# What kind of UPS do Europeans use? My modified sq wave UPS + PFC'edload = CRASH !

A

#### AC/DCdude17

Jan 1, 1970
0
X-No-Archive: Yes

I have an APC UPS that puts out modified squarewave (the marketing
people calls it modified sinewave) when on battery. I connected a
fluorescent fixture with an electronic ballast containing passive L-C
power factor correction. The UPS crashed and shut down as soon as it
switched over to battery. The UPS doesn't have issues with residential
electronic ballast such as CFLs which doesn't have a PFC circuit. It
must not be getting along with the PFC. UPS is rated at 600VA. Light
only takes 100VA.

Electronic ballasts have the identical rectififier+capacitor front end
as switchmode power supplies used in computers. Like electronic
ballasts, computer power supplies are available with PFC. They're rare
in the US, but EU legislations mandated power factor corrected power
supplies on new computers.

An example of power factor corrected PSU:
http://www.savastore.com/products/product.asp?catalog_name=Savastore&product_id=10262447&pid=44

The passive PFC on a power supply like the one linked above is the same
type as the one in my ballast. Since computer power supplies in the US
typically don't have a PFC, we're fine, but how do Europeans get their
PFC equipped computers to get along with their UPS? I searched "PFC
UPS" on Google and looks like I'm not the first one to have issues with
a UPS interecting badly with a PFC power supply.

My UPS shares the transformer between inverter operation and charging.
It's a steel core transformer with center tapped secondary. When
operating on battery, the center tap on secondary is tied to one of the
battery terminals and the MOSFETs switch the other terminals back and
forth between the two taps on the sides. A relatively common design.

P

#### [email protected]

Jan 1, 1970
0
| An example of power factor corrected PSU:
| http://www.savastore.com/products/product.asp?catalog_name=Savastore&product_id=10262447&pid=44

Do you have any schematics or design theory?

| The passive PFC on a power supply like the one linked above is the same
| type as the one in my ballast. Since computer power supplies in the US
| typically don't have a PFC, we're fine, but how do Europeans get their
| PFC equipped computers to get along with their UPS? I searched "PFC
| UPS" on Google and looks like I'm not the first one to have issues with
| a UPS interecting badly with a PFC power supply.

I would have thought that a proper power factor corrected load would be
indistinguishable from a linear resistive load. Maybe what they are
talking about is just the phase angle, rather than harmonics? Maybe
this kind of PFC is actually making harmonics worse?

| My UPS shares the transformer between inverter operation and charging.
| It's a steel core transformer with center tapped secondary. When
| operating on battery, the center tap on secondary is tied to one of the
| battery terminals and the MOSFETs switch the other terminals back and
| forth between the two taps on the sides. A relatively common design.

And that's not producing any backfeed?

A

#### AC/DCdude17

Jan 1, 1970
0
X-No-Archive: Yes

| An example of power factor corrected PSU:
| http://www.savastore.com/products/product.asp?catalog_name=Savastore&product_id=10262447&pid=44

Do you have any schematics or design theory?

I don't know about that specific power supply, but this is the general
schematic of a passive power factor correction that goes in the input
stage of a switch mode power supply or a ballast:

http://www.irf.com/technical-info/designtp/temp004.pdf

Look at figure 2 on page 2.

This L-C type circuit is tuned to work with sine wave of specific
frequency.

All ballast meant for commercial use have power factor correction of
some sort so that it will get at least PF >0.95 most likely because
there's some regulations mandating it.

Residential application ballast doesn't have any form of front end
circuit to correct PFC and they don't have a problem with my UPS.
Apparently the extremely high harmonic content modified squarewave
doesn't play nicely with L-C circuit tuned to work with sinewave.

The inductors are actually quite large. Despite the weight, it's the
preferred choice for many electronic ballasts, because it's the cheapest
design and works quite well.

| The passive PFC on a power supply like the one linked above is the same
| type as the one in my ballast. Since computer power supplies in the US
| typically don't have a PFC, we're fine, but how do Europeans get their
| PFC equipped computers to get along with their UPS? I searched "PFC
| UPS" on Google and looks like I'm not the first one to have issues with
| a UPS interecting badly with a PFC power supply.

