Author Topic: UPS Supercapacitor Conversion  (Read 24443 times)

0 Members and 1 Guest are viewing this topic.

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26751
  • Country: nl
    • NCT Developments
Re: UPS Supercapacitor Conversion
« Reply #50 on: June 20, 2018, 10:29:40 pm »
How about simply mounting a much bigger capacitor in the PSU of the device? That will probably give a much better return on investment. The problem with capacitors is that their voltage has a relation to their charge so you can only use a very small part of the energy stored before the under voltage protection kicks in. A PSU OTOH has been designed to deal with that.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 16792
  • Country: lv
Re: UPS Supercapacitor Conversion
« Reply #51 on: June 20, 2018, 10:35:34 pm »
I am still scratching my head over WHY this idea seems like it solves anything useful.  :scared:

Big, expensive, and complicated over the option of just using the typical SLA batteries? Is it possible? Sure.

I thought I made my goal clear.    Did you take the time to read any of the literature I linked from name brand companies that are offering it?
I hope your goal won't end up with house in flames.
 

Offline ArthurDent

  • Super Contributor
  • ***
  • Posts: 1193
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #52 on: June 20, 2018, 10:44:06 pm »
I'm curious if some of the posters here have seen any of the videos where a car battery has been replaced with supercaps with success.



Ebay
https://www.ebay.com/p/Gravity-600-Amp-Car-Battery-Capacitor-Gr-600bc/1985231668?iid=392043817345&chn=ps
« Last Edit: June 20, 2018, 10:47:07 pm by ArthurDent »
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 28134
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: UPS Supercapacitor Conversion
« Reply #53 on: June 20, 2018, 10:47:45 pm »
I'm curious if some of the posters here have seen any of the videos where a car battery has been replaced with supercaps with success.
Did you follow the link I posted (now fixed) in reply#48 ?
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #54 on: June 20, 2018, 10:50:55 pm »
I am still scratching my head over WHY this idea seems like it solves anything useful.  :scared:

Big, expensive, and complicated over the option of just using the typical SLA batteries? Is it possible? Sure.

I thought I made my goal clear.    Did you take the time to read any of the literature I linked from name brand companies that are offering it?
I hope your goal won't end up with house in flames.
How about simply mounting a much bigger capacitor in the PSU of the device? That will probably give a much better return on investment. The problem with capacitors is that their voltage has a relation to their charge so you can only use a very small part of the energy stored before the under voltage protection kicks in. A PSU OTOH has been designed to deal with that.

That would be a major undertaking.  But again, to be clear I am looking for peoples experiences with cheap super capacitors.   

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #55 on: June 20, 2018, 10:52:21 pm »
I am still scratching my head over WHY this idea seems like it solves anything useful.  :scared:

Big, expensive, and complicated over the option of just using the typical SLA batteries? Is it possible? Sure.

I thought I made my goal clear.    Did you take the time to read any of the literature I linked from name brand companies that are offering it?
I hope your goal won't end up with house in flames.

None of them have so far but I appreciate your concern.

Offline IDEngineer

  • Super Contributor
  • ***
  • Posts: 1923
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #56 on: June 20, 2018, 10:59:33 pm »
I'll add my limited but very real experience with commercial UPS's.

I once designed and installed redundant server farms that we placed in rental space at colocation facilities all over the western USA. We consumed a LOT of UPS's, mostly APC's expensive (and supposedly high end) SmartUPS rackmount units. This gave me a lot of exposure to UPS's in general and APC UPS's in particular.

The earlier comments were correct about APC charging circuits: They suck. They REALLY suck. They suck so terribly I've run out of superlatives to describe how bad they are. And this is across the entire product line, not just the supposedly high end SmartUPS.

Within 12 months the batteries would have swelled so much that they could not be physically slid out of the rackmount frame. We initially presumed it was low quality batteries so we started researching and buying the very best we could, any chemistry, any price. Didn't matter. No battery could stand up to the electrical abuse that the APC charging circuit dished out. I've lost count of how many times I had to remove a rackmount APC UPS from the rack and take the entire enclosure apart, including multiple internal walls, just so the batteries could be PRIED out and replaced.

