Author Topic: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter  (Read 108445 times)

0 Members and 2 Guests are viewing this topic.

Offline theoldwizard1

  • Regular Contributor
  • *
  • Posts: 171
Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
« Reply #25 on: December 20, 2014, 03:27:40 pm »
Great write up  !  Love the pictures !

I just finished Knurlgnar24's video series on this unit and am happy to see that someone else could successfully perform the upgrades he suggested.  As an inverter there are more efficient ones on the market, but you certainly won't be buying one for that kind of price !  Besides, most of the losses due to heat are from the transformer.

I have a couple of questions
  • Has the increase in battery wire size (to approximately 8 AWG) reduced the temperature rise in those wires ?  (Getting rid of the Anderson connector was a big "win")
  • Any good locations on the board where you could bolt a "lug" to the board and then bolt, say a 6 AWG cable ?
  • Did you ever add a second capacitor ? What brand, model, size ?
  • Did you do anything about improving air flow over the transformer ?
  • The site you posted the partial schematic from appears to be dead.  Can you re-post the entire schematic ?
  • Any other sites you can recommend for information on this and other APC units ?

Edit(gnif): Fixed the YouTube link.
« Last Edit: April 04, 2017, 12:12:41 pm by gnif »
 

Offline cte7dsTopic starter

  • Contributor
  • Posts: 47
  • Country: fi
  • hello world
    • my github
Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
« Reply #26 on: December 26, 2014, 05:11:03 pm »
    Great write up  !  Love the pictures !

    I just finished Knurlgnar24's video series on this unit and am happy to see that someone else could successfully perform the upgrades he suggested.  As an inverter there are more efficient ones on the market, but you certainly won't be buying one for that kind of price !  Besides, most of the losses due to heat are from the transformer.

    I have a couple of questions
    • Has the increase in battery wire size (to approximately 8 AWG) reduced the temperature rise in those wires ?  (Getting rid of the Anderson connector was a big "win")
    I approximately halved the resistance in the cabling by taking out the original battery wires and the Anderson connector. The 1-meter 10mm^2 cables get only slightly warm at full load.
    Quote
    • Any good locations on the board where you could bolt a "lug" to the board and then bolt, say a 6 AWG cable ?
    You probably could install PCB mount wire lug terminals like this:
    Quote
    • Did you ever add a second capacitor ? What brand, model, size ?
    I decided not to bother with the second capacitor because I didn't have a good quality one at hand. But if I ever need to open up the UPS again I might replace the cap with something like 2 x 4,700 - 10,000 uF @ 50-63 V low-esr Panasonic caps.
    Quote
    • Did you do anything about improving air flow over the transformer ?
    I couldn't think of any easy way to improve the air flow over the transformer so I didn't bother.
    Quote
    • The site you posted the partial schematic from appears to be dead.  Can you re-post the entire schematic ?
    I don't have the schematics anymore, and they were for the older 1400VA model anyways. This is the best I could find: http://master-tv.com/article/apc-back-ups/
    Quote
    • Any other sites you can recommend for information on this and other APC units ?
    Most of the links I could find were 404, But heres a few that worked:
    http://www.techno-fandom.org/~hobbit/power/dumbups/
    http://www.techno-fandom.org/~hobbit/power/dumbups/res.html
    Quote
    [/list]
     

    Offline theoldwizard1

    • Regular Contributor
    • *
    • Posts: 171
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #27 on: December 26, 2014, 06:30:51 pm »
    .
    .
    .
    One other modification idea that has crossed my mind is that if I ever find another Smart-UPS 1500 for really cheap, it would be interesting to take the transformer out of it and hook it up in parallel with the existing transformer. This would essentially convert the UPS into the XL (Extended Runtime) model, I think... (I remember seeing the internals of the XL model and it was just two transformers in parallel (but maybe those transformers need to be carefully matched?)). There would actually be room inside the UPS for another transformer because I don't have the internal batteries.
    Based on Dave's inspection video of a 2200XL, I think the main feature of the XL models us additional cooling.

    There is a good deal on a pair of functioning 3000XL units within a couple hour drive of my house.  The down side on them is 48V battery bank.  I would much prefer a 24V battery bank !