I would have thought that a proper power factor corrected load would be
indistinguishable from a linear resistive load. Maybe what they are
talking about is just the phase angle, rather than harmonics? Maybe
this kind of PFC is actually making harmonics worse?

That's the goal, but there's still harmonics. Again the L-C circuit is
designed to limit the current rise so that it doesn't abruptly rise
while the voltage sweeps in sinewave. Since in modified squarewave,
voltage goes from zero to full almost instantaneously, i'm sure the
situation is very different.

This is what modified squarewave looks like:

http://www.fords-mtm.com/electric images/waveforma.jpg
even though the sales likes to call that modified sinewave, it's not.

The only reason it's called modified is because there's a slight pause
before polarity is reversed.

This might qualify as modified sinewave though:
http://www.wojack.net/Our Solar System/MODIFIED_SINE_WAVE.GIF

| My UPS shares the transformer between inverter operation and charging.
| It's a steel core transformer with center tapped secondary. When
| operating on battery, the center tap on secondary is tied to one of the
| battery terminals and the MOSFETs switch the other terminals back and
| forth between the two taps on the sides. A relatively common design.

And that's not producing any backfeed?

Not sure, I haven't put the effort to reverse engineer the UPS any further.

W

#### w_tom

Jan 1, 1970
0
Because a computer power supply is so resilient, then these
plug-in UPSes don't normally affect the computer. However
that modified sine wave is a threat to other less robust
appliances such as small electric motors. UPS manufacturer
quietly note that even a plug-in surge protector on the output
of that UPS can cause serious problems when UPS is in battery
backup mode. The fluorescent ballast as a load would only
complicate the operation of that UPS in battery backup mode.

You have also demonstrated that UPS connects computer
directly to AC mains when not in battery backup mode which is
why problems only exist in battery backup mode.

PFC is not an issue. Different appliances will place
different loads at different frequencies (harmonics)
regardless of PFC circuits. That 'so called' sine wave output
is really power at many frequencies (harmonics). No problem
to computers. But a serious problem to small electric motors
manufacturers don't like the public informed that a typical
plug-in UPS outputs some of its dirtiest electricity when in
battery backup mode. Such information tends to undermine
myths that promote more UPS sales such as the resiliency of
computers.

A

#### AC/DCdude17

Jan 1, 1970
0
X-No-Archive: Yes

w_tom said:
Because a computer power supply is so resilient, then these
plug-in UPSes don't normally affect the computer. However
that modified sine wave is a threat to other less robust
appliances such as small electric motors. UPS manufacturer
quietly note that even a plug-in surge protector on the output
of that UPS can cause serious problems when UPS is in battery
backup mode.

You're missing the point entirely. I was not talking about potential
damage to the computer.
The fluorescent ballast as a load would only
complicate the operation of that UPS in battery backup mode.

An electronic fluorescent ballast has the same front end as a switch
mode poewr supply. The front end is made of rectifier and smoothing
capacitors.

P ower factor corrected ballasts and power supplies simply have an
inductor and a capacitor upstream of the rectifier input.

You have also demonstrated that UPS connects computer
directly to AC mains when not in battery backup mode which is
why problems only exist in battery backup mode.

PFC is not an issue. Different appliances will place
different loads at different frequencies (harmonics)
regardless of PFC circuits. That 'so called' sine wave output
is really power at many frequencies (harmonics). No problem
to computers.

Yes, the passive PFC is an issue with my UPS. Not because things smoke,
but the reactor/capacitor PFC circuit adversely reacts with the UPS
causing it to crash(power cuts out immediately) which defeats the
purpose of having an UNinterruptable power supply.

I know for fact it is the PFC front end that's causing the problem.
When I bypass the PFC by putting a rectifier and a capacitor before the
ballast (which makes the passive PFC a spectator since inductance only
responds to A/C), UPS doesn't crash. This is a good way to determine
the cause, but it's not a practical real life solution.