Meanwhile, I have many lead-acid and AGM batteries for other purposes that we maintain with Battery Tenders and similar aftermarket intelligent chargers and those batteries last years and years.

I will never again willingly purchase an APC UPS. I've since moved on in my career and don't have need for racks of UPS's so I don't have experience with other brands to share. But APC? Never again. Too many last-minute plane tickets and chasing around remote cities looking for a Batteries Plus and all-nighters disassembling and reassembling entire rackmount enclosures.

I agree with the other respondents who have openly wondered if the battery life is controlled by the UPS's charging circuit. Makes me wonder if one could run a quick experiment by taking an existing UPS, disconnecting its internal charging circuit, and substituting a respected external aftermarket battery charger/conditioner. Use the UPS to generate power, but rely on someone respected to recharge the batteries afterward. Just thinking out loud.
 

Offline ArthurDent

  • Super Contributor
  • ***
  • Posts: 1193
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #57 on: June 20, 2018, 11:05:27 pm »
I'm curious if some of the posters here have seen any of the videos where a car battery has been replaced with supercaps with success.
Did you follow the link I posted (now fixed) in reply#48 ?

Sorry, my remarks were not directed at you but SOME think its black magic.
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #58 on: June 20, 2018, 11:11:29 pm »
I'm curious if some of the posters here have seen any of the videos where a car battery has been replaced with supercaps with success.

I've seen a few of these.  Actually, the cranking time on that video you link is longer than what I normally experience for a line drop out and I would not be surprised if the car did not present a much larger load then I pull off the UPS.   The interesting thing about this video is the caps are mounted under the hood of a car.  That's a much worse environment than I plan to run them in.   

There are a few boxes of used Maxwells on eBay.  I have not looked to see what they are.  May be an option just to run some experiments with. 

Online tautech

  • Super Contributor
  • ***
  • Posts: 28134
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: UPS Supercapacitor Conversion
« Reply #59 on: June 20, 2018, 11:34:17 pm »
I'll add my limited but very real experience with commercial UPS's.

I once designed and installed redundant server farms that we placed in rental space at colocation facilities all over the western USA. We consumed a LOT of UPS's, mostly APC's expensive (and supposedly high end) SmartUPS rackmount units. This gave me a lot of exposure to UPS's in general and APC UPS's in particular.

The earlier comments were correct about APC charging circuits: They suck. They REALLY suck. They suck so terribly I've run out of superlatives to describe how bad they are. And this is across the entire product line, not just the supposedly high end SmartUPS.
:clap:
Some further hunting on this topic today reveals yours and my thoughts about APC charging are widespread.
A interesting link I just stumbled upon from chasing a hint madires gave us earlier:  hidden service menu
http://www.jjoseph.org/notes/apc_smartups_battery_float_voltage

I might have to dig out the APC 1000 I parked up years ago and have another look at it !  :scared:
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline The Soulman

  • Frequent Contributor
  • **
  • Posts: 949
  • Country: nl
  • The sky is the limit!
Re: UPS Supercapacitor Conversion
« Reply #60 on: June 20, 2018, 11:50:17 pm »
Yeah, those apc smart ups's they have set the floating voltage ridiculously high (29V if I remember correctly?) by default,
it can be lowered in some menu.
Had to replace the batteries on a couple of 19" units after 5 years of mostly standby use and maybe a handful of discharges.
The batteries where already completely deformed and swollen and took brute force to get them out.
I think apc's idea on this is that costumers should replace battery's every two or three years, $.
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: UPS Supercapacitor Conversion
« Reply #61 on: June 21, 2018, 12:40:02 am »
i have been dumping price and data on a spreadsheet for sometime, so i could meddle with some kWh capacity and voltage, and sort it to tell me some ranking data.