    Edit(gnif): Fixed the YouTube link.
    « Last Edit: April 04, 2017, 12:13:31 pm by gnif »
     

    Offline cte7dsTopic starter

    • Contributor
    • Posts: 47
    • Country: fi
    • hello world
      • my github
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #28 on: December 26, 2014, 08:55:59 pm »
    .
    .
    .
    One other modification idea that has crossed my mind is that if I ever find another Smart-UPS 1500 for really cheap, it would be interesting to take the transformer out of it and hook it up in parallel with the existing transformer. This would essentially convert the UPS into the XL (Extended Runtime) model, I think... (I remember seeing the internals of the XL model and it was just two transformers in parallel (but maybe those transformers need to be carefully matched?)). There would actually be room inside the UPS for another transformer because I don't have the internal batteries.
    Based on Dave's inspection video of a 2200XL, I think the main feature of the XL models us additional cooling.


    There is a good deal on a pair of functioning 3000XL units within a couple hour drive of my house.  The down side on them is 48V battery bank.  I would much prefer a 24V battery bank !
    Sure, a 48V bank is more awkward to set up and handle and it doesn't look as neat, but it also means you will be getting more energy out of your batteries because they are being discharged with a smaller current (Peukert's law).

    I haven't been able to confirm exactly how the XL models differ from the non-XL models, but at the very least it gives more assurances that the unit will not overheat during longer periods of use, and that's definitely nice to have.

    Also getting a pair of them is great because then you have a spare if something goes wrong. You probably should go for it if the price is right :)
     

    Offline theoldwizard1

    • Regular Contributor
    • *
    • Posts: 171
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #29 on: December 26, 2014, 10:37:56 pm »
    Also getting a pair of them is great because then you have a spare if something goes wrong. You probably should go for it if the price is right :)
    If I really had a need, I would be after them in a heartbeat !  Right now they would just sit around and gather dust.

    Did you watch Knurlgnar24's video on converting a pair of APC UPS into 240v split phase ?  I would love to see some 'scope traces and testing on that setup !!

    Edit(gnif): Fixed the YouTube link.
    « Last Edit: April 04, 2017, 12:13:50 pm by gnif »
     

    Offline johansen

    • Frequent Contributor
    • **
    • Posts: 977
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #30 on: January 17, 2015, 02:07:14 am »
    a worthwhile mod is to place an inductor in series with the low voltage side of the transformer.

    these machines use the leakage inductance of the transformer to filter the pwm and its very lossy.
    my apc1000 when placing an inductor in series with the transformer, it not only reduced the no load loss by 10 watts, but it also reduced the switching losses because it reduced the pwm frequency.

    another mod is taking the 20uf capacitor off the 120vac side of the transformer and replace it with 4.7uf. this saves you about 120va's of circulating current and about 1-2 watts in power loss.

    I used a ferrite core of 25mm diameter and 7 turns of wire, gapped with a couple sheets of paper.
    don't use an iron powder core.. i was actually able to increase power loss by attempting to use two dozen t-106 yellow cores, by slipping them over the wires of the transformer.

    the ferrite core from a crt flyback transformer could work.. but you would need a lot of copper and a big air gap for a 1kw load. --2 of them or 4 of them would be better.

    the inductor will also reduce the no load loss when charging the battery.. mine did not use the linear region of the fets, but rather used the h bridge as a boost converter to charge the battery. this requires pwm, and pwm means the transformer gets warm.

    i don't recall what the reduction at the outlet was, but the transformer probably ran 5C cooler.
    « Last Edit: January 17, 2015, 02:10:20 am by johansen »
     

    Offline darkyputz

    • Newbie
    • Posts: 6
    Hello out there...
    Just found this article...and i am very impressed...
    I have now 2 of these devices sitting in front of me, and i would like to get one apart to complete the other...with the right parts...of course...
    But my question is...what is the main benefit of doing that?
    More Long time stable?
    More Watt?
    Better cooling?
    Not quiet sure...
    Can please someone explain?
    And an another question...
    Is there a way to tell the device that i is 230 volts from now on and 50 hertz?
    Or is this a total different board design than?

