UPS suggestions for GW3000a

[Auchtermuchty Weather]

Now I can use the ecowitt_http driver with backfill, it seems worth giving the gateway a UPS, but I am totally clueless about what would be suitable. The longest I've been without power for is almost 2 days. That's exceptional but I think gives a benchmark for how long the UPS needs to keep the gateway going. 

[michael.k...@gmx.at]

Without knowing which devices you want to power with the UPS, it's difficult to give a precise recommendation.

Since you only want to power the GW3000 and then backfill, it seems realistic that many UPS models on the market would be sufficient—even for nearly two days—provided the battery is in good condition.

A more common approach is to use the UPS to support multiple devices, such as a network switch, NAS, computers running WeeWX, routers, or access points. Many UPS units and NAS devices support automatic shutdown when power is lost, right out of the box.

Here’s my setup:

I use an EATON Ellipse PRO 1200 to power a NAS, one access point, a network switch, some Ecowitt gateways, a router, and an n100 mini PC. The NAS is configured to shut down automatically after 10 minutes without power. I’m not sure how long the UPS could sustain all these devices, but our power grid is very stable. I can only recall outages lasting less than an hour, and they happen less than once a year.

I plan to install an automatic blackout switch. I have a 24 kWp PV system with a 22 kWh battery, so the UPS would only need to bridge the brief moment the blackout switch takes to activate.

[John Smith]

Get a UPS with a Lithium battery, better still get a LiFePo4 based battery, they don't catch fire like other chemistries.

Most, if not all, UPS still use deep cycle lead acid batteries (LAB), but these have a woefully small number of charge cycles before wearing out and then have to be tossed out. Although lead usually can be easily recycled, or melted down to make sinkers for fishing.

Quick web search said some LABs can do 200-500 charge cycles but I've never gotten anywhere near that number of charge cycles from UPS before they become door stops.

Lithium batteries do cost more than LABs, but can still have 80% usable capacity after 3,000 cycles, LiFePo4 batteries usually fall to 80% after 6,000 charge cycles. the Fe part of the chemistry means iron, so they don't use them in cars due to weight and slight increase in bulk, houses don't care though.

This is something to look out for, for example the batteries in Google phones are only rated for 80% after 1,000 charge cycles, where as Samsung is up about 3,000, so not all batteries are created the same, even when using similar chemistries.

I bought a couple of LiFePo4 135Ah 12v batteries off the net, my UPS is 24v and a 12v to 24v converter costs more than the 2nd battery, so then I just extended the cables from the existing UPS with the dead batteries.

I don't just have the weather console connected, but also have a switch, the router, the WNTD (4G fixed wireless) for the internet and a number of servers as well.

I estimated it should easily be good for a few days with the current load, and yes we've had a multi-day outage in the recent past as well, but thankfully since upgrading the UPS we have never had to test how long it'll actually last.

Also most UPS are only designed/rated to handle a brief supply outages and aren't great for more than 5mins due to the LAB, fully discharging LABs wear them out even quicker.

So don't go cheap, it won't end well if you are trying to have coverage for much longer outages, more so if they are frequent as well.

As a last resort we have a generator that plugs into the switch board to power the house, UPS do buy time to get the generator setup and running, and then cut over again when power is restored.

Still waiting for a house battery, with blackout protection, we finally got approval to install.

On Thu, 2 Oct 2025 at 17:05, Auchtermuchty Weather <shunr...@gmail.com> wrote:

Now I can use the ecowitt_http driver with backfill, it seems worth giving the gateway a UPS, but I am totally clueless about what would be suitable. The longest I've been without power for is almost 2 days. That's exceptional but I think gives a benchmark for how long the UPS needs to keep the gateway going. 

--
You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/weewx-user/b459bf18-bc64-417b-b12f-82e0572cfba5n%40googlegroups.com.