W

#### w_tom

Jan 1, 1970
0
My point apparently was not obvious. Because computers are
so resilient, then these plug-in UPSes "do it on the cheap";
output dirty, crappy electricity when in battery backup mode.
I was not talking about a UPS damaging computers. I
demonstrated why plug-in UPSes output very dirty power - power
that may be destructive to some small electric motors.
Because computers are so resilient and would not be harmed by
that "dirty" UPS electricity.

What is the frequency response of that PFC circuitry - if it
really is a PFC circuit. Is circuit the EMI filter that is
required by EU (and other) regulations? Then, does that
filter impedance increase or decrease with higher frequency?
These line interfaces vary in design so that only you can tell
us what that circuit really is. But a line interface circuit
must include a line filter. Does the line filter design cause
lower impedance (short out) those higher frequency
components? If so, then that is the answer of why a line
filter (or PFC circuit) caused UPS problems.

A UPS with high THD (too much power in higher frequency
harmonics) would see a short circuit in that filter.

Again, most plug-in UPSes are built as cheaply as possible.
So cheap because computers are so resilient; are not adversely
affected by poor quality UPS power when in battery backup
mode. If you can even find one, try a UPS with a Total
Harmonic Distortion of less than 5%. Good luck in that
search. Most plug-in UPSes tend to be on the order of 20 to
30% THD which is why there is so much power in harmonics and
why your ballast circuit (its line filter) would have problems
with that UPS.

Part of the problem finding that UPS - many manufacturers
fear to inform consumers since many UPSes are purchased on
mythical recommendations. Complete set of numerical
specifications for UPSes are difficult to locate.

A

#### AC/DCdude17

Jan 1, 1970
0
X-No-Archive: Yes

w_tom said:
My point apparently was not obvious. Because computers are
so resilient, then these plug-in UPSes "do it on the cheap";
output dirty, crappy electricity when in battery backup mode.
I was not talking about a UPS damaging computers. I
demonstrated why plug-in UPSes output very dirty power - power
that may be destructive to some small electric motors.
Because computers are so resilient and would not be harmed by
that "dirty" UPS electricity.