    Thx in advance...
     

    Offline johansen

    • Frequent Contributor
    • **
    • Posts: 977
    But my question is...what is the main benefit of doing that?

    benifit of doing what exactly.

    adding more mosfets to reduce the losses? the point is to reduce losses.
    adding an inductor works as well to significantly reduce switching losses and transformer iron and transformer copper losses.

    i do not believe you will be able to convert your inverters from 120v 60 hz to 230v 50hz.

    you could rewind the transformer to get a higher voltage.. but that is a lot of work.. you would have much better success winding a new toroidal transformer.
     

    Offline darkyputz

    • Newbie
    • Posts: 6
    Hello and thx for your answer...
    Yes...i meant putting the 8 missing pieces in and the regarding gates...and the cap...
    Or would you stay with two intact ones untouched instead of putting the parts from the second in the first to get it complete?
    And what did you do with yours in the reply prior mine?
    You said you added something somewhere...can you be some more detailed or have a pic?

    Thx in advance
     

    Offline darkyputz

    • Newbie
    • Posts: 6
    And onother question to this...
    Because i have 2 of the same ups...could i put the second trnsformer parrallel to the first and get some benefit from that?
    and if yes...would it be just connect the second to the same as the first?
     

    Offline johansen

    • Frequent Contributor
    • **
    • Posts: 977
    you can pull the mosfets from one board and install them on the other.. no problem there.
    I would also pull the electrolytic cap (its the largest one on the board) and install it on the other, as there is probably space for it.
    if not, solder it to the battery cables as close as possible to the h bridge.

    you can indeed parallel the two transformers, however this will not double the potential power output, as you will now be limited by the current capacity of the circuit board traces, the fuses.. and the 10 awg wires from the battery. however all of those things can be upgraded as well. it will also practically double the no load losses.

    for an inductor, i used 7 turns of wire on a 25mm diameter core. yes, that's very large, an inch in diameter. the air gap was 2 pieces of paper.

    you can use the ferrite core from a crt flyback transformer, but you would want 4 of those cores, not just one. although even one would help.
     

    Offline darkyputz

    • Newbie
    • Posts: 6
    Thx again...

    The extra cap amd the 8 parts were plan already...
    the second transformer just gave me hope to lower the load per each...so they are not getting as hot...
    Is that so? or just a miss leading thought?
    And i don't want to double the power...just have a straight 1000 watt pure sinus inverter for a long time...and that it should be able to, don't you think?
     

    Offline cte7dsTopic starter

    • Contributor
    • Posts: 47
    • Country: fi
    • hello world
      • my github
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #37 on: March 02, 2016, 04:44:27 pm »
    a worthwhile mod is to place an inductor in series with the low voltage side of the transformer.

    these machines use the leakage inductance of the transformer to filter the pwm and its very lossy.
    my apc1000 when placing an inductor in series with the transformer, it not only reduced the no load loss by 10 watts, but it also reduced the switching losses because it reduced the pwm frequency.

    I used a ferrite core of 25mm diameter and 7 turns of wire, gapped with a couple sheets of paper.
    don't use an iron powder core.. i was actually able to increase power loss by attempting to use two dozen t-106 yellow cores, by slipping them over the wires of the transformer.

    the inductor will also reduce the no load loss when charging the battery.. mine did not use the linear region of the fets, but rather used the h bridge as a boost converter to charge the battery. this requires pwm, and pwm means the transformer gets warm.

    i don't recall what the reduction at the outlet was, but the transformer probably ran 5C cooler.
    I'm pretty bummed I didn't know about this before I put my unit in to service. I still intend to experiment with this mod, if I manage to find another SmartUps for cheap (to serve as a guinea pig).
    Also, I'm curious, does/could this mod affect the buzzing noise that is present during battery charging and standby?

