Router UPS

ZipZwan

New member
Joined
Nov 3, 2019
Messages
14
I am currently working on a small UPS for Router. My Router is a 9v version, not 12v are most Router.
I am going for a version with 2x 18650.
Initialy this was going to be my step up:


image_avzyfq.jpg


After some reflection, I came to why use both component at the same time??
I mean from my understand and from what I found around this BMS can charged already the 2x 18650 connected to it, of course when a 8-9v supply is applied to the BMS, so why use TP5100 which is kind of a charging Board. This board is used to charged 2x18650 in serie,if it is the objectif requested...
I means, my guess is to use only one component, it seems kind of pointless to use both of them.
Plus the TP5100 model, I got is only a charging board, it do not have Ouputs pins, so not able to proctect batts from over discharge! Initially, I went for this board mainly for the visual infomation of batteries status: the led (charging and charged) cause BMS got not info about charge status.
So the BMS alone is maybe the best choose.
Of course all this, is my personnal analysis, so maybe you can give some better recommandation.
Thx you in advance
 
The bms doesn't "charge" the cells. it just monitors and makes adjustments when needed. The TP5100 will do a CC/CV charge curve to keep the cells healthy longer. The bms just keeps them from overcharge/discharge, over current protection and under/over voltage protection. But not charging. You still need a charger.


Your drawing is mostly correct. You should have a boost converter between the bms and the Router. The router will not like voltages dropping below 8V. It will become unstable.
I would actually recommend going with "3s" as you can then use a Buck converter, which is more energy efficient and will draw less current. A boost converter will keep increasing current load on the cells as the voltage drops.

I know this is a backup battery situation here, but depending on what you are doing and how long the power outage is, it could be pertinent to last longer than a few minutes.
 
Without the charging circuit, the empty batteries would try to suck ~10Amps or more from the AC/DC adapter. The adapter will hopefully detect an overload and shut down.... or overheat and burnout, and possibly start a fire.

You could use a special AC/DC wall adapter that already has a CC/CV charging circuit integrated, such as this one. Might have a TP5100 inside :)
https://www.aliexpress.com/item/4000174767703.html
 
Korishan said:
The bms doesn't "charge" the cells. it just monitors and makes adjustments when needed. The TP5100 will do a CC/CV charge curve to keep the cells healthy longer. The bms just keeps them from overcharge/discharge, over current protection and under/over voltage protection. But not charging. You still need a charger.
Thx you very much for you information. I read a bit about BMS, but there are so many info around that everytime I get a bitconfused. About the monitoring and adjustment to the batteries, I got it. Where I got losted is because I was following this project:

image_ikvifk.jpg

Router UPS
I know that the BMS used is different because it got balacing feature, and mine do not (I ll go for a "manual" balancing). But after further analysis, it is the CHARGER that he usedthat kind of the keyof his build!! According to you explanation combinedthe comment ofajw22, I came to this understanding. His build do not have a charging part, simply because the charging part you pointed/explained is handled by this special charger!!! Charger that cut out, when the load tells it, thats enough I am not hungry any more (I know fancy explanation), in contrary with ordinary charger will continu to pump power into the batteries... Very sad project with BMS do not point that special "intelligent" charger should be used with BMS!! They just say charger connected to BMS!!
Korishan said:
Your drawing is mostly correct. You should have a boost converter between the bms and the Router. The router will not like voltages dropping below 8V. It will become unstable.
I would actually recommend going with "3s" as you can then use a Buck converter, which is more energy efficient and will draw less current. A boost converter will keep increasing current load on the cells as the voltage drops.
About the Boost converter, I tried a simplier version as this one:


image_ushekm.jpg

https://www.instructables.com/id/Ups-for-Routers/
Only downside was that the Leds of the TP4056 were both blinking when the battery was charged. Maybe cause the battery was draining and charging at the same time. Anyway My Router did not mind, as 8,5v was still coming to it, so was working ok.