so i sized all units into some 14v modules with minimum 1Wh. i could see that supercaps that could nearly catch the tails of lead acids on my spreadsheet are something like 8.4v MAL219690118E3 vishay, this is around $25/Wh/year. the nearest lead acid (7.2AH) is at around $20/Wh/year. a 38AH is at around $9/Wh/year.

a 2.7v 400F cap (XV3560-2R7407) to compare, stands at around $45/kWh/year. i think this similar to what the youtube tripplite mod used 6x 2.7v? 67 farads?
if to compare with this tripplite mod in terms of similar amount of energy (1 Wh ?), then MAL219690117E3 or MAL219690118E3 might be useful for long lifespan.

but how much power outage can it bridge out of a "1 Wh battery"? if the usable UPS voltage is 14v to 10v, the UPS can only use about 28% of that 1 Wh from a supercapacitor. which means, bridge for about 3seconds for a 350w load? did i math this correctly?
or maybe, there is a statistics of brownouts etc to help size this?

on the other hand, the sustained current output of a "small" supercap stack (this vishay or string of 400F), i dont think could match that of a bulk leadacid. if i am seeing correctly, this (either the vishay or the 400F version) would need maybe alot more in parallel like 4 or 5?
by itself, the caps overall lifespan is around 23x more than the SLA, for about at least 2x the cost (just for 1 Wh, or comparatively 1/15 the capacity)
just some rough estimates, if the capacitor is around 50C/W in thermal resistance. a load dump for 3s could mean the end of life for the caps. 1 single use hmmm?


in any case, very interesting to dive inside some specs to find some numbers
« Last Edit: June 21, 2018, 01:17:08 am by 3roomlab »
 

Online Circlotron

  • Super Contributor
  • ***
  • Posts: 3166
  • Country: au
Re: UPS Supercapacitor Conversion
« Reply #62 on: June 21, 2018, 12:54:35 am »
Quote
As I mentioned in my first post, every two or three years I seem to be replacing my lead acids.
Either you buy crap batteries or charging circuit in your UPS sucks. Good long life batteries easily last 5+ years in a decent UPS. BTW if by chance you buy original batteries from APC, usually they are a cheap Chinese crap with very expensive APC sticker slapped on top.
Something nobody has mentioned yet is the Achilles heel of lead acid batteries - TEMPERATURE. The ups next to my work computer has the battery at easily 30deg C. I’ll bring in the temp probe tomorrow and see. The 13.65V charging voltage is only relevant at 20 deg C and should decline with increasing temperature.

The optimum operating temperature for the lead-acid battery is 25°C (77°F). As a guideline, every 8°C (15°F) rise in temperature will cut the battery life in half. VRLA, which would last for 10 years at 25°C (77°F), will only be good for 5 years if operated at 33°C (95°F). Theoretically the same battery would endure a little more than one year at a desert temperature of 42°C (107°F).

http://prod.sandia.gov/techlib/access-control.cgi/2004/043149.pdf
Okay... got into my little tin shed office this morning and the ambient temp was 8.2 deg C. UPS has a single 12V 7AH SLA. Underside of the battery was 18 deg, side was 25 deg. Top was 33 deg. Battery housing surface next to the top of the battery was 37 deg C. Remember, this is in 8 deg C ambient so add another 15-25 deg to those figures. The battery was 2-3/4 years old and with a 2R2 ohm load dropped to 10.5 volts immediately. Also had some corrosion around one of the terminals, so basically it was way dead.

Edit -> After replacing the battery I've now stood the unit up vertically so that the battery is at the bottom and hopefully the internal heat from the transformer will now exit the top.
« Last Edit: June 21, 2018, 12:57:09 am by Circlotron »
 

Online BradC

  • Super Contributor
  • ***
  • Posts: 2104
  • Country: au
Re: UPS Supercapacitor Conversion
« Reply #63 on: June 21, 2018, 02:22:09 am »

A interesting link I just stumbled upon from chasing a hint madires gave us earlier:  hidden service menu
http://www.jjoseph.org/notes/apc_smartups_battery_float_voltage

It's not a "hidden service menu". It's a set of commands for factory calibration of the units that isn't documented unless you have the factory calibration notes. That page (and most of the net) is quite unclear about the float voltage on these units, so here are some real take home facts.