    Someone actually offered me a very clean looking SmartUps 1500 for 50 EUR including shipping (exceptionally cheap imo), but sadly I wasn't able to justify the expenditure at the time (you know, student life in poverty, *long sigh*)

    you can indeed parallel the two transformers, however this will not double the potential power output, as you will now be limited by the current capacity of the circuit board traces, the fuses.. and the 10 awg wires from the battery. however all of those things can be upgraded as well. it will also practically double the no load losses.
    Yeah, I remember seeing parallel transformers in some old APC XL UPS teardown pictures.

    And of course this guy:  8)
    - APC smart UPS from 1000VA to 3000VA with two parallel transformers!

    Would also love to try this one out myself, if I find more than one smart ups.

    for an inductor, i used 7 turns of wire on a 25mm diameter core. yes, that's very large, an inch in diameter. the air gap was 2 pieces of paper.
    I don't have any experience with magnetics, so I have to ask; What type of core was it? Toroidal, shell, cylindrical, etc..?
    Would something like this work?

    https://www.digikey.com/product-detail/en/epcos-tdk/B64290L618X38/495-3861-ND/1830191
    And perhaps wrap some Kapton tape around it, to provide some gap?

    Aaanyways...
    The reason why I set out to revive this old topic was because I realized people on this thread might be interested of the tool I created for logging and graphing my smart ups statistics (via apcupsd / apcaccess)...
    Just a simple Bash script leveraging rrdtool to generate a HTML gallery of graphs like this:

    "1000 hours in rrdtool-graph... I'm almost not kidding"
    The HTML has thumbnails to select different resolutions and time periods.

    The code with some "instructions" can be found at: https://github.com/fld/graph-apcupsd
    I know it's quite a crude solution, but so far it has worked for me pretty good!

    Enjoy! ..or don't...or make something better!
    « Last Edit: March 02, 2016, 04:54:46 pm by cte7ds »
     

    Offline johansen

    • Frequent Contributor
    • **
    • Posts: 977
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #38 on: March 03, 2016, 12:46:10 am »
    Quote
    Also, I'm curious, does/could this mod affect the buzzing noise that is present during battery charging and standby?

    It should reduce the noise considerably if that noise is from the carrier frequency of the inverter. if you have a high line voltage your transformer may be slightly saturated and that will make both noise and heat.
     

    Offline marcone

    • Contributor
    • Posts: 11
    • Country: ro
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #39 on: January 06, 2017, 11:56:44 am »
    I got to do some modding on a SUA1500i recently so I will share my experience.
    I needed a power back-up system capable of delivering 1300W at peak load to power a pellet heating furnace.
    Going for a 2200VA unit and 4 batteries was out of the budget so I decided to go with the 1500VA unit and 2x65Ah deep cycle batteries.

    In modding the unit I:
    *ripped out original battery wiring, Anderson connector and battery switch out back and replaced all that with 2x1m 10mm2 cables straight from the board to the battery.
    * got rid of the network card box which was impeding air flow in the case
    * added a 3rd MOSFET to each arm of the H bridge and an additional 3300uF/50Vcap
    *using the SMART protocol (Serial port out back + APC serial cable) I adjusted charging voltage to 27.2V and adjusted current sensing for 1300W max.

    Unit works just fine, no extra heat but it's not outputting the 1300W continuously, most of the time the load is around 300W.
    If I would go for more than 1000W continuously, I would add a second transformer where batteries normally sit, 2200VA units use 2x24 transformers in series.

    Great sources of information for this project were https://www.youtube.com/user/FFcossag and https://www.youtube.com/user/knurlgnar24
    « Last Edit: January 06, 2017, 11:59:57 am by marcone »
     

    Offline BradC

    • Super Contributor
    • ***
    • Posts: 2104
    • Country: au
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #40 on: January 06, 2017, 01:28:44 pm »
    Late to the party, but some info :

    - The SU(A)1400/1500 XL are 48V thus they have 2 transformers. The non XL are 24V (as you found out). The XL are basically a de-rated 2200.

    - The reason you couldn't calibrate the runtime correctly using the battery constant was you need to tell the unit it has extra batteries. They all use standard firmware, so even the non-XL units can be tweaked up with extra batteries. If you do a quick calculation of the estimated battery life and correct for the puekerts constant for your batteries, you crank the ext-battery count up until your estimated runtime exceeds the calculated runtime, then decrease the battery constant until it gets close. Then when you perform a runtime cal it'll be pretty close.

