OLD CS1's Blog

Did you check for any resistors embedded in the "sense" wire that might be covered by heat shrink in the harness of the BP24BPG?

from viewing other UPS mods on the web, this would be something to look for.

I am trying to do a similar mod, but instead i want to connect 2 large AGM deep cycle batteries to the BR1500G, while getting the unit to display the correct runtime with the added batteries...

do you have any photos of the BP24BPG harness in detail?

Edited January 22, 2021 by Peter W

On 1/22/2021 at 10:41 AM, Peter W said:

Did you check for any resistors embedded in the "sense" wire that might be covered by heat shrink in the harness of the BP24BPG?

from viewing other UPS mods on the web, this would be something to look for.

I am trying to do a similar mod, but instead i want to connect 2 large AGM deep cycle batteries to the BR1500G, while getting the unit to display the correct runtime with the added batteries...

do you have any photos of the BP24BPG harness in detail?

Yes.  I detected no resistor in the sense wire of the BP24BPG.  I do not have any additional photos of the external battery pack harness, but I can do some if needed.

 

I expect that the unit will display the correct run-time when using the deep-cycle batteries.  With no special configuration, the UPS is able to calculate the difference between the two-battery and four-battery packs.  The former it is not designed to work with.

 

Sorry for such a late reply -- I do not recall getting a notification of your comment.  Hope all has gone well with your project.

 

Oh, BTW, it is not a shrink-wrapped wiring harness but rather it is molded.

I just stumbled across this blog post; it was a super interesting read!

 

Our org almost exclusively uses the rack-mount flavor of the APC UPS chassis, most of them are SMT1500RM2U (or similar).  Prior to me transitioning to a different group internally, we had started the push to the li/on models versus the lead-acid ones.

@digdugnate I use the SM1500RM2U at the data center for my servers, and two of the SMT1500RM2U like yours on my home office rack.  The newer models with the LCD panel use a different power scheme internally which does not keep the main transformer energized all of the time.  This cuts heat output enough to change the ambient temperature in my office from around 82°F to 78°F -- the office used to stay around 6-8° hotter than the house, but now only about 2-3°.

 

I looked at the LiIon stuff.  They seem kind-of pricey for the same power output as the SLAs.  How have they been for you guys?

42 minutes ago, OLD CS1 said:

@digdugnate I use the SM1500RM2U at the data center for my servers, and two of the SMT1500RM2U like yours on my home office rack.  The newer models with the LCD panel use a different power scheme internally which does not keep the main transformer energized all of the time.  This cuts heat output enough to change the ambient temperature in my office from around 82°F to 78°F -- the office used to stay around 6-8° hotter than the house, but now only about 2-3°.

 

I looked at the LiIon stuff.  They seem kind-of pricey for the same power output as the SLAs.  How have they been for you guys?

The cost is a fair bit higher than the LAs right now but doing the analysis it looks like we'd come out ahead over five to seven years of the lifecycle of the UPS. 

 

I LOVE the li/ons from an installation standpoint because they're shorter in depth (which is great for us because not all our IDF cabinets are standardized sizes) and weigh significantly less.  We've had them in production only about six to eight months but so far they seem to hold up pretty well to temperature extremes.

 

The batteries didn't used to be user-replaceable but now they're considered CRUs.  Even though APC will try to steer you towards the AP9640 network management cards, you can still use the old AP9630/31 if you were so inclined.  (we manage all our UPSes inhouse instead of using APC's software).

 

I managed our UPSes for our org for several years to standardize installs and configs but was more than happy to pass the baton to someone else.  I still do some small work with ours.

@digdugnate Yeah, I still use 9630s with SNMP and other monitoring utils (apcd and NUT.)  I am not a big fan of the outside system monitoring, especially after I spent so much time developing my own in-house resources.  They are not perfect, but they work well enough for me.

 

I can see the shallow chassis being a benefit in my environments, as well as the decreased weight.  The SLA packs are good for about three to five years, usually on the high end for me as the UPSes are only there to keep the rack going until the building generator spins up.  I would like to see how long, in practice, the LiIon packs last.  It costs me about $140 to replace the batteries in one unit, so every three years or so that is not terrible, but if I can get longer service from the lithium batteries for a comparable price, it might be worth it.  Possibly at home, as well, to help keep the heat down.

expected battery life for li/ons is five to ten years, according to their data sheets, compared to four to six years for lead/acid. 

 

from my notes:

 

Difference in battery life from lead/acid to li/on - 4 years greater for li/on

Difference in chassis weight from lead/acid to li/on - 20-30 lbs lighter for li/on

Difference in maximum depth from lead/acid to li/on - 5-6 inches shorter for li/on

Total cost for lead/acid battery replacement - $542.14

Total cost for lead/acid chassis replacement - $1637.15

Difference in cost for chassis lead/acid to li/on - $559.99

7 minutes ago, digdugnate said:

Total cost for lead/acid chassis replacement - $1637.15

Oh, they are having a giggle with this.  I am running the same chassis I bought 10 to 15 years ago   I guess that means replacing the SLA unit with an LiIon?

