Smart-UPS 700 Load% Calibration

[BillP]

First post here so please go easy. 🙂

I've just recently got into the world of Smart UPS 'tweaking' after acquiring two (2) Smart UPS 700 (SU700INT), practically from the trash bin. I cleaned them up (both have suffered severe battery leaks) and fixed the corroded battery connectors, I gave them new batteries and now they're back from the dead. 

Now, it wasn't that smooth, to be honest. I had to play around with the battery calibration register ('0') so the units will report a more accurate runtime duration with the new batteries. I also adjusted the float voltage using the 'B' register so they'll read 27.33V. They were previously set to 27.60 and 27.73, respectively. 

One last issue I see is that, both identical units are reporting different Load % even when I applied the exact same load in both of them. I used three 100W bulbs, which reads approx. 318W at 230V with a power factor of 0.99-1.00 on the power meter.

So, with the PF=1 or unity, VA in this case is equals to W. Doing the math, 318W / 700VA yields 45.4% Load. On UPS1, the Load reported is 53% while UPS2 reported it as 77%.

I then played around with the 'P' register to get the Load % down to 49.8% (I gave it about +4% buffer from the 45.4%).

The question is, is this at least a valid way of calibrating the Load % offset?

I am of course, referring to information found here: http://networkupstools.org/protocols/apcsmart.html

Thanks in advance!

[BillP · Accepted solution]

Nobody?

Anyway, I came to a realization that the Smart UPS 700 is rated for 450W. So I re-calibrated the Load% to use 450W (as opposed to 700VA) as the max load reference point.

Assuming now that 450W is 100% load, using the same test load I used before (318W at 230V), I calculated the Load% using the following formula: (Measured Watts / System Max Watts) x 100 = Load%

So,

(318W / 450W) x 100 = 70.6%. That means 318W is taking up approx. 70% of the UPS load capacity.

With the load still connected, I dialed in 70.8% in the 'B' register and now, the Load% and LED indicators properly reflect the actual load.

I understand that this calibration is only as good as the power meter's accuracy (I used a Kill-a-Watt clone), but I'm fairly confident that it is 'good enough', given that I used three (3) 100W incandescent bulbs, the measurements were close enough.