The APC UPS always target the upper end of the float voltage range for the batteries to get maximum runtime at the expense of ultimate battery life. The plethora of reports about APC SmartUPS cooking batteries isn't because APC are deliberately doing anything to cause this, it's because a combination of tolerances drift significantly as the units age and the divider that measures the battery voltage tends to drift low, forcing the charging ASIC to increase the voltage. It seems to happen to most of them, but then those 3G units are all >15 years old now. An occasional service isn't out of the question for a unit that has > 100,000 hours on the clock.

The SmartUPS family comes in 5 basic generations at this point. The most commonly found are G3 (beige SU series) and G4 (black SUA series). Some of the later G3 variants were black also, but they're kind of a half-breed (mostly G3 with some G4ish improvements but not the G4 charging smarts).

Generation 1&2 were configured with trimpots and had RS232 status line reporting. I've never seen one in the flesh, only the service manuals.

Generation 3 (SU) has the ability to report and alter the reported battery voltage in firmware. They have no ability to alter the *real* battery voltage in firmware other than to apply a boost during recharge (and that boost is an on/off thing only).  To alter the charge voltage you must alter the resistive divider at the input to the charger ASIC. There is no way to alter the float voltage in firmware. All you are doing by "re-calibrating" the UPS is to change the voltage the UPS *thinks* the battery is at.

G4 -> (SUA) has the ability to alter the reported battery voltage in firmware. The processor uses the reported battery voltage to alter a PWM value to provide trimming to the charger ASIC which does alter the *actual* battery float voltage. This ability is primarily so the UPS can apply temperature compensation to the float voltage, but it has the side affect of being able to change the float voltage by altering what the UPS *thinks* the battery voltage is. As you change the battery voltage constant, the UPS alters the charge voltage to keep it where it thinks it should be based on the battery temperature (NTC hanging in the back of the battery box). There is no way of changing the target setpoint and changing the divider means the UPS isn't accurately reporting the battery voltage, but at the end of the day you are changing the float voltage.

The G4 also encompasses the new G5 (SMX/SMT/...) as the actual UPS back end is relatively the same. There is a second UI processor on those that handles the front panel and the Microlink interface, but if you dig in you'll find that processor is polling an essentially unmodified G4 back end using the same old UPSLink protocol. So if you get between those two you can tweak the UPS like you could with the older generations.

Each generation has a series of sub-generations. I've seen at least 3 variants of the SU series, the latest of which was entirely SMD and with a dedicated FET driver (the original through hole 3G SU series had a discrete FET driver that was awesome at cooking the board on the 48V models). No matter which variant, they all rely on a fixed resistive divider to set the battery voltage.

The reality with the UPS units is the batteries live inside the box with the electronics. That makes the batteries much warmer than they would be in a separate enclosure and *that* is the major compromising factor in their life. I've got a couple of units modified for external batteries, and that takes the battery life from ~3 years to ~5 years. No change in voltages or battery management strategy.
I also modify the SU series to run the fans 24/7 at half speed. That drops the internal battery temperature by about 5C off the bat. The SUA series and up do that out of the box.

As for quality, things certainly started to dive after the Schneider buyout. I don't think there is any dispute of that.
 
The following users thanked this post: 3roomlab, mmagin, tautech, RossS, The Soulman

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #64 on: June 21, 2018, 10:58:49 am »
I think the first UPS I used was made by Sola and used ni-cads. Maybe.  That was early 80s.   

The first APC I bought for home has turned yellow over the years.  Its still works fine and is use 24/7.  This unit stays in a part of the house where it is cooler and has the battery that I mentioned that's the oldest. It looks like these:
http://www.mysynergyups.com/m7/2%20-%20BK600--two-apc-back-ups-600.html

I bought a newer APC that was gray/black and rated for higher power.  This unit developed a problem after several years and I junked it.  I had some no name brand, space age name.  It could not do the one thing it was sold for.   Simple transients could cause a disruption in the output.  After the batteries failed, I junked it. 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #65 on: June 22, 2018, 10:38:57 pm »
I've added an SSR to my AC line power breakout box to do line dropout tests.  This will give me an easy way to test the UPS.   