    - Be careful that your heatsink compound does not increase the resistance between the FET tab and the heatsink. It's a major current route.

    - The main DC filter cap is subjected to a high stress environment (seriously), which is why it is physically so big for the capacity/voltage. They are an extremely high quality cap and one part I've never seen fail yet.

    - Only the SUA models (generation 4 and above) can alter the battery voltage in software. The earlier models require hardware tweaking of a resistive divider. The SUA uses a PWM output from the micro to sway the 5V reference voltage on the charger. They do this to provide temperature compensated battery charging which none of the previous generations had. It actually works passably, but I still wind about 800mV off the charge voltage by default. You are changing the voltage the charger actually provides, not the voltage the UPS reads, so the UPS still thinks it's hitting its target voltage.

    - The MOSFETS don't operate in linear mode while charging the battery. They momentarily short the low side of the transformer H-Bridge to allow a build up of magnetic field which then discharges via the fets body diodes to charge the battery. It's a crude form of boost converter as the raw transformed low voltage is not high enough to charge the battery. This explains the noise from the transformer while charging.

    - 48V models can be converted from 110 to 230 and back by re configuring the transformer cabling and changing some MOV's / Caps on the board. It's a lot of work though.

    - You don't need to change the CT burden resistor to alter the power calibration as long as you stay within about 80% of the original power rating (in VA). There is enough headroom in the software calibration to cope. Anything more than that and you are asking for a transformer failure anyway.

    - The biggest difference in the XL units is the cooling. The also have a thermal cut-out on the heatsink that breaks the main logic supply rail to the MOSFETS & driver if it overheats, so it goes dead.

    - Battery cabling resistance is critical on these units. They have a set of constants (4,5 & 6) that are calibrated for each model and characterize the battery system. These allow for instantaneous drop when switching to load, volt drop in the cabling (to properly estimate the actual battery voltage) and battery internal resistance. Increasing the battery cable resistance does bad things to the correct runtime estimation.

    - A good result for external batteries is to put a new set of internals in and supplement them with externals. This keeps the internal resistance within original spec.

    Common failure points on pre-SUA models :
    - On 48V models the bootstrap caps on the FET drivers
    - On 48V models the 1K resistors in the FET drivers
    - The caps on the -8V supply
    - The caps on the 24V logic SMPS
    - The battery voltage divider drifts

    I've not had SUA models fail to a point they need repair yet.

    There are schematics out there if you look hard enough for the 2, 3 & 3.5G units (SU). I have found one schematic for an early SUA1500 which gives enough detail on the altered FET drivers & PWM reference tweak on the charger ASIC to be useful on the remainder of the SUA units. I believe the later SMT & SMX units are basically an SUA with a pretty front panel and a protocol translator to prevent accessing the internal UPS-Link interface. There are Youtube videos on how to bypass that if you look around.
     

    Offline marcone

    • Contributor
    • Posts: 11
    • Country: ro
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #41 on: January 06, 2017, 03:58:15 pm »
    Thanks, that reminds me of another software tweak I performed, telling the UPS that it had 3 external battery packs in order to make up for the 65Ah batteries as opposed to the original 18Ah.
     

    Offline BradC

    • Super Contributor
    • ***
    • Posts: 2104
    • Country: au
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #42 on: January 06, 2017, 10:24:01 pm »
    Thanks, that reminds me of another software tweak I performed, telling the UPS that it had 3 external battery packs in order to make up for the 65Ah batteries as opposed to the original 18Ah.

    It's worth playing with. Some ups count 18ah as an external pack, and some count 36ah so you have to experiment to figure out which does what. I have standard internals + 55ah of external on my current units.
     

