Mini DC UPS packs

G

gpn

Member
Joined
Jan 21, 2018
Messages
263
I want to build some small packs to run some of my DC items if the power goes off. The main things I want to run are the wireless wifi mesh nodes for my network and a wireless hub for my weatherflow tempest wx station. Ideally I could keep the internet and wx online in a power outage. The weather hub runs on a simple 5v 1amp micro usb plug and only uses a few wh per day. The mesh nodes run on 12v and I haven't measured their power consumption yet.

The pack would have to tolerate being plugged in 24/7. I was thinking a 4s and maybe 2 or 3 in parallel. I would use dc-dc converters for each device that I need to power. I'd like to keep these small and simple. I can 3d print the cases if I need to do that. It might even make sense to use a jehu style pcb mount for this.

My only real concern is having a charger plugged in all the time and not pulling power off the cells.
 
How about a smallUPS on ebay and build a 18650 battery to match - probably 7s.

Maybe something like this -https://www.ebay.com/itm/APC-UPS-45...tion&brand=APC&_trksid=p2047675.c101195.m1851

It will automatically have charging, grid->ups failover etc. You can easily convert 120v output to suit what you need.

I use 18650 DIY battery packs on several APC UPSs and they just sit there...oldest one is 2 years and going strong :)
 
I just figured skip the AC -> DC -> AC -> DC (wall wart) conversion since one runs on usb and the other a 12v barrel plug.
 
Interesting idea/plan.
To get around the always plugged in problem, you could add a little extra circuitry. Have a couple diodes on the input side of main power feed. This keeps backfeeding to the sensor. Then on the power side of the diodes use an SSR to turn on/off the battery pack in case the grid goes down. On the load side of the diodes a capacitor to help keep any dips from happening.
For charging, i suppose any of the cheaper chargers would work. Though, matching 12V is hard on standard Li-Ions.

Basically this would become a cheap/simple DC ups ;)

Of course the idea and general plan is crude. That would need to be worked out through R&D :p
 
gpn said:
I just figured skip the AC -> DC -> AC -> DC (wall wart) conversion since one runs on usb and the other a 12v barrel plug.
Click to expand...
You could tap the battery (of the UPS)directly - no AC involved except UPS keeping the battery charged. Maybe the key is you want to build something from scratch - and I get that :)
 
If any group understands doing it yourself I would hope it is this one! :D

The SSR is an interesting idea. They have super fast switching times and the load is minimal so I don't need to worry about heat sinking or super large SSRs.
 
keep your vpc near 4.0v and let it float.
holding at 4.2 greatly increases degradation.
limiting to 4.0 drops an insignificant amount of capacity in exchange for extended life.
 
Here is a slightly different question. Lets say I used something like a tp4056 to charge a 1s2p cell pack to power the weather station only. Can I have it plugged into a 5v source as well as running the load at the same time? Or is there a simple charging circuit for something like that?
 
Can someone point me to discussions about a multi-TP4056 board please? I know I read it but can't find it. I also recollect a mention of an upgraded version. A nudge would be helpful here to please.
 
Close to what I'm looking for. I'm sure I've seen a design where the circuitry for 4 or 5 cells was all on one board, and powered by a single supply connection.

Wolf - What do you use yours for... is it just to pre-charge cells to increase the throughput of your testers? I've been doing just that with my singular D4. With only 2 testers, its a slooow process. Darn COVID has meant shipping from China is taking over 6 weeks.
 
You could make packs & run the gear from them via suitable buck step downs & keep the pack charged by running it from a larger CCCV charger/supply set to eg 4.0V/cell per suggestions?
Eg mains > charger/PSU > cells > buck > device.

This way, when the mains goes down, the devices don't even notice & continue on, when mains comes back, PSU recharges cell & powers device again....?
 
Redpacket said:
You could make packs & run the gear from them via suitable buck step downs & keep the pack charged by running it from a larger CCCV charger/supply set to eg 4.0V/cell per suggestions?
Eg mains > charger/PSU > cells > buck > device.

This way, when the mains goes down, the devices don't even notice & continue on, when mains comes back, PSU recharges cell & powers device again....?
Click to expand...

That was my original plan and I apparently asked the question poorly. How bad is it for the cells to be held at the top end? I'd be happy to overbuild the capacity and then knock down the top charge to even something lower like 3.9. This time of year we have frequent power outages due to heavy monsoon storms. Otherwise our power supply is fairly stable for the other 8 or so months of the year. The cells would likely spend around 98-99% of the time on mains power.
 