 

Probably worth having another look.  I imagine the heat output is even less, I might be able to bring my office in-line with the rest of the house.  Two degrees hotter in practice, one degree on average (night temp at 70°F kind-of skews the average,) is worth the electrical/energy costs alone.

 

(Blue = house, green = office)

 

I appreciate the info.  I will spend some time with my vendor in the coming weeks.

it does.   

 

The li/ons don't put out diddly for heat, near as I can tell, either.  Happy to help!  Funny that we got together on AA of all places to talk about UPSes, lol.

One issue I've seen with SLA-form-factor 12V LiFEP04 packs purchased as a replacement for original lead-acid batteries is that the old SLAs (typically rated at 12V @ 9AH) can provide a lot more current than 9A.  However the LiFeP04 batteries have a battery management system (BMS) to protect the cells from over/under charge and over current.  They may only provid 11-12A of maxiumum current (compared to *WAY MORE* with a old-fashioned lead-acid battery.  The observed behavior is that a UPS with a BMS-based Lithium pack may fall quite a bit short of being able to deliver the full load that the UPS was able to with the SLAs.  A big UPS might offer 600W continuous load capability.  At the internal 24V battery terminals thats >25A  (let's ignore conversion losses, it will be even higher with them).  You aren't getting 25A out of a LiFeP04 SLA-form-factor replacement battery from Amazon, the UPS will just shut off without warning.

They must have changed something since yours and mine. Just tried to rig up 24 VDC worth of AGM gel cells to the external jack using spade lugs, I connected to the bottom terminal, negative, when I went to attach the sense line (jumpered to the negative lead), sparks! Luckily it was quick enough it didnt seem to have done any damage to the UPS.

Any ideas?

On 6/9/2024 at 6:57 PM, K5KTF said:

They must have changed something since yours and mine. Just tried to rig up 24 VDC worth of AGM gel cells to the external jack using spade lugs, I connected to the bottom terminal, negative, when I went to attach the sense line (jumpered to the negative lead), sparks! Luckily it was quick enough it didnt seem to have done any damage to the UPS.

Any ideas?

No idea.  Same model UPS as mine?  I suppose it is possible Schneider made some changes, but that would make new UPSes incompatible with the old battery packs.  Is there any voltage on the sense connection in the UPS port?

Hello, I have the same UPS (APC BR1500g) and some time ago I was searching the internet for information about the external battery pack connector and I came across your post. It was really interesting and now that I had to put in new batteries I decided to risk doing something similar, putting two batteries in the normal way inside and two outside (all exactly the same and new). Several things I noticed. First, the rear connector is capable of using the external batteries without the internal ones being inserted and without the need for the upper cable (i.e. the yellow one that acts as a sensor), but when you also put the internal ones in it only calculates the time as if it only had 2 (a pack), it is only when you connect the sensor cable that it beeps and increases the time (also without the sensor cable the fan does not turn on to charge the batteries, but it does when it is connected as you described). I imagine that even if it does not show the time it should last twice as long and not as you report, but oh well. Another thing I noticed is that when I connect the sensor cable the time increases but almost 4.5 times (remember that I only put two other 12v 9ah batteries) and in a false way because evidently it does not last that amount. And my question is. Could you verify if the time that it told you both with the original pack and with the modified one was correct (or at least close to it). I am thinking of doing a calibration with the modified pack connected (with the + 30% charge etc.) and see if it improves. If you have any advice I would be happy to receive it.

@MrCas2 Those are interesting results.  I can confirm that the time my unit calculated with the two external batteries, connected sensor, and internal batteries was indeed correct.  Since doing this I have had numerous power outages due to hurricanes, and a couple which were due to local environmental anomalies (in particular, a large Mylar unicorn balloon escaped its handler and shorted out two distribution phases around the corner.)

 

Most recently, I acquired a proper battery pack, which contains four batteries, for the unit running my freezer.  As before, the calculated run-time increased roughly commensurate to the number of batteries: from ~230 minutes to ~332 minutes.

 

I cannot quickly surmise the reasoning behind why your unit, while obviously utilizing the external batteries, made calculation changes once the wire was activated.  Nor why my unit properly calculates run-time with the four-battery external pack versus the two-battery external pack.

 

In all cases, I am using 9Ah batteries.  The newest configuration provides three 24V 9Ah battery groups, externally offering 18Ah, and with the internal batteries, assuming they are in parallel to the external, a total of 27Ah capacity.

 

While I believe these units are capable of a run-time calibration, I am not certain how much these drift or the usefulness of calibration.  My two SUA1500s greatly benefited from the calibration I ran on the recently when I replaced their battery packs, so I would want to trust the same outcome for these.

Thank you very much for your answer. Yes, I have to perform a calibration with the external batteries connected and the sensor cable also connected to see if the time is more precise. Well, with the 4 batteries connected and the sensor cable also connected, the unit without connected equipment shows a time of 999 and without the sensor cable (if I remember correctly) 264, when it should show 528, that is, double since there are two packs. So I think it is a calibration problem. Also, when the sensor cable is connected the fan starts, I think it is because the unit, when sensing the external pack, increases the charging power (which according to the APC website is 23w) in order to charge the external pack without overheating. I will have to check this with a clamp meter, which I do not have at the moment.