    Offline NJM

    • Newbie
    • Posts: 5
    • Country: it
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #43 on: April 04, 2017, 10:59:08 am »
    Hi , i'm new in this forum, i have a APC 3000 48v to 220 50hz 60hz , i would like use it as a power inverter. I Have a few questions; is it possible work  on 110ac, can you tell me what i have to do, toimprove the runtime,i do not need more then 500w continous. But i need to use it for long period. I was thinking in easy modification,becouse i' completely  uneexperint in elettronics. Thanks in advance. NJM
     

    Offline johansen

    • Frequent Contributor
    • **
    • Posts: 977
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #44 on: April 04, 2017, 08:20:37 pm »
    Hi , i'm new in this forum, i have a APC 3000 48v to 220 50hz 60hz , i would like use it as a power inverter. I Have a few questions; is it possible work  on 110ac, can you tell me what i have to do, toimprove the runtime,i do not need more then 500w continous. But i need to use it for long period. I was thinking in easy modification,becouse i' completely  uneexperint in elettronics. Thanks in advance. NJM

    the largest parasitic loss in that system may be the transformer.

    as such winding a custom transformer may not be such a bad idea. i would suggest a toroidal transformer but there is one problem. those inverters do not ensure zero dc volts output. they actually saturate the transformer arbitrarily by passing dc current through the low voltage side. but i have not verified this with a 48v system.

    so what i might suggest is you rewind the existing transformer for 110vac output. you will need to wind an additional 220v coil so that the inverter thinks its producing 220v. this coil can be of relatively low capacity.
     

    Offline NJM

    • Newbie
    • Posts: 5
    • Country: it
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #45 on: April 04, 2017, 08:45:09 pm »
    Hi johansen, I did the tests on the transformers using the serial lamp, from black +white i found 110v, i'm not sure but i was looking under the main board, it seens the two transformers are wired in series to, as the primary, my tought is they use just one model of transformers for many models of ups. just changing the wires. I'm not hundred % secure. Another thing, i wired its transformer on Lf main board  24v to 110v and it worked. but i would use them on APC. Thank you NJM
     

    Offline BradC

    • Super Contributor
    • ***
    • Posts: 2104
    • Country: au
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #46 on: April 05, 2017, 07:41:14 am »
    Hi johansen, I did the tests on the transformers using the serial lamp, from black +white i found 110v, i'm not sure but i was looking under the main board, it seens the two transformers are wired in series to, as the primary, my tought is they use just one model of transformers for many models of ups. just changing the wires. I'm not hundred % secure. Another thing, i wired its transformer on Lf main board  24v to 110v and it worked. but i would use them on APC. Thank you NJM

    You don't need to rewind the transformer as you've already found. The larger units use two identical transformers with the HV side in series for 230V and parallel for 110V. You need to re-wire those. The UPS senses mains voltage using small transformers on the PCB, so your unit will be set up for 230V. You'll need to figure out some sort of other arrangement to change the mains output sense voltage to get that to work. It wouldn't surprise me if they were a relatively "off the shelf" PCB mount transformer. If you have some more details on your unit (like generation, PCB configuration) I might be able to assist with some more info. Start with the label on the top of the main control uC. That'll have UPS generation and firmware revision.

    You'll also need to change the divider resistors on the output current transformer. I can't help with the values on a 110V unit, but might be able to come close by examining some of the domestic model schematics. There's a pretty wide range of tolerance on both the voltage and current sense in the software calibration.

    I've certainly never seen a SmartUPS deliberately put *any* DC into the transformer while on inverter. It certainly does when used as a charger as it uses the transformer as the inductor in a boost converter.

    What you are asking is doable, but there are a few little things you need to look out for. It'd be easier to flog that one on E-bay and pick up a 110V unit.
     

    Offline johansen

    • Frequent Contributor
    • **
    • Posts: 977
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #47 on: April 05, 2017, 11:21:21 pm »
    I've certainly never seen a SmartUPS deliberately put *any* DC into the transformer while on inverter. It certainly does when used as a charger as it uses the transformer as the inductor in a boost converter.