Isn't it sad that the UPS that's giving me a headache is one of the
smaller APC SMART UPS? This isn't supposed to be like the $29.99 junk. What is the frequency response of that PFC circuitry - if it really is a PFC circuit. Is circuit the EMI filter that is required by EU (and other) regulations? I don't know the frequency response. The series reactor extends the current rise time of the typical rectified load and it is tuned to give a PF of >0.95 when used with a constant load on 60Hz 120V line. The EU regulation now requires PFC on computer power supplies. EMI filter was most likely mandated long before that. Then, does that filter impedance increase or decrease with higher frequency? These line interfaces vary in design so that only you can tell us what that circuit really is. But a line interface circuit must include a line filter. Does the line filter design cause lower impedance (short out) those higher frequency components? If so, then that is the answer of why a line filter (or PFC circuit) caused UPS problems. See the PDF file in my previous post. It shows you how the typical PFC is setup. I don't have any specific information on the actual components in ballast. A UPS with high THD (too much power in higher frequency harmonics) would see a short circuit in that filter. hmm mine says 28% total harmonics, 21% single harmonic. W #### w_tom Jan 1, 1970 0 Example UPS output is a modified sine wave. Output under light load is two 200 volt 'very sharp' square waves with a 270 volt spike between those square waves. That is called a 120 VAC modified sine wave. That would create hell with most everything but a computer. Why do replacement batteries cost almost as much as the entire UPS? Plug-in UPSes for computers are made a cheaply as possible because computers are so resilient and because consumers so often 'wish' those UPSes do things they really cannot do. Look at what is called a modified since wave in the example UPS. Two 200 volt square waves. People have even claimed that could not be true because it said 'sine wave' on the box. Manufacturer did not lie. Manufacturer simply demonstrated propaganda. A square wave is nothing more than lots of different sine waves simultaneously. The only person who was lying was the consumer who wanted to believe something different. Find, if you can, a UPS with low THD. Clearly you have the typically worst THD I have seen in UPSes. It sounds like you may have a passive PFC circuit which is really little more than a filter. A filter that also provides EMI protection. There are many ways to meet PFC requirements - both active and passive. Incidentally, I believe the EU has since weakened its PFC requirements. The typical plug-in UPS is good for protecting data from blackout and extreme brownouts. It is why they are so cheap and why they are dangerous to small motors and power strip surge protectors. And yet still many will even insist that UPS also provides hardware protection to 'sensitive' computers. You have now seen the numbers AND have experienced what those numbers mean - "28% total harmonics, 21% single harmonic". This is how we learn by obtaining both the theory (specifications) and experimental evidence (your ballast connect to that UPS). A #### AC/DCdude17 Jan 1, 1970 0 X-No-Archive: Yes w_tom said: Example UPS output is a modified sine wave. Output under light load is two 200 volt 'very sharp' square waves with a 270 volt spike between those square waves. That is called a 120 VAC modified sine wave. That would create hell with most everything but a computer. It's the same type waveform as a typical 12V-110V travel inverter. Why do replacement batteries cost almost as much as the entire UPS? Plug-in UPSes for computers are made a cheaply as Not on mine. http://www.controlcable.com/details/item.asp?id=35511018|AP112| Battery is about 1/4 to 1/6 the retail price of the UPS. possible because computers are so resilient and because consumers so often 'wish' those UPSes do things they really cannot do. Look at what is called a modified since wave in the example UPS. Two 200 volt square waves. Actually it's more like 150V. The peak is lower than the sinusoidal peak of 120V. Find, if you can, a UPS with low THD. Clearly you have the typically worst THD I have seen in UPSes. Which would be the sinewave Smart UPS. So going back original question. Do Europeans buy sinewave UPS just so it would be compatible with their PFC computer power supplies? They're VERY expensive. It sounds like you may have a passive PFC circuit which is really little more than a filter. A filter that also provides EMI protection. There are many ways to meet PFC requirements - both active and passive. Incidentally, I believe the EU has since weakened its PFC requirements. The typical plug-in UPS is good for protecting data from blackout and extreme brownouts. It is why they are so cheap and why they are dangerous to small motors and power strip surge protectors. And yet still many will even insist that UPS also provides hardware protection to 'sensitive' computers. You have now seen the numbers AND have experienced what those numbers mean - "28% total harmonics, 21% single harmonic". This is how we learn by obtaining both the theory (specifications) and experimental evidence (your ballast connect to that UPS). APC advertises their Smart UPS as for Network Computers and Servers.... B #### bushbadee Jan 1, 1970 0 Let me explain something about sine wave ouput converters. First the input to the chopping transistors is a DC. This is now chopped up and you have a square wave . NOw to get a sine wave output, (Other than useing a bunch of choppers and makeing a stepped wave which costs even more) You must have an LC filter on the output . Generally this is a parrellel resonent filter that reduces higher order harmonics. Now you have a square wave coming out of the chopper and a sine wave output. For this to be, there must be a series element of some kind to absorb the difference between the square wave output of the choper and the sine wave output. Generally an inductor is used to perform this function, being relitively losseless. But the fact is that the Inductor to absorb the volt seconds and to carry the DC generally winds up being about half the size of the output transformer. There is no getting around this. The inductor or some series element must be there to act as a voltage absorber between the sine wave output and the square wave. Now there is another way to acomplish this function. The Inductor can be put in series between the input voltage and the center tap of the output transformer so the voltage upon the switching transistors is a partial sine wave. There are some draw backs to this. The voltage on the transistors will now be higher. But there is another advantage. The output transformer can now become part of the LC filter. But there is a problem with this. If the sine wave is to have less than about 5% harmonic distortion the current in the output side of the transformer must be about 5 times the load current or it must have a Q of about 5. But if the output transformer was not part of the tuned circuit, the ouput filter inductor would still have to carry about 5 times load current and have a Q of about 5. So there is little added weight with the much bigger output transformer. Now to get the harmonic distortion down below about 3% you generally need a series LC fiter across the output. But in general this is quite small. The use of the inductor ahead of the transformer gives you a lot of advantages if you want short circuit protection. By using a tapped inductor, you can arrange to have 0 volts out at other than 0 pulse width. This reduces the pulse width range that the converter has to work over between full output voltage and 0 volts into a short circuit. Such converters were built for the military by a company called Electrosolids some years ago. For more information on these types of converters see my series on Magnetics and Converters. M #### Mike Tomlinson Jan 1, 1970 0 AC/DCdude17 said: Since computer power supplies in the US typically don't have a PFC, we're fine, but how do Europeans get their PFC equipped computers to get along with their UPS? I run an APC SmartUPS 600 and a PC with a Fortron/Source power supply with passive PFC without any problems. Just a data point. A #### AC/DCdude17 Jan 1, 1970 0 X-No-Archive: Yes Mike Tomlinson wrote: I run an APC SmartUPS 600 and a PC with a Fortron/Source power supply with passive PFC without any problems. Just a data point. Hello, You have a nicer UPS than I do. I did some data digging on APC site and found that: The your UPS has a SINEWAVE inverter. Your APC SmartUPS 600VA: http://sturgeon.apcc.com/techref.nsf/partnum/990-0600A/$FILE/0600-9.pdf

