Can Pi be powered by stepped sine wave power generated by a UPS?

[crocket]

When a relatively affordable UPS like APC BE-550 loses mains power, it resorts to battery power. The waveform of the battery power is stepped sine wave.

Raspberry Pi 3 B+ could be shut down when it encounters tiny power cuts contained in stepped sine wave. A properly designed Active PFC power supply can work with stepped sine wave.

Can it work with stepped sine wave power output?

[W. H. Heydt]

I routinely run Pis connected through commercial "SOHO" UPSes. I get occasional power outages, all the way from "the lights blink" to multiple hours. The Pis do just fine. Bear in mind that you are not running the Pi from an AC source. You are running it from a DC source. If the "wall wart" PSU is any good at all, it will never notice the minor glitches in the stepped sine wave being fed into it. I would also note that the Pis don't have problems with the momentary period when the UPS is shifting from mains power to battery power. Most such modern UPSes are "standby"--that is, the batteries don't cut in until the line drops--and not "continuous" where the power is supplied by the batteries while the line current is charging them.

[crocket]

I get occasional power outages, all the way from "the lights blink" to multiple hours. The Pis do just fine. Bear in mind that you are not running the Pi from an AC source. You are running it from a DC source. If the "wall wart" PSU is any good at all, it will never notice the minor glitches in the stepped sine wave being fed into it.

You mean this 5V 3A AC adaptor that I ordered from aliexpress should work with stepped sine wave.

I experienced a momentary brownout once in my lifetime. I never experienced any noticeable power surges that my speakers should have picked up. I experience planned power outages once per 4~5 years. There was only one unexpected power outage in the last 20 years. Since power outages are usually planned, I am not going to buy a UPS for now. I might buy one once I start doing critical business work on my computer.

[W. H. Heydt]

I get occasional power outages, all the way from "the lights blink" to multiple hours. The Pis do just fine. Bear in mind that you are not running the Pi from an AC source. You are running it from a DC source. If the "wall wart" PSU is any good at all, it will never notice the minor glitches in the stepped sine wave being fed into it.

You mean this 5V 3A AC adaptor that I ordered from aliexpress should work with stepped sine wave.

I experienced a momentary brownout once in my lifetime. I never experienced any noticeable power surges that my speakers should have picked up. I experience planned power outages once per 4~5 years. There was only one unexpected power outage in the last 20 years. Since power outages are usually planned, I am not going to buy a UPS for now. I might buy one once I start doing critical business work on my computer.

I use the RPF PSU, see the green "Buy" tag at the top of the page. It's been many years since I had a "planned" (i.e. local utility said they were going to do work that took the power out at a set time for a set duration), but I have had--not exactly numerous, but they do happen--outages occur.

one memorable evening, the UPS alarms went off and on checking, the input voltage was way over spec, so I called the electric company to fix it. (Nominal voltage here is around 177v to 120v. The UPSes were seeing 140v.)

[crocket]

one memorable evening, the UPS alarms went off and on checking, the input voltage was way over spec, so I called the electric company to fix it. (Nominal voltage here is around 177v to 120v. The UPSes were seeing 140v.)

Can a typical consumer UPS like APC BE-550 protect equipments against constantly high voltage levels? But, that doesn't seem like a surge.

[drgeoff]

A RPi cannot be powered by an AC supply. Period. Pure sine or stepped sine wave is irrelevant to the RPi.

A RPi should be powered by 5 volts DC. So there must be something, typically called a power supply unit (PSU) between the AC supply and the RPi. How that something reacts to stepped sine wave is what the OP should have asked.

Clearly no two designs of PSU are the same but most switching types should be OK with the "modified sine wave" from a UPS. I hesitate to pontificate about PSUs with a mains frequency transformer but those are rarely used with a RPi.

[pfletch101]

Can a typical consumer UPS like APC BE-550 protect equipments against constantly high voltage levels? But, that doesn't seem like a surge.

Yes, but! It does it by switching to battery power, so the consequences for connected equipment are the same as for a power outage - you or the UPS will need to shut everything down cleanly before the UPS's battery runs out.

[rpiMike]

You mean this 5V 3A AC adaptor that I ordered from aliexpress should work with stepped sine wave.

That power supply does not look good - the switch will introduce resistance and I can't believe 3Amps can be delivered through that cable.

Buy the official power supply 5.1V 2.5A 18AWG cable, no switch. https://www.raspberrypi.org/products/ra ... er-supply/

In my experience these often keep the Pi running during very brief power outages.

[W. H. Heydt]

one memorable evening, the UPS alarms went off and on checking, the input voltage was way over spec, so I called the electric company to fix it. (Nominal voltage here is around 177v to 120v. The UPSes were seeing 140v.)

