 So a bit about me and my brother. So I'm Anand from India. I'm pretty new here, also pretty new to Hacker spaces. So I'll just make my best effort to like explain this one. So yeah, I mean, so my brother had this home server where he hosted a few self-hosted a few services. We tried like handle like self-hosting an email server. And then we, for a brief period of time, we also had this new social where we like self-hosted a social network inside. And yeah, we also self-hosted our blocks in using an Intel MUC and just an ADSL broadband connection. So our, so why do we need an UPS for this? So this is back in India where we have like frequent power cuts, like it can go up to even 12 hours at some point of time. So, so our question is like, how do we give a backup for backup or supply for our server? So for that, we put up one of the standard UPS which basically which like converts your AC power packing and stores the battery. And then wherever it goes, where it becomes a DC and then it goes back gets converted into an AC power back into a 230 volts AC power. So our idea is like, how do we make it like simpler without having doing all of those tests? So that's where we thought like we should probably try to make a DC UPS so that we can give like a backup for our server. So, yeah, so a bit about this Intel MUC. So it is actually a really a low power server. Like it's, it had a processor which is Pentium quad core which is like actually does like similar to this Intel Atom series consuming like really low as six watts. So we actually needed like a really small battery for it. So then for some time we were using a standard UPS battery which is like 12 volts, seven A heads which you can see in most UPSs. And that could give up to like 12 hours of backup for this server. So what I tried here is to like design a DC UPS to give a like backup for the server. So I mean, like there are like a lot of off the shelf UPS which we can go, but we added like an additional objective here. So like how do we use the most simplest available components? So how do we, I mean like we could use like an SMD, SMD electronic components and like really high end chips which would battery management chips are available but those would not be available in a common store. So in a common store, what you would get is through-hole resistors, then you would get the most common microcontroller which is at mega 328P. So yeah, and a very limited selection of characters. So that was one of our objectives. So to use the most commonly available components and make the circuit design simple to the point that where we could actually hand assemble it in a perf code. We didn't get to that point yet, but that was our objective. And lithium batteries were out of the question. So we went with sealed lead acid battery. One thing at that point was little expensive and with lithium batteries, you really need, you need like good battery management systems. Otherwise, I mean, things could go really wrong. Yes. So yeah, so that's what, so this is what we decided. And then yeah, one of the advantages of this BCOPS could be that could be like lighter than a standard EPS designed properly and also like more efficient. So this was the objective. So I mean, I'll just like the given way of how the design how we didn't kick it. So what we had as a power source was in the top left corner which is in this Intel NUC adapter, which gives like 19 volts. And through the Intel NUC port, we had to give this particular NUC fortunately could accept like a wide range of voltage. It like 12 volts to actually yeah, 12 volts to 19 volts. So the output of this EPS is like 12 volts. And fortunately it is closer to the battery. So the battery's voltage is typically it varies anywhere from like 10.4 volts to 14.4 volts. So that's something. So then I had a buck converter, which like to convert stepped on the 19 volts into the battery's voltage. Then I then for the microcontroller we used at mega 328P to like manage the battery charging. So for battery charging, we used we planned to use like a three stage charging algorithm. So a very standard algorithm, which will be like implemented by the microcontroller. And then like, so when there is, so then there are like other production and scatter circuits around this. So there is this manual bypass. You can see in the top right, which can help us to like bypass the entire board. If something goes wrong, then we had the undervoltage lockout system. Then in case the battery voltage is really low, we need to like cut it off. So we had this experience in which the battery voltage went really low and our service, our disk was damaged, like corrupted. So that undervoltage lockout system. Then, then there is a gate right as like, so for this also like we use the most commonly available PWM IC that is in the local shop, which was like TL 494. Then we had like status and alert system, which we had like a series of LEDs to display the charge, how much charge was there in the battery. And then a warning beeper like in case the battery is low. So yeah, this is the basic technical idea of it. So I managed to get up to the point where we did have one iteration of this board. So I mean, unfortunately I don't have any photos because I kind of lost it when the phone was gone. So I mean, this is a 3D view of from kicker run out of the board. So yeah, the deserve it looked like with all the through-hold resisted components. And then we had this microcontroller here. And then I mean the input was supposed to be here and the output is here and the battery is supposed to be connected here. And then we had a series of LEDs to like show the battery so yeah, this is what we had done. So this is what we have gone till this point. So yeah, I just thought like I would share it here.