I am sorry, I did not explaned my needs from this UPS from the beginning. I have a generator, but it take few to kick in and another few for router to be back online. So my needs are just few minutes, idially 10min.


So at the end I came up to this 2 options:
1- the same of above (initial drawing) combined with a Boost Converter

or

2- UPS using TP4056 as this:


image_qussze.jpg
 
ZipZwan said:
Korishan said:
The bms doesn't "charge" the cells. it just monitors and makes adjustments when needed. The TP5100 will do a CC/CV charge curve to keep the cells healthy longer. The bms just keeps them from overcharge/discharge, over current protection and under/over voltage protection. But not charging. You still need a charger.
Thx you very much for you information. I read a bit about BMS, but there are so many info around that everytime I get a bitconfused. About the monitoring and adjustment to the batteries, I got it. Where I got losted is because I was following this project:

image_ikvifk.jpg

Router UPS

I know that the BMS used is different because it got balacing feature, and mine do not (I ll go for a "manual" balancing). But after further analysis, it is the CHARGER that he usedthat kind of the keyof his build!! According to you explanation combinedthe comment ofajw22, I came to this understanding. His build do not have a charging part, simply because the charging part you pointed/explained is handled by this special charger!!! Charger that cut out, when the load tells it, thats enough I am not hungry any more (I know fancy explanation), in contrary with ordinary charger will continu to pump power into the batteries... Very sad project with BMS do not point that special "intelligent" charger should be used with BMS!! They just say charger connected to BMS!!

LOL nice video. Love the humor.

However, he is mistaken in the video. That BMS does not have a charging circuit. HX-2S-JH20 reverse-engineered (2S-10A-Balance-Li-ion-Protection-PCB)

He is relying on the BMS's ability to cut the power input when cells reach 4.2V. This is not a good idea. Dangerous in fact, as the above article mentions during balancing procedures.
I suppose he's also relying on the wall charger as well.

There is (and I'm going out on a limb here) NO cheap BMS unit that charging capability built in. If you want charging built in, you are looking at much higher priced units.


ZipZwan said:
Korishan said:
Your drawing is mostly correct. You should have a boost converter between the bms and the Router. The router will not like voltages dropping below 8V. It will become unstable.
I would actually recommend going with "3s" as you can then use a Buck converter, which is more energy efficient and will draw less current. A boost converter will keep increasing current load on the cells as the voltage drops.
About the Boost converter, I tried a simplier version as this one:


image_ushekm.jpg

https://www.instructables.com/id/Ups-for-Routers/
Only downside was that the Leds of the TP4056 were both blinking when the battery was charged. Maybe cause the battery was draining and charging at the same time. Anyway My Router did not mind, as 8,5v was still coming to it, so was working ok.

I am sorry, I did not explaned my needs from this UPS from the beginning. I have a generator, but it take few to kick in and another few for router to be back online. So my needs are just few minutes, idially 10min.


So at the end I came up to this 2 options:
1- the same of above (initial drawing) combined with a Boost Converter

or

2- UPS using TP4056 as this:


image_qussze.jpg

Yup, that'll work. The TP unit will only charge back up once the cells drop to a certain level. However there is one major issue here. Will you be powering the TP board 100% of the time and relying on that input to power the router?? If so, this isn't a good idea as the TP4056 can only output a max of 1A and it only does this when the cells are around <4.0-4.1V, then it starts to taper off on current output.
The TP5100 on the other hand can do 2A.

However, what I would recommend is adding a little more circuitry to have the DC power from the wall supply also power the router while it is powered up. Then, when the supply is disconnected, the battery portion is connected to keep the router up. Once the power is restored, it gets flipped back over to the wall charger.

Also it would not be good to run even 2 cells in parallel to boost the voltage up to 8.5V. I still recommend 2s and go with a buck converter. Less stress on the cells that way.
Example: If your router does 8.5V @ 1A = 8.5W. 8.5W / 4.2V = 2.02A on the battery side. Laptop cells are no designed to run at 1A. They like to be more around 500mA or less.