Can a typical consumer UPS like APC BE-550 protect equipments against constantly high voltage levels? But, that doesn't seem like a surge.

Within limits, yes. Check the industry specifications for normal/acceptable voltage range for your line (aka "mains") voltage. In the US, the nominal voltage is 120v,+/-5%, so 114v to 126v. That's why my UPSes threw fits when the input was 140v. Mind you, the *output* voltage at that point was 120v, so the UPSes did, in fact, protect everything behind them from the overvoltage.

There are three levels of protection from power line problems. The first is surge protection. Depending on the quality of the surge protector, it will protect devices from voltage spikes or surges. If a surge protector stops a serious spike (say, a lightning strike on a near by section of power line), it should be replaced. The next level is a line conditioner. A line conditioner will correct for undervolt (brownout) conditions and overvolt conditions. Again, there are limits. The documentation on a given line conditioner will specify just how low it can boost and high high it can cut down. Line conditioners include surge protectors.

The top level is the uninteruptable power supply (UPS). When the power fails, batteries pick up the load and provide power. As noted in this thread, they may not output a "pure" sine wave when running on the batteries. This should not matter to the protected equipment. UPSes include line conditioners...which include surge protectors. A UPS is going to have limited run time that will depend on the supported load. It will also have a specified load limit, which I will address next because it is at the intersection of technical specifications and marketing exaggeration.

In general, "Small Office/Home" (aka "SOHO") UPSes list their output rating as "volt-amperes". Regular readers will have seen some of repeatedly try to drive home the point that power (Watts) is voltage (volts) times current (Amperes), of W=vA. That is true when dealing with direct current (DC). It is not true when dealing with alternating current (AC). The output of a UPS is AC, as is the line/mains input power.

For AC power, the power goes as W = (2*-2)/2 * v * A. That is, take the volt-ampere rating and multiply by one half of the square root of 2. In practice, just multiply by 0.7. Therefore a 1000vA (aka 1KvA) UPS output should be 700W. As a practical matter, the power rating will be somewhat lower than this. The 1.5KvA UPSes that I have should--theoretically--be rated at 1050W (1500vA * 0.7), but are actually rated at 900W.

This is a pet peeve of mine. The volt-ampere rating is real, but most people aren't sufficiently trained in electrical engineering to understand what it really means in practice. You have to go through the fine print to find the rating in Watts. So far as I can tell, this is done for marketing purposes--the volt-ampere rating is a bigger number. After all, how many people faced with one device labeled "700W" (or less) and one labeled "1000vA" where the price is the same, would buy the "700W" unit, even though they are identical?

[W. H. Heydt]

Can a typical consumer UPS like APC BE-550 protect equipments against constantly high voltage levels? But, that doesn't seem like a surge.

Yes, but! It does it by switching to battery power, so the consequences for connected equipment are the same as for a power outage - you or the UPS will need to shut everything down cleanly before the UPS's battery runs out.

That depends on how high the overvoltage is. The line conditioning circuits should handle it up to a point. See my experience with 140v input on a 120v line.

[W. H. Heydt]

You mean this 5V 3A AC adaptor that I ordered from aliexpress should work with stepped sine wave.

That power supply does not look good - the switch will introduce resistance and I can't believe 3Amps can be delivered through that cable.

Buy the official power supply 5.1V 2.5A 18AWG cable, no switch. https://www.raspberrypi.org/products/ra ... er-supply/

In my experience these often keep the Pi running during very brief power outages.

And besides, 3A is above the polyfuse limit for even a Pi3B+.

[pfletch101]

Can a typical consumer UPS like APC BE-550 protect equipments against constantly high voltage levels? But, that doesn't seem like a surge.

Yes, but! It does it by switching to battery power, so the consequences for connected equipment are the same as for a power outage - you or the UPS will need to shut everything down cleanly before the UPS's battery runs out.

That depends on how high the overvoltage is. The line conditioning circuits should handle it up to a point. See my experience with 140v input on a 120v line.

From APC's Help file: "When your AC utility power source fails (a blackout) or falls outside the normal voltage range, your battery backup immediately begins supplying power to your computer system using its internal battery." {my emphasis}. Obviously, if the overvoltage is below the top of the safe range you have chosen, only the spike protection circuits will be active.

It is my understanding that typical SOHO UPSes do not have the power conditioning circuits that are found in higher-spec commercial devices. Basically, in lower-end devices, any serious power disturbances except transient spikes are handled by switching to battery power. Higher end devices actively maintain their output voltage and waveform within much tighter limits - often by operating their inverters continuously, normally supplied by DC power derived from the AC mains. This does, however, need a much beefier AC to DC conversion setup than is required if you are only using this to charge batteries, and that costs money.