    i can post a video of this if you want, but here's a photo of the current flowing through the low voltage side of an apc1500 24v ups.
    i have an inductor in the circuit, and the resistor used to measure the current is 16.7mOhms, so two divisions is one volt and that is 60 amps.
    rms amps should be about 53 amps, the primary is 14 vac nominal, 120vac output into 750 watt resistive load. you can see the dc offset clearly.
    http://johansense.com/bulk/sinewaveinvertermodwiinductor17mohm750wattload.JPG
    in this photo the inductor has been removed.
    http://johansense.com/bulk/sinewaveinvertermodwoinductor17mohm750wattload.JPG



    in this photo is the current through the 14vac side of the transformer, at no load, running as an inverter. there is a small inductor in series with the resistor.
    http://johansense.com/bulk/sinewaveinvertermodwiinductor17mohm.JPG
    the transformer is fairly saturated at this point.


    the dc voltage offset varies with time, it is not random though. i have verified this across at least 4 smart APC ups's.

    the cuprit is the custom asic or whatever it is, a 16 pin dip that handles the pwm generation I believe it also has a pll in it to bring the output back in sync with the ac grid before switching back to ac mains.

    I have tried inserting a sufficiently large capacitor in series with the transformer to cancel out that dc offset but it doesn't work with a capacitor in series (perhaps i need a few farads rather than 100,000uf)

    One method to fix this might be to measure the dc offset and then add it to the output voltage sense circuitry before it gets back to the chip that generates the pwm.


    when the h bridge is shorted out to use the transformer as a boost converter, the volt amps are not equal on both sides of the sine wave either. but its not as bad as when its used as an inverter.
    « Last Edit: April 05, 2017, 11:26:24 pm by johansen »
     

    Offline BradC

    • Super Contributor
    • ***
    • Posts: 2104
    • Country: au
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #48 on: April 06, 2017, 01:02:59 am »

    i can post a video of this if you want, but here's a photo of the current flowing through the low voltage side of an apc1500 24v ups.
    i have an inductor in the circuit, and the resistor used to measure the current is 16.7mOhms, so two divisions is one volt and that is 60 amps.
    rms amps should be about 53 amps, the primary is 14 vac nominal, 120vac output into 750 watt resistive load. you can see the dc offset clearly.
    http://johansense.com/bulk/sinewaveinvertermodwiinductor17mohm750wattload.JPG
    in this photo the inductor has been removed.
    http://johansense.com/bulk/sinewaveinvertermodwoinductor17mohm750wattload.JPG

    the dc voltage offset varies with time, it is not random though. i have verified this across at least 4 smart APC ups's.

    Novel. Well, if it's on the 24V models it'll be the same on the 48V models. I must say it's not behaviour I've seen, but then I've never looked at it from this perspective. If it's not random, *and* it varies with time, have you figured out what the correlation is?

    the cuprit is the custom asic or whatever it is, a 16 pin dip that handles the pwm generation I believe it also has a pll in it to bring the output back in sync with the ac grid before switching back to ac mains.

    Not quite. There are two interlinked asics. The first manages the charging and mains phase comparison, the second manages the PWM. They don't have any internal power smarts and receive their sine wave reference from a DAC connected to the main CPU. The CPU generates the sine wave reference, and slews it over 4 seconds to ensure it's in sync before they switch back to mains. Aside from the funky boost charging, pretty much all the smarts in these units is in firmware. It also changes the inverter frequency depending on the incoming mains. It'd be interesting to compare that generated sine to see if there is a DC offset in there relative to what hits the H-Bridge.

    In *theory* (as documented in some internal APC technical docs that were found on a Russian web page), these units all default to 60Hz regardless of voltage setup. So a cold start from battery with no incoming mains should have them generating 60Hz. I've used this feature a lot, but I've never actually tested the generated mains frequency. I'll do that when I next power one up.
     

    Offline NJM

    • Newbie
    • Posts: 5
    • Country: it
    Re: Cool tutorial: Converting an APC UPS into a powerful sinewave inverter
    « Reply #49 on: April 06, 2017, 07:27:34 am »
    Hi the problem is; I d'nt have the that board to connect to my computer,to change configuration, i have the RJ jack. I tried to connect, but when i connect the cable, the ups gos off i'm not sure if can use the software without that pc card. What about to change the dc imput from 48to 24dc? I'm telling that becouse i' making some meccanichcs modification. Mounting another power inverter in the same box.thanks.NJM
     


    Share me

    Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
    Smf