even at that miniscule load it's dropped a nominal 12V supply by 0.1V, and at higher loads it drops it even more. So the only thing you can safely use it for is to measure the no-load voltage, once you start running current through it to measure power use, and particularly at higher currents, it's affecting the power it's being fed.

A problem with this is that PZEM-031s are everywhere, half the inline voltage meters are this design. Which means that either, like the CMoy headphone amp, there are an endless variety of bad copies being sold and this is one of them, or the design itself is fatally flawed. In any case caveat emptor if you're buying a PZEM-031 for power monitoring.

mclean

581 posts

Ultimate Geek

Subscriber

#2291797 7-Aug-2019 17:06

Not at all sure about your conclusions there. A wiring diagram showing how the meters wired up in relation to the supply and the load would be helpful.

neb

11294 posts

Uber Geek

Trusted

Lifetime subscriber

#2291815 7-Aug-2019 17:22

The top meter is being fed by the supply, the bottom by the top meter, and the load sits behind the bottom meter.

tripper1000

1617 posts

Uber Geek

#2292233 8-Aug-2019 12:14

Have you reversed the meter positions to prove the theory and disprove meter error?

I'm not sure sure about your conclusion either for several reasons.

1) If 100mV dropped across your meter is enough to disrupt your circuit, the design is flawed as it is sailing way to close to the wind to be reliable.

2) 100mV drop across the meter at 140mA means the sense resistor is 700 milli ohms - which doesn't add up - for starters that isn't a preferred value (eg its a funny number) and these things generally have sense resistors of 100 milli ohm or 10 milli ohm - even the cheap knock-off, they have to, otherwise they generate way too much heat and quickly fail. The current measurement accuracy is 1.0, meaning 1% of 20 amps or 200 mA, meaning you load current reading (140 mA) could be out by 1/2 or double. The more likely explanation is that the volt meters on your power meter simply can't measure to that resolution (also likely to be 1% of 100 volts) or +/- 1 volt, so you can't bank on those numbers.

Can you give us a bit more detail on that UPS? Input voltage requirements etc?

1 | 2