I'm sipping my coffee as I type this, so I may be off on a few things. Hopefully others will chime in with some other ideas as well
 
Korishan said:
LOL nice video. Love the humor.

However, he is mistaken in the video. That BMS does not have a charging circuit.HX-2S-JH20 reverse-engineered (2S-10A-Balance-Li-ion-Protection-PCB)

He is relying on the BMS's ability to cut the power input when cells reach 4.2V. This is not a good idea. Dangerous in fact, as the above article mentions during balancing procedures.
I suppose he's also relying on the wall charger as well.

There is (and I'm going out on a limb here) NO cheap BMS unit that charging capability built in. If you want charging built in, you are looking at much higher priced units.
thx you, and you are correct, he got nice humor and some not so bad project, at least compared to the crap you can find on yT...
I totally agree with you about the BMS, least be realist and talk proper word BMS is Battery Management System, and not Battery Charging Managing System!!!
Maybe a bit weird, but just to say that I agree with you, charging should be handle by a proper component and not be handle by BMS... btw, nicely found this reverse-engineering about the BMS

Korishan said:
Yup, that'll work. The TP unit will only charge back up once the cells drop to a certain level. However there is one major issue here. Will you be powering the TP board 100% of the time and relying on that input to power the router?? If so, this isn't a good idea as the TP4056 can only output a max of 1A and it only does this when the cells are around <4.0-4.1V, then it starts to taper off on current output.
The TP5100 on the other hand can do 2A.

However, what I would recommend is adding a little more circuitry to have the DC power from the wall supply also power the router while it is powered up. Then, when the supply is disconnected, the battery portion is connected to keep the router up. Once the power is restored, it gets flipped back over to the wall charger.

Also it would not be good to run even 2 cells in parallel to boost the voltage up to 8.5V. I still recommend 2s and go with a buck converter. Less stress on the cells that way.
Example: If your router does 8.5V @ 1A = 8.5W. 8.5W / 4.2V = 2.02A on the battery side. Laptop cells are no designed to run at 1A. They like to be more around 500mA or less.
LOL, that is exactly what I was working on before to read your reply!! A pure Backup up, that kick in, when there is power failure!! because in the build, it is a kind of OnLine UPS, so Power always drain from Batt, and during my testing when the battery gets full charging both led of TP4056 start to blink, it is kind of mind wash for the poor TP4056 IC lol
So I came with a schematic with a relay, and when power fail the draining from the battery starts, and when the power is back, the TP is charging lonely the Batt without to be disturb, while main power adapter is supplying the Router.
About Amp consumption my Router is rated 850mA, and during my testing I barely reached 300mA!! so I am far from the 1A you refered to, but of course it is always beter to stay safe than be sorry...

Here my schematic:

image_vghthl.jpg
 
Looks like it'd work, as long as the capacitor can handle the load long enough while the relay switches. There are plenty of 9V capacitors (I'd use 16V, tho. Lots of them on old computer motherboards) that are around 1000uF. So you could parallel them if need be.

300mA is pretty good, but is that while it's idle? Try doing a test with all possible devices connected downloading files. This will cause the current draw to jump. The 850mA rating is under full load, this includes WiFi if it has that feature.

Buuuuut, if you use 2x 2000mAh cells, making a 4000mAh battery, this would give you about 13 hours of run time, estimated (if I did my math right: 4000 / 300).


If you could go with a solid state relay, that'd be even better for several reasons. 1) a lot less power draw while switched, meaning less heat generated 2) faster switching

Another thing to suggest, if using the magnetic relay, use the NC while on battery power. This would help negate #1 above. So when you plug in the wall supply power, it would power on the relay and force it over to that source. This would mean that the relay would "always" be powered while on this source, though. Sooo, you would need to take this into consideration as well.

But I suppose if it's only going to be on battery for about 10 mins max, then using battery to activate the relay wouldn't be so bad. Just take that into consideration when choosing cells to drop in.