On page 38, it says "on battery waveshape: sinewave"
Frequency tolerance on your unit is +/- 0.1 Hz

Your UPS is no longer made and APC actually recommends my UPS as a
direct replacement (SmartUPS 620VA) (
http://www.apc.com/support/tech_refs.cfm ) which is a seriously
economized version. It doesn't have all the bells and whistle (i.e.
sensitivity adjustment) and the output waveform is MODIFIED
SQUAREWAVE!!! AAARGH!!!

Manual for SmartUPS 620VA

http://sturgeon.apcc.com/techref.nsf/partnum/990-7041/$FILE/D7041-1E.pdf "On battery waveshape: Stepped sinewave" This isn't one of the better stepped sinewave that kinda goes up and down like a stairway. It's a "modified squarewave" that goes from zero to full at once which I verified on my oscilloscope. I think this model puts the SmartUPS name to shame. I personally feel this unit should have been put on BackUPS line. I believe you having a sinewave and me having a modified squarewave unit is a good explanation for the discrepancy in our result. A #### AC/DCdude17 Jan 1, 1970 0 X-No-Archive: Yes Mike Tomlinson wrote: I run an APC SmartUPS 600 and a PC with a Fortron/Source power supply with passive PFC without any problems. Just a data point. Hello, You have a nicer UPS than I do. I did some data digging on APC site and found that: Your UPS has a SINEWAVE inverter. Your APC SmartUPS 600VA: http://sturgeon.apcc.com/techref.nsf/partnum/990-0600A/$FILE/0600-9.pdf

On page 38, it says "on battery waveshape: sinewave"
Frequency tolerance on your unit is ± 0.1 Hz

Your UPS is no longer made and APC actually recommends my UPS as a
direct replacement (SmartUPS 620VA) (
http://www.apc.com/support/tech_refs.cfm ) which is a seriously
economized version. It doesn't have all the bells and whistle (i.e.
sensitivity adjustment) and the output waveform is MODIFIED
SQUAREWAVE!!! AAARGH!!!

Manual for SmartUPS 620VA

http://sturgeon.apcc.com/techref.nsf/partnum/990-7041/\$FILE/D7041-1E.pdf

"On battery waveshape: Stepped sinewave" This isn't one of the better
stepped sinewave that kinda goes up and down like a stairway. It's a
"modified squarewave" that goes from zero to full at once which I
verified on my oscilloscope.

I think this model puts the SmartUPS name to shame. I personally feel
this unit should have been put on BackUPS line.

I believe you having a sinewave and me having a modified squarewave unit
is a good explanation for the discrepancy in our result.

M

#### Mike Tomlinson

Jan 1, 1970
0
AC/DCdude17 said:
You have a nicer UPS than I do. I did some data digging on APC site and
found that:

Your UPS has a SINEWAVE inverter.

Thank you. Handy to know.

and thanks for that. Grabbed for future reference (I bought the unit
secondhand, hence no manual, and replaced the batteries.)
I believe you having a sinewave and me having a modified squarewave unit
is a good explanation for the discrepancy in our result.

Possibly. I don't have enough direct experience to comment.

Replies
2
Views
447
Replies
8
Views
1K
Replies
0
Views
1K
Replies
0
Views
2K
Replies
3
Views
1K