[Graham Eddy]

sensor usually has battery backup for months.

sensor controller (eg davis console) usually has battery backup for weeks.

weewx server backup, you are probably looking at SOHO UPS, for minutes to hours - more than that takes something more sophisticated/expensive like a portable power station (lithium) or full-blown premises battery backup [i now have home battery on grid/solar for blackouts]

⊣GE⊢

On 2 Oct 2025, at 5:05 pm, Auchtermuchty Weather <shunr...@gmail.com> wrote:

Now I can use the ecowitt_http driver with backfill, it seems worth giving the gateway a UPS, but I am totally clueless about what would be suitable. The longest I've been without power for is almost 2 days. That's exceptional but I think gives a benchmark for how long the UPS needs to keep the gateway going. 

[michael.k...@gmx.at]

@John Smith:
There’s a very good reason why lead-acid batteries are commonly used in UPS systems. Lead-acid batteries have a long lifespan when their state of charge (SoC) is consistently kept high. In a typical UPS setup, the battery is not subjected to frequent charge/discharge cycles. Instead, it remains fully charged and is only used in rare cases—such as during an AC power outage. Therefore, the low number of charge cycles is not a disadvantage for lead-acid batteries in this context.

On the other hand, maintaining a 100% SoC is one of the fastest ways to degrade a LiPo battery. LiFePO₄ batteries are a better alternative in this regard—they’re more tolerant of high SoC levels and have a lower risk of fire. However, I’m skeptical that their lifespan would exceed that of lead-acid batteries in a typical UPS scenario.

If your use case involves frequent and prolonged power outages, then a UPS with LiFePO₄ batteries would be a better choice. Otherwise, a lead-acid battery will serve you perfectly well.

[Greg Troxel]

First, I recommend against "lithium ion" aka "lithium nickel managanses
cobalt" as too unstable to be in a house in greater than phone/laptop
quantity.

UPS units are designed for much shorter runtimes. Power is needed to
run the UPS and the inverter even at tiny load. I have never seen one
that would go for more than few hours.

If someone has a UPS that can go 24h with a 10W load, please post
details of the UPS and data from your experiment.

If your load that you want to keep up is 12V or 5V, then I would look
into a LiFePO4 battery and a small dc-dc converter to e.g. 5V USB, with
a LiFePO4-specific charger. For extra points, get a 50W or more solar
panel permanently mounted outside with a LiFePO4-specific solar charge
controller, and then you should be able to keep your weather logger up
indefinitely.

The first step is to get a power meter and see what you need, and then
to understand the rest.

It's cool to hear how many of this (not representative of the general
population surely) group have home batteries integrated with solar.
They are fairly rare around me (the subsidies in the US have been
focused on generation not storage, and few people seem to value
being able to use their solar during grid outages), but becoming
increasingly common. I know one person with it operational and another
whose system is about to be commissioned.

[John Smith]

On Thu, 2 Oct 2025 at 19:48, 'michael.k...@gmx.at' via weewx-user <weewx...@googlegroups.com> wrote:

On the other hand, maintaining a 100% SoC is one of the fastest ways to degrade a LiPo battery. LiFePO₄ batteries are a better alternative in this regard—they’re more tolerant of high SoC levels and have a lower risk of fire. However, I’m skeptical that their lifespan would exceed that of lead-acid batteries in a typical UPS scenario.

I've had a UPS with a 20Ah LiFePo4 battery in operation to keep a WiFi router in bridge mode powered up for an IP camera with a dodgy WiFi chip or something that constantly dropped out and it was pot luck if it would re-connect but works fine on Ethernet, since before COVID happily chugging along.

I've bought other UPSs with LABs since then, and even replaced the batteries in one with fresh LABs. The last LAB replacement was to LiFePo4s though.

If your use case involves frequent and prolonged power outages, then a UPS with LiFePO₄ batteries would be a better choice. Otherwise, a lead-acid battery will serve you perfectly well.

This is exactly my situation and we just keep wearing out LABs as a result.