If the router can take up to 850mA, add on the power consumption from the relay, this is your theoretical max load you would need to design for. And obviously if you "actually" use half that, so much the better. But also try to plan for the heaviest possible load condition and size accordingly
 
Can not thx you enough, so many great informations!!


Korishan said:
[size=small]Looks like it'd work, as long as the capacitor can handle the load long enough while the relay switches. There are plenty of 9V capacitors (I'd use 16V, tho. Lots of them on old computer motherboards) that are around 1000uF. So you could parallel them if need be.[/size]
yep I got a bunch of capacitor and use to add them to my builds output to keep the output stable. I got some 470uF and probably some 1000uF, and of course always go for a higher rating, so 16v minimum!



Korishan said:
[size=small]300mA is pretty good, but is that while it's idle? Try doing a test with all possible devices connected downloading files. This will cause the current draw to jump. The 850mA rating is under full load, this includes WiFi if it has that feature.[/size]
hmmmm you right, 300mA was just my phone connected. At full load, I did not reach to that part during my testing LOL. But my Router got only one LAN device to care about, I mean only my PC is connected by cable, the rest is Wifi device, around 4-5. I will again look at numbers during next testing and add some loads (devices).



Korishan said:
[size=small][size=small][size=small]Buuuuut, if you use 2x 2000mAh cells, making a 4000mAh battery, this would give you about 13 hours of run time, estimated (if I did my math right: 4000 / 300).[/size][/size][/size]
Right, even after some perso reflection, I said why bother with 2 batts, I only need it for few min, and during my testing with 1 batt, it lasted 30min very well, I means not lasted, I only runned the test for 30-40min and then disconnected the stuffs (was tired)... and checked status of the battery it was around 3.95V, not bad I think... So I will initially go with 1 battery and if more backup time is needed , add the 2nd to increase capacity, of course balance them before to plug them, safety first!



Korishan said:
[size=small][size=small][size=small][size=small][size=small]If you could go with a solid state relay, that'd be even better for several reasons. 1) a lot less power draw while switched, meaning less heat generated 2) faster switching[/size]
[/size][/size]
[/size][/size]

[size=small][size=small][size=small][size=small][size=small]Another thing to suggest, if using the magnetic relay, use the NC while on battery power. This would help negate #1 above. So when you plug in the wall supply power, it would power on the relay and force it over to that source. This would mean that the relay would "always" be powered while on this source, though. Sooo, you would need to take this into consideration as well.[/size][/size][/size][/size][/size]
About solid state relay, I only have old popular relay. I mean the model we easily find everywhere. About the switching time, I also had some doubt during my drawing... will it kick fast enough for the router to dont turn off, and be like a reboot? Only way to be sure is testing the theory, and simulate a power outage? power failure

I decide to go with the TP4056 because I got a bunch off them + did not receive the TP5100 yet, and after deep reflection I only need few min of Backup...
I will connect everything on breadboard, use my oldy router for the time being, cause I need practical testig with the main dude (router), enough of theory lol

I am very greatfull for all the support and the infomations shared!
 
remember, having 2 cells in parallel as opposed to a single cell is not only about capacity. It's also about shared current flow. You want these to last a long time with minimal wear.

So, looking at it another way, 8.5V @ 400mA = 3.4W. 4.0V / 3.4W = 1.18A
That's still a lot for a laptop cell. Now, if you are using a powertool cell, that's a different story. They are designed for high amp load, though they are lower on the capacity.
 
Korishan said:
remember, having 2 cells in parallel as opposed to a single cell is not only about capacity. It's also about shared current flow. You want these to last a long time with minimal wear.
you are right , I forget took take care of the poor 18650 LOL, using 2 cell mean shared drain, instead of drain from only one. So less wear in time... longevity...

Thx for the learning math, so if I follow, If I go for 2 cells:
1.18A /2 = 590mA drained from each cells, accordig to your previous math... So better than to drain 1.18A from a single cell.
And yep, I am using laptop savaged cells, around 1400-1600mA capacity, not very bad. At least for my need they do the job, so got to take take of them :D
 
Back
Top