diy solar

diy solar

Building a UPS for my home lab

windham

New Member
Joined
Dec 29, 2019
Messages
4
I am a complete newbie about building battery systems. I have a home lab on an equipment rack of all my home servers (network equipments/NAS) consuming about 400W. I also have a work computer system with 500W max during use. I want to build a LFP battery based UPS to power my entire system for at least an hour in case of any power outage.

I have been watching Will's videos. It occurs to me that something like a MPP Solar's 12V all-in-one system plus a stock battery (like the SOK battery or Battery Born) is a good combination for this purpose. Has anyone built a similar system? Is there any quirks and beginner traps that I need to watch out for?

Appreciate your response in advance!
 
You need about 900W of power, for one hour.
That's a fairly small amount. Presumably infrequent occurrence (at least for an entire hour's use.)
It could be satisfied with a single LA battery, either gel or AGM.
Ignoring Peukert effect, 100 Ah 12V would be drained 75% delivering 900 Wh.
A bit more than that for inverter inefficiency, and size it larger due to the 1C discharge rate.

You could build a DIY lithium battery for this, but a single AGM battery would be an off-the-shelf solution at moderate cost.

UPS - most have some transfer time, and most computer stuff will ride through. An "on-line" UPS is the way to have zero dropouts, if that matters.

Most computer loads have an ugly current draw, rectifying AC onto a capacitor so far from 1.0 PF. Whatever inverter has to be OK with that. Maybe a "modified sine wave" inverter would actually be better than pure sine wave in that case? I'm not sure.
 
Ignoring Peukert effect, 100 Ah 12V would be drained 75% delivering 900 Wh.
A bit more than that for inverter inefficiency, and size it larger due to the 1C discharge rate.
Thanks for the explanation. Most likely I don't need constant 900W, more likely ~700W or so, though I must watch out for the inverter efficiency for sure.

You could build a DIY lithium battery for this, but a single AGM battery would be an off-the-shelf solution at moderate cost.
Looks like the case! Just did some quick research. AGM battery costs much less than LFP and its maximum discharge can be over 1C for some amount of time. Thanks for the tip.

UPS - most have some transfer time, and most computer stuff will ride through. An "on-line" UPS is the way to have zero dropouts, if that matters.
1012LV-MS's datasheet (https://www.mppsolar.com/v3/catalogs/PIP-LV.pdf) shows a UPS mode with <10ms transfer time. While typically a consumer-grade UPS has about 8ms transfer time, (eg., https://www.cyberpowersystems.com/product/ups/smart-app-lcd/or1500lcdrm1u/) the MPP Solar model I refer to seems to be a good fit. Not sure how it would perform though. I suppose it isn't really built for a UPS use case.

Most computer loads have an ugly current draw, rectifying AC onto a capacitor so far from 1.0 PF. Whatever inverter has to be OK with that. Maybe a "modified sine wave" inverter would actually be better than pure sine wave in that case? I'm not sure.
Not sure -- I have no idea about this at all. The all-in-one system I was reading about has pure sine wave output. I should probably do a bit more research.
 
Some of your equipment may prefer pure sine wave.

From Dell:


"Pure sine wave output while on battery provides the highest degree of compatibility for active PFC (power factor corrected) servers and sensitive electronics"


"Good point as Dell does not recommend modified square wave with a number of their workstations. All Dell UPS (500W through 5600W) have a true sine wave on utility (mains) and inverter power."


"The system uses a very efficient Active Power Factor Correction (APFC) power supply. Dell recommends only Universal Power Supplies (UPS) based on Sine Wave output for APFC PSUs, not an approximation of a Sine Wave, Square Wave or quasi-Square Wave (see UPS Technical Specifications). If you have questions, please contact the manufacturer to confirm the output type."

The other issue will be whether a (probably transformerless) pure sine wave inverter will object to driving loads consisting entirely of non-power factor corrected supplies.
I have a VFD for my 2 HP well pump, and when operating off-grid a 10 kW transformerless PV inverter gave an error. I figured that could be due to the poor PF it saw.
 
Some of your equipment may prefer pure sine wave.
Yeah true. I heard that from friends as well.

I have a VFD for my 2 HP well pump, and when operating off-grid a 10 kW transformerless PV inverter gave an error. I figured that could be due to the poor PF it saw.
That's quite crazy. I would imagine 2 HP is just about 2kVA if from a pure sine wave input.
 
Yeah true. I heard that from friends as well.


That's quite crazy. I would imagine 2 HP is just about 2kVA if from a pure sine wave input.

I think the original split-phase motor ran on a 15A 230V breaker, but that occasionally tripped supplying VFD for 3-phase motor so I changed it to 20A. Maybe it wasn't more watts or VA, but with all the current being drawn in a shorter time rather than at least some early and late in the sine wave, I^2R of breaker was a bit too high.

The current draw to recharge capacitor only when ac voltage exceeds it is a funny load. Your computers may look the same, depending on whether power-factor corrected.

Transformer-type pure sine wave inverter might be best and handling both PFC and non-PFC loads. It might run hotter but would attenuate harmonics so not as much seen by the electronics.
 
Just wanna report here that MPP Solar confirms in email that their 12V unit actually has 5ms transfer time in UPS mode. I bought a unit from them directly -- they shipped the unit from Taiwan with Fedex after their Chinese New Year holiday. I received it within 3 days!

I will post some photos here once I have my system set up. :)
 
Back
Top