When the power drops out here we usually get 2 quick outages followed by a 3rd much longer one that won't come back for 3-4hrs while the sparkies locate and rectify the fault. The 2 quick ones are probably an automated circuit breaker somewhere resetting and then giving up after failing to clear whatever is causing a short.

We're almost at the end of the section of grid we're connected to and there is no substation connecting to multiple circuits.

The 2 short drops in power would probably be dealt with ok by a UPS with a LAB, the 3rd time it drops for hours on end not only needs the higher energy density of lithium, but the ability for lithium batteries to be ok with a full discharge.

The electronics in lithium batteries, apart from balancing the charge between cells, should prevent a lithium battery from actually fully discharging, as that will kill lithium batteries.

In any case the thread author wanted protection for long, possibly multi-day outages, means they will need to get LiFePo4 batteries.

[John Smith]

On Thu, 2 Oct 2025 at 21:54, Greg Troxel <g...@lexort.com> wrote:

It's cool to hear how many of this (not representative of the general
population surely) group have home batteries integrated with solar.
They are fairly rare around me (the subsidies in the US have been
focused on generation not storage, and few people seem to value
being able to use their solar during grid outages), but becoming
increasingly common.  I know one person with it operational and another
whose system is about to be commissioned.

 We had a Federal election in Australia in May, one of the election promises was a 30% subsidy on house batteries. They told us before the election it'd be good for the country as a whole, and not just people willing and able to get one installed that.

This is because everyone will benefit from the programme due to lower peak demand in the evening as people either self consume excess solar produced during the day, or because they can get paid $1/kWHr to export to the grid during peak time.

Just waiting for ours to be installed, along with the rest of the country except Graham Eddy, who already has his.

Although I've never come across this situation before, the installed price dropped more than 20% from when we placed our order in June, you'd expect the company to absorb the extra as profit, can't say I'm complaining.

[Francisco Alonso]

I have been monitoring the power draw on the GW3000 for a few weeks now and it sits solid in 162 mAh of current consumption, 5.25 V. Any UPS with a greater capacity than 10000 mAh should last at least 2 days. At least, my power bank did when I tested it.

--

You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/weewx-user/CAGTinV67GNhxk_tdNGyAV_Ds81cOYWLKpm1cKO-F7YYB8g4%2BdQ%40mail.gmail.com.

--

Saludos/Regards,

Francisco Javier Alonso

[Greg Troxel]

Francisco Alonso <fj.alons...@gmail.com> writes:

> I have been monitoring the power draw on the GW3000 for a few weeks now and
> it sits solid in 162 mAh of current consumption, 5.25 V. Any UPS with a
> greater capacity than 10000 mAh should last at least 2 days. At least, my
> power bank did when I tested it.

Yes, but that's output capacity of 10000 mAh at 5V, not "contains a
battery rated at 10000 mAh". A UPS that produces 120V (or 230V, 240V)
is running an inverter, and you have to consider the usage of the UPS
control circuit and the power required to keep the inverter going even
if the only load is a USB power supply. My point is that you can't
reason about linear loads with real 120V UPS units.

A power bank, that is low-power DC only, and designed to be efficient at
low loads, is another story. It's great that you actually tested!

[Auchtermuchty Weather]

It's only for the gateway. Anything else I might want a UPS for is the other side of the house. 

It is the gateway in the house, not the sensors in the garden that I want a UPS for. It's not a Davies Console - it's an Ecowitt so it doesn't have any built-in batteries.

I see some people in WXForum are using a power bank. That might be the right route for me to go - I see the power connector is USB-C - so will look in that direction.

[John Smith]

Then do what Greg suggested and get a Lithium based device that outputs DC. Such as this, but it's only 12-24v.

https://powershield.com.au/product/powershield-dc-mini-n36-ups/

--

You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/weewx-user/ea999cf1-7d44-4911-bd19-b1ee0d70c8a0n%40googlegroups.com.