Upgrade APC Smart-UPS 1500 (SUA1500i) to LiFePo4

[Hyper.nl]

Hm, I used to have an account at the APC.com forums, but it seems to have been deleted. (Last login around 2018)

I started a small project for myself to upgrade my older APC Smart-UPS 1500 (SUA1500i) from AGM SLA batteries to LiFePO4 batteries with advanced BMS. Normally UPS units that are designed to work with SLA batteries are not compatible with LiFePO4 batteries, some battery manufacturers have designed an advanced BMS (Battery management system) that allow LiFePO4 batteries to be used in any application where SLA batteries are normally used.

I used 12v 30Ah PowerBrick+ batteries from PowerTech systems. These are fairly expensive but they should last a lot longer than traditional SLA batteries. These batteries can have 3000 cycles at 100% DoD or 4500 cycles at 80% DoD.

This guide should work for any Smart-UPS that uses the same batteries and same connector (APC RBC7)

https://youtu.be/tSfOPcbxAqk

This video is not sponsored. I bought the batteries myself.
Using LiFePO4 batteries in APC units made for SLA batteries is NOT supported by APC.

See video description for important information.

[Javik_apc]

I am very interested in these LiFePO4 substitution projects, because lead-acid is just a horrible technology, and APC's battery monitoring and failure detection has been technically deficient forever. A LiFePO4 BMS substitution may not be built to APC's intended usage, but their "intended usage" is terrible also.

At this very moment I have an APC 1500va LED with a dead APC RBC7 battery pack in a small server room, but which the UPS believes there are no problems with the RBC7. If I run a manual self-test it drops the load and the Dell PowerEdge server R630 suddenly turns off!

Nice work there APC, I thought you fixed this faulty technical implementation and moved over to using resistive simulation loads for self testing, when your products changed from beige to black.

,

So anyway, a specific problem that you may experience with your LiFePO4 is that the onboard BMS needs to pretend it is a lead-acid battery, and slope the output voltage up and down as the LiFePO4 discharges.

You do NOT want a BMS that outputs a steady "full-charge" voltage of 13.6v until the lithium battery is discharged, at which point the BMS suddenly turns off.

This defeats the UPS from being able to determine the state of discharge by monitoring the battery voltage as it discharges. The UPS will have no way of knowing when to enter the "low-battery" phase and send the alert signal that the server should start the automatic shutdown or suspend procedure, to prepare for loss of power.

Ideally the BMS output voltage should slope down to approximate the declining state of charge of a lead-acid battery as the LiFePO4 discharges.

Also it may potentially be better to use a 24 volt LiFePO4 with the same approximate size as the two 18ah lead-acid batteries next to each other. This allows for bigger cooling heatsinks in the combined BMS, fewer BMS components to fail, and also larger lithium cells in the unpartitioned bottom casing.

[Hyper.nl]

The usage of a single 24V LiFePO4 battery is actually quite a nice idea. It may help to prevent any balance issues, if any. I just took 2x 12V replacements to stay the nearest to what the APC expects.

PowerTech Systems also has a 24v variant. The connectors will be at a different place and you need to check if the physical dimensions are right for this unit.
https://www.powertechsystems.eu/wp-content/uploads/specs/PowerBrick_PRO+_24V_32Ah_Lithium-Ion_batter...

[Javik_apc]

Also there is another problem with these LiFePO4 substitutions. You need to carefully examine the specifications for the substitute battery to make sure the onboard BMS can actually output the full capacity of the original lead-acid battery pack, for the maximum load watts for the UPS.

In the UPS I am discussing here: APC SMT1500 - Maximum load 1000 watts, 1500 VA

Lead acid battery, general characteristics:

• Full charge = 13.6v * 2 = 27.2 volts
• Full discharge = 10.5v * 2 = 21.0 volts

As any kind of battery discharges, gradually the amperage increases:

• 1000 watts / 27.2 volts = 36.7 amps
• 1000 watts / 21.0 volts = 47.6 amps

The PowerBrick 12v / 30ah Standard will not work for this, but the PowerBrick Pro+ can do it... nearly.

(Whether the Pro+ actually drops to 21.0v at end of lithium discharge, I don't know.)

https://www.mylithiumbattery.com/shop/12v-lithium-ion-battery-packs/lithium-battery-12v/lithium-ion-...

Standard:

• Maximum continuous current: 30 amps
• 30 second peak current: 45 amps

Pro+:

• Maximum continuous current: 45 amps
• 30 second peak current: 70 amps

Meanwhile this lower-quality 12v LiFePO4 of a similar physical size from ExpertPower is pretty much useless:

https://www.expertpower.us/products/ep1220-20ah

ExpertPower EP1220

• Continuous current: 20 amps
• Permanent discharge: 45 amps for 2 seconds

You would have to use two of the much larger ExpertPower battery that won't physically fit in the APC 1500VA battery compartment.

https://www.expertpower.us/products/ep1250-50ah

ExpertPower EP1250

• Continuous current: 50 amps
• Permanent discharge: 100 amps for 10 seconds

For people that are not buying quality LiFEPO4 with a built-in drop-in 12v/24v BMS, there is the potential to be severely disappointed if their cheap PongHuan Fzdagbu Amazon battery for US $50 fails at the worst possible moment because it can't handle more that 20% the UPS maximum rated load capacity... lol

[alvesjc]

I just converted my very old but great SmartUPS 1400RMXLNET to LiFePo4.

I used the Daly BMS plus active balancer pack, and the battery pack is 16S3P of 32650 6AH cells.

It's working fine but had some fights to start the UPS, it took a while to let me start the UPS.

What I'm noticing here, is that the battery pack is not being fully charged, the cels are a 3.40V aprox and aren't charged anymore.

From what I've understood, the UPS only trickle charges to 13.6V per 12V battery what matches the 54.4V I currently have in the pack.

Is there a way to change this voltage to a bit higher value to have a better usage of the cells capacity?
The BMS cuts at 3.65V per cell, If I could make the UPS charge up to 3.6V would be great, and still before the BMS cuts the pack.

[DZelmer]

There might be look up the smart signaling for the ups you might beable to change the float voltage my logging in with a terminal. Need the smart signaling cable. 2nd option is do a little creative soldering and adding a 20 turn potentiometer resister to the float voltage circuit. I did that to several apc ups's incliluding some matrix units and think the otheres were smart ups 1000 lcd and some pre lcd ones. They seem to run good other than the runtime and battery% are way off. But they run good otherwise.

[alvesjc]

Hi @DZelmer 

I got a schematic from it, can you tell me what resistor you're talking about?
For the cable, is that a standard null modem cable connected in the rear port ?

Thanks for the inputs.

[alvesjc]

Hi @DZelmer 

are you refering to voltage devider formed by R216 and R184 in the schematic ? 
Picture attached.

Thanks