AC Power - Using a UPS

CNC Machines Shapeoko
1

I’m a computer guy, and I have uninterruptible power supplies (UPS) all over my house. I buy new ones for my computers, and the old ones find homes on TVs, clocks, etc.

I’ve owned my Nomad for over 5 years, and without even thinking about it I have always had it plugged into one of the hand-me-down UPS. I am currently in the process of building my Shapeoko, which will be in the garage in it’s own enclosure. I’ve been working on the enclosure and Shapeoko for nearly a month now, and I am finally at the tail end of the build to the point that I should run my first project this weekend. But back to the point of this conversion, the UPS…

I put an older UPS (with new batteries) inside this enclosure, and one of the things I have in the enclosure is a simple AC powered fan. When I was working on the wiring for the enclosure I turned on the fan, and then unplugged the UPS from the wall (to simulate a power failure). The UPS switched over to batteries and the fan (and light) continued to work just like it should have. Except I (easily) noticed the fan SOUNDED different, slower. I could even feel the air flow difference. I plugged the UPS back in and the fan spun faster, unplugged and it spun slower.

This sent me down a rabbit hole, but I can tell you this observation was not unique to that specific UPS unit. I observed this on 4 of the 5 UPS I tested. But to simplify this discussion, here are the results I recorded when running the Carbide router off that original UPS (router power/speed to low/1).

When the UPS was plugged in, running off AC power. Here are the numbers I recorded…
118.4 volts
2.41 amps
76 watts
276 volt amps
59.9 hz

… and when the UPS switched to battery (pulled the plug).
104.8 volts
3.16 amps
95 watts
344 volt amps
hz showed an error msg (not a ture sine wave ups).

And the sound, the router was not consistent, it sounded slower and it sounded like it was struggling. I stopped the test, and was disappointed. These were the results using an $120 class UPS, my initial testing (with the fan) got different results on my most expensive UPS ($400 range). So I used that most expensive UPS and the Carbide router, and of course it did significantly better, no struggling, no audible difference in power (as my credit card screamed in pain).

So what is this telling me. On one hand I’m assuming that ANY UPS is still better than no UPS, especially if the power outage is very short (or a brown out situation). But I didn’t expect the routers performance to drop. A badly timed power droop or blink could still result in a CNC failure. But do I use the older UPS knowing this, or do I spend $400 on a UPS… probably not.

I am not an electrical engineer, would LOVE to hear some options of those that know more about this stuff.

3 Likes
2

There’s been a bit of discussion on this in the past — apparently there’s a difference in the electronics and it takes a fairly expensive UPS which has specific electrical characteristics to support the machines well.

3

First thing I would ask is what provided the numbers? The UPS? Or a separate power meter? If the UPS, the frequency meter may be showing the input power, not the output. Pull the plug, no input.

Several other things that come into play:

An AC fan will likely not do well when the supply isn’t a pretty good sine wave. The harmonics above the 60HZ principal do little to nothing driving it, but the first several will heat it, and the second at 120Hz will REALLY change the motor sound. No surprise it sounded weird. This is one of several reasons why you don’t want to plug most refrigerators or AC units into a UPS, even if it is nominally capable power-wise.Some modern load-sensing AC and 'frige controls actually do well on UPS’s, since they use VFD drives, as long as the surge load is in bound.

The current increase is unsurprising, as well, for several reasons. Any inherently AC load, like the fan, will have greater waste for non-sine-wave, and most electronic loads- switching supplies, for example- draw roughly constant power, with somewhat greater loss at lower supply voltage. Many also have input filtering that shunts the harmonics, so current at the harmonic frequencies is waste power.

The router getting funky could be a number of things, and I’d need to know what the unit is, and probably look at scope traces, to do more than guess. The lower speed would be typical of lower voltage for a universal-type (brushed) motor, though, and poor stability of the UPS with that type of load would also not be surprising.

Why the output voltage drops to 105 is a different question. I would expect 115V, no lower than 108V, as so much equipment is brownout sensitive, but you’d need to read the specs for the unit to know. (North America is 120V nominal, +/-6%, at the supply, but ill defined at point of use, though assumed to be at most about 5% lower.)