[JohnsUPS]

So long as the power supply connected to the Pi puts out a clean 5v DC when power by a UPS running on its batteries, you'll be OK.
Due to the nature of how a switching power supply works, I believe that the vast majority of the supplies out there will work just fine.
Unless the UPS is a seriously inexpensive and possibly a poorly designed one emitting a stepped output with, say, lots of ringing on the edges, you should be fine.
As for a switch inline with the 5v side, any contact resistance can be ignored.


I have always known AC wattage to be:
V * I * cosine of the voltage/current phase angle (theta).
Cos .707 = .999.

[drgeoff]

I have always known AC wattage to be:
V * I * cosine of the voltage/current phase angle (theta).
Cos .707 = .999.

@JohnsUPS
Where has that .707 come from? The "half of root 2" in an earlier post is not an angle.

Also you are wrong when you say "As for a switch inline with the 5v side, any contact resistance can be ignored."

[JohnsUPS]

JohnsUPS wrote: ↑
Mon Mar 11, 2019 5:52 pm
I have always known AC wattage to be:
V * I * cosine of the voltage/current phase angle (theta).
Cos .707 = .999.
@JohnsUPS
Where has that .707 come from? The "half of root 2" in an earlier post is not an angle.

Also you are wrong when you say "As for a switch inline with the 5v side, any contact resistance can be ignored."

Unless you are passing a lot of current, contact resistance of a switch can be ignored, or the switch contact rating has been sized incorrectly. In my years of engineering (much of it analog), I have never seen this as an issue. Passing 2A isn't much current at all.

The 1/2SQRT of 2 (.707) was from an earlier post. I believe that the formula presented was probably to allow for a margin (.7PF) when sizing a UPS. Otherwise, at unity power factor, W=VA.

[W. H. Heydt]

The 1/2SQRT of 2 (.707) was from an earlier post. I believe that the formula presented was probably to allow for a margin (.7PF) when sizing a UPS. Otherwise, at unity power factor, W=VA.

AC power is generally quoted as peak-to-peak value. The 1/2 * SQRT(2) (0.707, rounded to 0.7) is to get the root mean square (RMS) value. I didn't think a dissertation on the calculus of how you get from a p-to-p sine wave to an RMS value was warranted. I'll leave that to someone whose math is rather fresher in mind than mine.

[drgeoff]

JohnsUPS wrote: ↑
Mon Mar 11, 2019 5:52 pm
I have always known AC wattage to be:
V * I * cosine of the voltage/current phase angle (theta).
Cos .707 = .999.
@JohnsUPS
Where has that .707 come from? The "half of root 2" in an earlier post is not an angle.

Also you are wrong when you say "As for a switch inline with the 5v side, any contact resistance can be ignored."

Unless you are passing a lot of current, contact resistance of a switch can be ignored, or the switch contact rating has been sized incorrectly. In my years of engineering (much of it analog), I have never seen this as an issue. Passing 2A isn't much current at all.

The 1/2SQRT of 2 (.707) was from an earlier post. I believe that the formula presented was probably to allow for a margin (.7PF) when sizing a UPS. Otherwise, at unity power factor, W=VA.

IIRC there have been posts where people found that in-line switches on the 5 volt wires to a RPi did drop enough volts to cause an issue.

You have not explained why you wrote "Cos .707 = .999". Cos is applied to angles. What angle in the discussion is 0.707 degrees?

[drgeoff]

The 1/2SQRT of 2 (.707) was from an earlier post. I believe that the formula presented was probably to allow for a margin (.7PF) when sizing a UPS. Otherwise, at unity power factor, W=VA.

AC power is generally quoted as peak-to-peak value. The 1/2 * SQRT(2) (0.707, rounded to 0.7) is to get the root mean square (RMS) value. I didn't think a dissertation on the calculus of how you get from a p-to-p sine wave to an RMS value was warranted. I'll leave that to someone whose math is rather fresher in mind than mine.

https://www.electronics-tutorials.ws/ac ... ltage.html

The section headed "RMS Voltage Analytical Method" has the mathematical derivation.

[pfletch101]

Unless you are passing a lot of current, contact resistance of a switch can be ignored, or the switch contact rating has been sized incorrectly. In my years of engineering (much of it analog), I have never seen this as an issue. Passing 2A isn't much current at all.

The contact resistance of a high-quality, correctly specified, switch in a power circuit should not be an issue. Unfortunately, many of the inline switches that come with cheap USB 'power supplies' are anything but high-quality. Even if they do not start off with enough contact resistance to cause problems, corrosion on the contact surfaces may cause this to develop in use. I suspect that in your professional life you will have been dealing with switches whose cost was (at least) a significant fraction of that of a Pi, for which your expectations would be entirely reasonable.