Beware that many off-the-shelf consumer-level DC UPS suffer from various problems.

Some are bogus U = Uninterruptible because the switch time is too long for many devices.

Most also provide no control over the standby charge level and simply maintain the battery at 100% SOC (or close). While this yields the max runtime on battery it also yields the worst battery life because Li-ion battery life is seriously degraded by the amount of time they spend at extreme SOCs and temps (e.g. NASA's deep spacemissions use very shallow cycles around 50% SOC, which may increase cycle/calendar life bymore than an order of magnitude- possibly lasting a couple decades). The difference can be quite extreme - e.g. only 6 months battery life when always fully charged on standby vs. more than 10 years when intelligently maintained.

Generally the rule of thumb is that to maximize Li-ion cell life we should strive to minimize the time that the cell spends at extreme voltages and temperatures (which accelerate internal parasitic reactions and other degradation processes). Because these degradation processes are typically nonlinear (even exponential) even a small change away fromthe extremes can reap large rewards in battery lifetime. See these slides for a very nice summary of theprimary degradation / aging mechanisms in Li-ion batteries (and see the further discussion in surrounding posts there).

For some real-world numbers regarding such see post #18 below, andthe graphs I postedhere, which highlighthowLi-ion chemistry behaves so differently in this regard vs.other common chemistries(e.g. lead acid, whose cumulative lifetime energy delivered does not exhibit such strongdependence on said discharge depth and itsbalancing around 50% SOC). This deserves to be much better known since it can yield tremendous improvements in battery lifetime when competently exploited.

A common mistake is that many folks try to transfer their intuition from other chemistries to Li-ion, which fails miserably in this case (and in otherstoo, e.g.vaperpants-on-fire incidents caused due to lack of knowledge that - unlike other common cells (e.g. alkaline AA) - for cyclindrical Li-ion cells the entire 18650 can is negativeso even a small metallic object like a key can short the cell),
 
[quote pid='64461' dateline='1593703006']


A common mistake is that many folks try to transfer their intuition from other chemistries to Li-ion, which fails miserably in this case (and in otherstoo, e.g.vaperpants-on-fire incidents caused due to lack of knowledge that - unlike other common cells (e.g. alkaline AA) - for cyclindrical Li-ion cells the entire 18650 can is negativeso even a small metallic object like a key can short the cell),
[/quote]


Jokingly - Nooobody here knows what the magic smokes is , we like ozone,bright lights and burnt nose hair. the other guy did it. :D
 
Wayne\ said:
Wolf - What do you use yours for... is it just to pre-charge cells to increase the throughput of your testers? I've been doing just that with my singular D4. With only 2 testers, its a slooow process. Darn COVID has meant shipping from China is taking over 6 weeks.
Click to expand...

I just sits in my shed now not being used. Too bad you are on the other side ofthe world otherwise I'd ship it to you as I only used it in the beginning of my harvesting adventure. This wasbefore I found out about IR, and was testing every cell that came across my grubby little fingers.
Now the ones that don't pass the "IR" test don't need to be pre charged.
I pretty much know what I'm getting now and straight to the tester they go.
Wolf

Waynes World and OP
Just realized this is kinda hijacking the OP's original post. Sorry
 
Apologies from me 2 :p
 
To expand on my remarks in post #14, here is some data on the degradation caused by float-charging and storing at high SOC and high temperatures. Below is a graph from this 2014 study of calendar life degradation (no cycling) for Sanyo UR18650E cells stored at various SOCs at 50C. As the study shows, you get only 107 days lifetime storing at 100% SOC, but 2.7x that (288 days) storing at 95% vs. 100% SOC, and 14x that (4 years by extrapolationstoring at 0% SOC. Many other studies show analogous results, so this is a general rule of thumb to keep in mind for prolonging the life of your Li-ion batteries.

image_riabwe.jpg

There they write: "During calendar aging tests cells were stored at different SOC and temperatures. Fig. 6 shows capacity fade and resistance increase for cells stored at 50C at different SOC. As expected, cells stored at lower SOC exhibit a longer lifetime. Especially cells stored at 100% SOC show a much faster degradation compared to other SOC. These cells only reach a lifetime of 107 days at 50C until capacity reaches 80% of the initial value (linear interpolation). At this time the resistance increased by a factor of 1.3. For comparison, cells stored at 95% SOC reach 288 days until end of life (linear interpolation), whereas a cell stored at 0% SOC has an extrapolated lifetime of about 4 years (linear extrapolation of the last three measured data points)"
 
You must log in or register to reply here.
Back
Top