3 Likes
4

I am also NOT an expert on UPS technology, except to know that most inexpensive UPS systems used “stepped” sine waves, which works. . . But is not the best. The APC “Smart UPS” line use something that is probably not a true sine wave, but is really close. I use the Smart UPS line on my 3D printers and SO3, but the regular UPS line on my TV, stereo, XBOX. . .

Here is a post of me running my SO3 and router on my APC UPS and just pulling the plug on the UPS. The Shapeoko doesn’t even notice it.

5 Likes
5

I had a power meter between the Carbide router and the UPS. So NOT the specs reported by the UPS, for just the reasons you stated.

From the entire experiment, the 104.8 volts is what concerned me the most. I’m going to move forward with my build/assembly using to current (questionable) UPS. Once I finish the assembly I will redo this test with the full setup including the bit runner, using the Carbide router at higher settings.

APC UPS’s are sold using “VA” (Volt Amp) stats. The ones I tested ranged from 1000VA to 1500VA, all well above the measured 344va, so spending more $$ doesn’t always mean this problem will be fixed. I suspect this is something in the electronics of the UPS, just wish I knew what buzz word to look for when shopping for a UPS. But that is no replacement for real world tests.

Just wanted to share this information as the results were not what I thought they would be.

3 Likes
6

THANK YOU,

When I said that the fan did NOT sound different on 1 of the 5 UPS units I tested, that unit was an APC Smart UPS… so that’s probably what I need to use.

For those interested it’s $280 to $350 depending on VA capacity.

2 Likes
7

The VA spec is used as the primary spec, rather than watts, because many loads they are expected to run do not look like resistors. The current and voltage are not in phase. Since they are typically designed for powering electronic loads, which may look from moderately inductive to slightly capacitive, and after startup usually have power factors fairly close to 1, I would be unsurprised by stability issues with a significantly inductive load.

I am a number of years out of date here on UPS state of the art, but the core EE stuff hasn’t changed. I can’t really help much more, but would love to know what you come up with.

Side note, since I can’t resist after how many months cooped up: Back in the day, a fun demo to do used a black box in series with an ammeter and an incandescent lamp. The box had a button on it. Push the button, and the meter dropped, but the lamp got brighter. Dropped a lot (maybe 20 to 30%) and the lamp got a LOT brighter.

This goes back well before modern electronics. The secret? Pushing the button changed the primary winding on the transformer, forcing the core into saturation and the output was decreased a bit in peak. This produced an output pretty close to a square wave. More power delivered to the lamp, but the mean EDIT: current dropped. Really drove home to a lecture hall full of EE second years the difference between RMS power and the product of ‘average’ (mean) current and voltage.

2 Likes
8

What you’re looking for to get optimal performance is an on-line UPS with pure sine wave output. You’ll pay a big premium for that though, so you might get by ok with a line interactive UPS. It’s able to handle brownouts and overvoltage without switching to battery. You can also save some money by going with a TrippLite or Cyber Power instead of APC. Here’s a decent article describing the differences: https://blog.tripplite.com/line-interactive-vs-on-line-network-ups-systems-and-which-should-you-choose/

Here’s another that goes into a little more detail: https://minutemanups.com/line-interactive-vs-online-ups/

6 Likes
9

The buzz words you are looking for are " pure sine wave’" and along with that you will see a big change in $$.

Switching power supplies draw spikes of current and therefor the average current can be much smaller than the spikes. The inverter must be able to handle the max current in the spikes. There is a great write up on this in an inverter manual - Samlex PST-300-12. You can get it on donrowe.com.

Transformers don’t always like the modified square wave inverters as the cores are not always designed to handle the frequency of the harmonics. Depending on the router you have, it may be saturating due to the harmonics.

1 Like
10

It might be useful if there was a way to have the UPS monitoring software automatically pause carbide motion/whatever sender you are using and switch the router off if the power drops for more then 30 seconds or something. Depending on what you are cutting things could go downhill if your dust collection isn’t on battery backup as well (deep slots filled with chips).

Of course implementing something like that isn’t trivial, especially if you don’t have the bitrunner or some other router control already tied into your controller/sender (assuming the sender turns off the router on pause, I don’t know if that is the case with CM).

1 Like
11

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.