[crocket]

Unfortunately, many of the inline switches that come with cheap USB 'power supplies' are anything but high-quality. Even if they do not start off with enough contact resistance to cause problems, corrosion on the contact surfaces may cause this to develop in use. I suspect that in your professional life you will have been dealing with switches whose cost was (at least) a significant fraction of that of a Pi, for which your expectations would be entirely reasonable.

Would this 5V 3A microUSB cable AC adaptor be ok with APC BE-550?

[Ernst]

Unfortunately, many of the inline switches that come with cheap USB 'power supplies' are anything but high-quality. Even if they do not start off with enough contact resistance to cause problems, corrosion on the contact surfaces may cause this to develop in use. I suspect that in your professional life you will have been dealing with switches whose cost was (at least) a significant fraction of that of a Pi, for which your expectations would be entirely reasonable.

Would this 5V 3A microUSB cable AC adaptor be ok with APC BE-550?

Yes, this power supply should not be a problem with an APC BE-550.

But before you order please post another question to ask if this power supply would be ok for a Raspberry Pi.

[crocket]

But before you order please post another question to ask if this power supply would be ok for a Raspberry Pi.

Is it ok for Raspberry Pi 3 B+?

[Ernst]

But before you order please post another question to ask if this power supply would be ok for a Raspberry Pi.

Is it ok for Raspberry Pi 3 B+?

No. You could have come to this answer your self if you had read the previous posts and you had done some research in the forum.

- This "power supply" has a switch of unknown quality inserted into the cable.
- The Pi3 B+ requires a reliable power supply to satisfy the consumption, this "adapter" looks a bit weak
- The cable on this adapter looks a bit thin, not good.

If you buy this "power supply" and later run into problems then you have to spend more money to get a better power supply, you will use valuable time to find the problems, may have to restore memory cards from backup after disk corruption.

when starting with Raspberry Pi it is highly recommended to use the approved power supply:
https://www.raspberrypi.org/products/ra ... er-supply/

[crocket]

No. You could have come to this answer your self if you had read the previous posts and you had done some research in the forum.

- This "power supply" has a switch of unknown quality inserted into the cable.
- The Pi3 B+ requires a reliable power supply to satisfy the consumption, this "adapter" looks a bit weak
- The cable on this adapter looks a bit thin, not good.

If you buy this "power supply" and later run into problems then you have to spend more money to get a better power supply, you will use valuable time to find the problems, may have to restore memory cards from backup after disk corruption.

when starting with Raspberry Pi it is highly recommended to use the approved power supply:
https://www.raspberrypi.org/products/ra ... er-supply/

I was confused by words like "switching type", so I didn't understand when I read the following quote in a previous comment

the switch will introduce resistance and I can't believe 3Amps can be delivered through that cable.

[JohnsUPS]

The 1/2SQRT of 2 (.707) was from an earlier post. I believe that the formula presented was probably to allow for a margin (.7PF) when sizing a UPS. Otherwise, at unity power factor, W=VA.

AC power is generally quoted as peak-to-peak value. The 1/2 * SQRT(2) (0.707, rounded to 0.7) is to get the root mean square (RMS) value. I didn't think a dissertation on the calculus of how you get from a p-to-p sine wave to an RMS value was warranted. I'll leave that to someone whose math is rather fresher in mind than mine.

https://www.electronics-tutorials.ws/ac ... ltage.html

The section headed "RMS Voltage Analytical Method" has the mathematical derivation.

AC power in my experience is usually expressed as the result of the product of two RMS values times the power factor (phase angle). Except in certain circumstances, it isn't generally quoted as a peak value. That is the same as saying your house outlets are 169 volts. They are, in peak terms, but that isn't the convention when discussing 120V AC house circuits (well, except for maybe this and a few other technical forums). The same holds true for wattage. Since most of the time the calculations are done using RMS values, the result isn't going to be a peak value.

As for the .7PF margin mentioned above, most UPS units are sized using VA, which is just Apparent power (V*I), not taking phase shift into account. The load may shift the phase between voltage and current, and the manufacturer has no control over that. It seems that the manufacturers prefer not to quote a wattage value because once a reactive component is introduced (because of the load), this quoted watt value isn't going to do the work that many would anticipate. Seemingly they are leaving that part to the consumer.

The .7PF above also had nothing to do with obtaining an RMS value. The values going in were RMS. Obtaining RMS from P-P is trivial (pure sinusoidal waveforms). In the previous post, the Apparent power was calculated, then multiplied by .707 which essentially provides a margin of overhead for the UPS. The number could easily have been .85, or .65 or whatever.

So yes, a Pi will work from a UPS emitting a stepped sinusoidal output, so long as the power supply that is in the middle does not mind.