I put together a small 70000uF capacitor bank w/ alum elec.   The UPS has no problem charging them from a short.   You can see the holdup time in the attached plots change.

Trace1: Vin to the UPS
Trace2: Current through the UPS
Trace3: SSR control Signal
Trace4: Voltage across load

I'll start but looking at a few low cost caps and see how they perform.   
 
The following users thanked this post: 3roomlab

Online tautech

  • Super Contributor
  • ***
  • Posts: 28134
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: UPS Supercapacitor Conversion
« Reply #66 on: June 22, 2018, 10:45:45 pm »
New channel in the making Joe ?
Watching with interest.  ;)
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: UPS Supercapacitor Conversion
« Reply #67 on: June 22, 2018, 11:25:59 pm »
im not sure if i interprete correctly, so the 0.7F cap allows a mitigation of around 0.2ms to 0.25ms? we are subtracting the difference from the "no battery"? but this is without load right?
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #68 on: June 22, 2018, 11:50:43 pm »
There's not much to interpret.  It will be a while before I can do anything meaningful with it.

New channel in the making Joe ?
Watching with interest.  ;)

I wasn't planning on making a video about it but may post some data once things start coming together.   It may be a week or so. 
 
The following users thanked this post: 3roomlab

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #69 on: June 24, 2018, 05:23:12 pm »
About half way in he talks about the parts being counterfeits.


Offline ArthurDent

  • Super Contributor
  • ***
  • Posts: 1193
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #70 on: June 24, 2018, 06:27:53 pm »
The results shown in the video in post #69 pretty much agrees with what I found with my limited testing. Of the 6 'Green-Cap' caps in the bank I bought, 2 were almost useless but 4 of them looked quite good. The video may be correct that these caps are rejects for any number of reasons and the sellers aren't too fussy about how good the caps are that they use.  This would explain why I found 2 out of 6 to be defective.

I didn't check for capacity because the bank of 6 would give me far more power-up time than I would need but I would need them to hold their charge for a reasonable length of time. After I replace the 2 defective caps with good ones I plan to solder a resistor/LED across the equalizing resistors to give me a visual indication that each protective circuits is working properly.

I knew when I bought these cheap caps that there might be problems but I wanted to give them a try to see if I could use supercaps in this application. If you're looking to buy a bank of supercaps that you can just drop in a circuit and have them work with no problems, I wouldn't recommend these at all.
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #71 on: June 24, 2018, 08:33:31 pm »
Some of those prices are so low, they just draw you in...  :-DD   

In the case of the UPS, it's always on so leakage shouldn't cause much of a problem out side the added heat.

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #72 on: June 28, 2018, 01:46:22 am »
Soldered up the bank and cycled it several times.  Installed into the UPS.  Showing the input voltage/current, trigger and output voltage with 300W resistive load.

105 is showing a 100ms dropout with the same load.   Better than the 75000uF caps anyway.   







Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #73 on: June 28, 2018, 01:48:01 am »
One second dropout, 300W resistive load.   

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: UPS Supercapacitor Conversion
« Reply #74 on: June 28, 2018, 01:57:45 am »
I am guessing with a 500W load, if it could ride out 5 seconds it would cover the vast majority of the dropouts we have.  Most of the time, I am using about 200W and I would guess the dropouts are around a second.  It's enough to have the lights flash and upset any tests I am running. 

Looks like it can hold the 300W for roughly 17 seconds.  In the ballpark anyway.   After reading the datasheets,  I had decided to run the caps well below their max operating voltage.   I think the next step is just button the thing up and see what happens over the next 5 years.  I will program the UPS to run a test on the bank every week or so.   


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf