Portable Power Stations

[wingnut771]

I've got a Sonoff powr2 and a Sonoff Elite in series before and after the UPS.

powr2 gives daily kWh and Elite gives hourly kWh

I started the test at 00:00 and when UPS tripped, I compared powr2 and Elite data, which was within 2% of each other

I think you need to add a 3rd device (shunt) on the battery terminal to measure actual battery capacity but it does sound like there a losses. What is the VA and W spec of the inverter?
Have you tried different charging speeds?

Something like this:

 

[soulie]

I realise this is perhaps too technical or 'whatever' for many, but I'm a bit concerned about the efficiency of my trolley UPS.
At this stage I'm not sure if it is to be expected, or if there is a problem with my UPS or one/both of my batteries (2x100Ah 12V Lifepo4).

I did a loadtest on it yesterday and could only get 1.52kWh out of it before it tripped. Load was only 12% of full load.
My batteries are supposed to be 12.7x2x100=2.54kWh.
It took 3.47kWh to fully recharge afterwards without a load.
My rate in Tshwane is about R2.70/kWh.
So it cost me 2.70x3.47=R9.34 to charge and only provided 2.70x1.52=R4.10.
That seems terribly inefficient.

Losses in the UPS should be conversion and cooling losses. Surely these can't be more than 50%?!

Attached is an article about the issue, for those interested

1 amp consumption on some is 10 amps draw from battery .ive measured on inverters from 3amp to 10 amp an a 1 amp load ie 220 watt .some are ridiculous , it all boils down to what mosfets they are using and if the output transformers are copper or aluminiun .some use circuits that turn off the one mosfet before turning on another others cycle continuously

 

[Bundu]

I think you need to add a 3rd device (shunt) on the battery terminal to measure actual battery capacity but it does sound like there a losses. What is the VA and W spec of the inverter?
Have you tried different charging speeds?

Something like this:
View attachment 1478321

nah, I doubt I'll be able to measure much more accurately and I doubt the problem is a measuring error.

Ignoring the charging inefficiency, I'm only getting about 1.5kWh out of 2.5kWh batteries

 

[Bundu]

1 amp consumption on some is 10 amps draw from battery .ive measured on inverters from 3amp to 10 amp an a 1 amp load ie 220 watt .some are ridiculous , it all boils down to what mosfets they are using and if the output transformers are copper or aluminiun .some use circuits that turn off the one mosfet before turning on another others cycle continuously

yes, at no load, fully charged, the UPS still draws about 40W, i.e. nearly 1kWh per day

 

[wingnut771]

nah, I doubt I'll be able to measure much more accurately and I doubt the problem is a measuring error.

Ignoring the charging inefficiency, I'm only getting about 1.5kWh out of 2.5kWh batteries

because you're measuring after the inverter when its already converted dc into ac. If you had a shunt on the battery, you would probably read 2.5kWh pulled out of the battery. Your inverter is probably 60% efficient.

I'm more worried about the recharging. LFP is supposed to be 95% efficient. 3.5kWh seems high. How long does it take to recharge? Are there different amp settings?

How many watts does your inverter take when idle?

 

[leon.davibe]

It would be JUST to support my PC, laptop and LTE router + Mikrotik fibre router

If you want to get a worst case scenario look at the power supplies of all the devices you want to power ie the router would use a 12v 1A/1000ma you multiply the amps with the voltage now you have the watts
ie 12watt

the laptop depends on the specs ie you get laptops that use 35w /65w etc
If you say game i assume it has a discrete gpu those depends on model some can go to 180w and higher the power brick will give you worst case scenario ie multiply the 19v with 7.7A or whatever values yours is ie mine is 146.3watt

Count up all the watts now you have your peak load ie how many watt inverter you need
Try and find an inverter with some headroom

then multiply with the amount of hrs you want back up for then you know how many watts you need in battery power obviously divide by 1000 for kw

Now yes the psu of the pc is a misnomer ie you can't say i have a 650w psu so it's gonna use 650w pwr hour
The hardware matters ie
A 1060 gaming would use about 100w add the cpu etc you would most likely end just under 200w

A 1080 uses about 200w

A 3070ti uses over 325w

These are just the gpu
Add a 12700/12900 and it could add under extreme load add over 200watt more

If you say the gaming pc specs i can guesstimate its usage

Or you can check youtube many test the power usage when reviewing them

 

[leon.davibe]

1 amp consumption on some is 10 amps draw from battery .ive measured on inverters from 3amp to 10 amp an a 1 amp load ie 220 watt .some are ridiculous , it all boils down to what mosfets they are using and if the output transformers are copper or aluminiun .some use circuits that turn off the one mosfet before turning on another others cycle continuously

You gotta keep in mind the amps on the 220 side and on the 12/24/48 isn't the same thing

ie a load of 1amp on 220 means 220Watt

Now pulling 220w from 24volt

Divide 220/24 =9.16amps nothing abnormal is happening amps wise

A watt is a watt on all voltages (there will be some losses with conversion)

An amp isn't an amp on different voltages

 

[soulie]

You gotta keep in mind the amps on the 220 side and on the 12/24/48 isn't the same thing

ie a load of 1amp on 220 means 220Watt

Now pulling 220w from 24volt

Divide 220/24 =9.16amps nothing abnormal is happening amps wise

A watt is a watt on all voltages (there will be some losses with conversion)

An amp isn't an amp on different voltages

that is correct , unfortunately people arent educated properly about what they are fitting and get a rude awakening when they find out about the losses during conversion .they still think 220 volt and not that they are now supplied by a battery which is not a finite source of power and the conversion uses up lots .

 

[leon.davibe]

I realise this is perhaps too technical or 'whatever' for many, but I'm a bit concerned about the efficiency of my trolley UPS.
At this stage I'm not sure if it is to be expected, or if there is a problem with my UPS or one/both of my batteries (2x100Ah 12V Lifepo4).

I did a loadtest on it yesterday and could only get 1.52kWh out of it before it tripped. Load was only 12% of full load.
My batteries are supposed to be 12.7x2x100=2.54kWh.
It took 3.47kWh to fully recharge afterwards without a load.
My rate in Tshwane is about R2.70/kWh.
So it cost me 2.70x3.47=R9.34 to charge and only provided 2.70x1.52=R4.10.
That seems terribly inefficient.

Losses in the UPS should be conversion and cooling losses. Surely these can't be more than 50%?!

Attached is an article about the issue, for those interested

You are 100% sure the batteries was fully charged before commencing the capacity test

ie what was the voltage

Also what was the voltage of the batteries when it switched off

Also were the voltages of the 2 batteries the same when it died

How long have you had it
ie the lithiums connected in parralel will equalise ,

it will take time though if they don't spend a lot of time at the full charge state

ie the balance current is based on voltage difference

Naturally this is quicker at the bottom and top of soc curve

Slower in the middle

 

[leon.davibe]

You are 100% sure the batteries was fully charged before commencing the capacity test

ie what was the voltage

Also what was the voltage of the batteries when it switched off

Also were the voltages of the 2 batteries the same when it died

How long have you had it
ie the lithiums connected in parralel will equalise ,

it will take time though if they don't spend a lot of time at the full charge state

ie the balance current is based on voltage difference

Naturally this is quicker at the bottom and top of soc curve

Slower in the middle

Keep in mind this trolley will take 10.5 hours to fully charge these batteries if you don't have solar
on the 20a setting and almost a day on the 10a

Do you have solar?

the voltage on lifep04 is finicky
ie lets say your inverter indicates its full when voltage is 13v for example those lithiums is nowhere near full
ie you have to have at least 13.4 - 13.8 resting to regard as full depending how long its been resting after charge
and while charging 14-14,6v

 

[Bundu]

because you're measuring after the inverter when its already converted dc into ac. If you had a shunt on the battery, you would probably read 2.5kWh pulled out of the battery. Your inverter is probably 60% efficient.

I'm more worried about the recharging. LFP is supposed to be 95% efficient. 3.5kWh seems high. How long does it take to recharge? Are there different amp settings?

How many watts does your inverter take when idle?

@wingnut771 thanks for your interest in my post - appreciated!
Hope I'm not derailing this thread, but efficiency should be relevant to all these options.

Aaah I see what you mean! DC from batteries to check battery capacity - I'll have to see if I can get one of those meters, or else remove the batteries and do a DC discharge test on them - I've got a meter that can do 10A, so 10 hours discharge. Have tested some Blue Nova batteries with it and they were spot on. (between 97 and 103% of stated capacity)

Yes, the massive charging energy required worries me as well - It took about 4.5 hours with the UPS set at 20A (it has a 10A setting as well) - on the AC side, the power was 0.83kW over the charging time, which again looks weird, as 830/20=41.5 and not 2x12.7? Things don't make sense, but my total consumption has gone UP

when idle, without a load, the UPS draws 40W

 

[leon.davibe]

that is correct , unfortunately people arent educated properly about what they are fitting and get a rude awakening when they find out about the losses during conversion .they still think 220 volt and not that they are now supplied by a battery which is not a finite source of power and the conversion uses up lots .

obviously not all devices are equal ie some would be 80% some 90%

this is where the victrons make up a bit for their cost
but in my opinion not enough to justify the cost

this cost /reward calculation is dependant on your wallet

some pay the premium with a smile

no idea where the sunsynk and other models fit in

 

[Bundu]

You are 100% sure the batteries was fully charged before commencing the capacity test

ie what was the voltage

Also what was the voltage of the batteries when it switched off

Also were the voltages of the 2 batteries the same when it died

How long have you had it
ie the lithiums connected in parralel will equalise ,

it will take time though if they don't spend a lot of time at the full charge state

ie the balance current is based on voltage difference

Naturally this is quicker at the bottom and top of soc curve

Slower in the middle

yes, I assume the batteries were fully charged, as the UPS input consumption was near the output delivery and the voltage on the one battery was a flat curve - see attached

the trolley switched off when the one battery I was monitoring was at 11.2V - I then physically measured the other battery which was 11.6V - not happy with that

I've had the trolley for 2 months, but it hasn't worked hard - batteries are in series, so 24V system

 

[wingnut771]

@wingnut771 thanks for your interest in my post - appreciated!
Hope I'm not derailing this thread, but efficiency should be relevant to all these options.

Aaah I see what you mean! DC from batteries to check battery capacity - I'll have to see if I can get one of those meters, or else remove the batteries and do a DC discharge test on them - I've got a meter that can do 10A, so 10 hours discharge. Have tested some Blue Nova batteries with it and they were spot on. (between 97 and 103% of stated capacity)

Yes, the massive charging energy required worries me as well - It took about 4.5 hours with the UPS set at 20A (it has a 10A setting as well) - on the AC side, the power was 0.83kW over the charging time, which again looks weird, as 830/20=41.5 and not 2x12.7? Things don't make sense, but my total consumption has gone UP

when idle, without a load, the UPS draws 40W

No worries

What batteries are these again? LFP? Are they in series (24V)? Is there a balancer for the series pack?

20A @ 24V is 480W so should take about 5.5 hours to charge a LFP battery from 0% to 100% so charging time sounds about right. Not sure I understand your 0.83kW comment. What happened to the 3.5kWh to recharge comment?

 

[wingnut771]

yes, I assume the batteries were fully charged, as the UPS input consumption was near the output delivery and the voltage on the one battery was a flat curve - see attached

the trolley switched off when the one battery I was monitoring was at 11.2V - I then physically measured the other battery which was 11.6V - not happy with that

I've had the trolley for 2 months, but it hasn't worked hard - batteries are in series, so 24V system

You need one of these:


What's the max voltage you see when charging?

 

[Bundu]

Keep in mind this trolley will take 10.5 hours to fully charge these batteries if you don't have solar
on the 20a setting and almost a day on the 10a

Do you have solar?

the voltage on lifep04 is finicky
ie lets say your inverter indicates its full when voltage is 13v for example those lithiums is nowhere near full
ie you have to have at least 13.4 - 13.8 resting to regard as full depending how long its been resting after charge
and while charging 14-14,6v

no solar

at 20A in series I think 5hrs should be fine - 20A is 20% of 100A, so 5 times rating time

 

[Bundu]

No worries

What batteries are these again? LFP? Are they in series (24V)? Is there a balancer for the series pack?

20A @ 24V is 480W so should take about 5.5 hours to charge a LFP battery from 0% to 100% so charging time sounds about right. Not sure I understand your 0.83kW comment. What happened to the 3.5kWh to recharge comment?

batteries are AOKLY LiFePO4 Battery ALFP-12100(12.8V 100Ah)

the 830W was on the AC side, so ignoring losses (hmmmmmmmmmmmmmmmmmmm......) it could have been 830/(2x14.4)=28A on the DC side, but charge was set to 20A, but who knows? I've got tong tester.... maybe try that?

3.5/0.83=4.2hours

 

[Bundu]

my apologies again to the guys that are simply looking for a UPS....

 

[soulie]

obviously not all devices are equal ie some would be 80% some 90%

this is where the victrons make up a bit for their cost
but in my opinion not enough to justify the cost

this cost /reward calculation is dependant on your wallet

some pay the premium with a smile

no idea where the sunsynk and other models fit in

we have tested some really strange makes and models from brick type to mounted type and most are similar ,i think the more expensive units will use better quality mosfets ,possibly matched sets and decent copper transformers ,faster switching ,and more stable .

 

[leon.davibe]

yes, I assume the batteries were fully charged, as the UPS input consumption was near the output delivery and the voltage on the one battery was a flat curve - see attached

the trolley switched off when the one battery I was monitoring was at 11.2V - I then physically measured the other battery which was 11.6V - not happy with that

I've had the trolley for 2 months, but it hasn't worked hard - batteries are in series, so 24V system

ok my mistake if its a 24 v system all the charge times must be halved
i assume you dont have a balacer?

now yes based on those voltages when it died they aren't balanced
wasn't to far appart

was that voltage resting for a while or immediately when they died ie the one i would think had about 10% capacity left, but like i said the voltage graphs on the lifep04 is finicky

so you should check the voltage on the 2 batteries again when fully charged if the one is higher you can connect a 12 car bulb
12v down light ie halogen that use more power to pull down the one with higher state of charge

as what will happen is the one that gets full first will hiy it's steep curve and make the voltage jump causing the charger to stop charging while the other battery is still jot fully charged

ideally what you want is that these batteries must be the same voltage when fully charged or depleted

anything in the middle you won't be able to see if they are balanced
you can clip the bulb on the high battery while it's charging it will slow down it's charging allowing the lower state battery to catch up