diy solar

diy solar

16S, 272Ah Lishen + TinyBMS - Build Thread

Most battery monitors are counting Coulombs. They do this by mathematical integration. The algorithm is going to take samples which are both low and high current on that curve, but they will all average together. So any well designed battery monitor should show accurate values.

Though some BMS have a SOC monitor feature, it can be inaccurate, as some have this feature kind of slapped on at the last second.
That's what I figured. According to the manual, it's sampling all signals every 100 milliseconds. At that rate, it's hard to image that the BMS wouldn't be calculating a relatively accurate mean. I'll keep a close eye on SoC values vs. rested voltage values and hopfully things don't start drifting apart.

So, fingers crossed this is all 'normal' and I just need to ignore the readings displayed on the BMS software...

Thanks everyone for the quick feedback on this.
 
Some BMS have an "average" current value which is a mathematical average of several samples.
 
Step 5 - Battery Case: With testing out of the way (mostly), it's on to final assembly. My plan is to arrange the cells in a 4x4 configuration, as shown in the first picture below. For compression/stabilization, each block of 4 cells will have 1/4" aluminum plates on either end with stainless steel ties holding them together. For the case, I've ordered a 12"x16"x30" Welder's Box. The box will be insulated with 1" XPS all around. Under each cell will be a 12V, 25W silicone heating pad (run in series off the 48v supply, still working out the control strategy).

The whole package will be on its side (the door of the welder's box will swing down to reveal the cells sitting on their sides). This will keep the depth to 12 inches. I'm going to mount the box on the wall on a sheet of 3/4" plywood supported by some heavy duty shelf brackets. The two Growatt inverters will be mounted side-by-side directly above the box.

I'm interested in feedback on this cell arrangement. Only drawback I can see is that I'll need two different bus bar lengths. Big advantage for me is that I can compress the cells in groups of 4, allowing me to move them around in their final, compressed form.
Screen Shot 2021-02-01 at 5.27.44 PM.pngScreen Shot 2021-02-01 at 5.28.25 PM.png

Here's a few pics of the first compressed pack. I put two strips of mounting tape between the cells, and between the plates and the cells, to keep everything secure. I used two clamps to pre-compress the cells before strapping them down as tight as I could get them. The straps seem to be staying nice and tight (they sound like guitar strings when plucked...). I'm not too worried about the actual compression force, as i don't plan to push these cells anywhere close to their limit.

IMG_7570.JPG IMG_7574.JPG IMG_7577.JPG
 
You might want to consider repositioning the welders box. From the Lishen spec sheet.

c)During the course of storage or usage, keep the cells upright
 

Attachments

  • Lishen 272ah battery date sheet.pdf
    207.9 KB · Views: 15
You might want to consider repositioning the welders box. From the Lishen spec sheet.

c)During the course of storage or usage, keep the cells upright

Yeah, saw that in the sheet and decided to interpret it as more of a recommendation than a requirement... Seriously though, I've read quite a few post on this subject, and general consensus seems to be that these prismatic cells are pretty comfortable on their sides. I may start with them on the sides, and if I observe any odd behavior (green ooze pouring out of the vents), I'll flip them back upright...
 
EVE has said it's ok to mount the cells on the large sides but not the edge side. I know people are going to debate this. The reason it's best to keep them upright is because of the jellyroll within the cell. You will never see green stuff coming out of the vent...lol. But you may end up with some partially dried up jellyrolls.

Sometimes people interpret spec sheets to fit their own needs. But they are your cells.
 
Setting up the cells in blocks of four will certainly make life easier, looks like a good setup IMO.
The case you're installing into is metal, so do remember to line the case itself with a non-metalic / non-conductive material, you never want to allow the case to become a conductor.
 
Battery Case Update: I got the remaining 'blocks' of cells assembled today. Overall, I'm really pleased with how this part turned out. They cells are completely immobile, and the stainless steel bands seem to be doing a great job keeping them compressed. With all four blocks together, it's starting to look like like a serious power supply. I might even say it looks like a pro job...from about 25 feet away...

I cut and drilled all the bus bars today. This ended up being easier than expected. I purchased 3/4" x 1/8" bar stock from Midwest Steel and Aluminum. I should note that I got the Aluminum compression plates - cut to size - from these guys as well, and was very happy with the quality of the material/cuts. 80" of bar stock was $44 and the eight aluminum plates were $42. With shipping and tax, the whole order came out just under $100.

I cut the short bars to 91mm, and the longer 'bridge' bars to 106mm. I used my 10" miter saw to cut them to length. I wasn't sure this was going to work (safely...), but with everything clamped down, and applying minimal pressure, the carbide blade ate right through them. The cuts were super clean. Drilling went fine as well. After marking and punching the hole locations (72.5mm spacing for the short bars, 87.25mm for the long bars), I went directly to a 1/4" bit (no step-wise drilling necessary in copper). The 1/4" holes fit the 6mm with just the right amount of play (IMO). Still need to clean them up and add heat shrink.

IMG_7578.JPG IMG_7581.JPG IMG_7582.JPG
 
Battery Case Update:

Looks great! Just keep in mind that the tightness of the fixture varies with SoC. I recommend discharging to 30% SoC or less and then checking to see how tight everything is (or isn't).
 
Battery Case Update: I got the remaining 'blocks' of cells assembled today. Overall, I'm really pleased with how this part turned out. They cells are completely immobile, and the stainless steel bands seem to be doing a great job keeping them compressed. With all four blocks together, it's starting to look like like a serious power supply. I might even say it looks like a pro job...from about 25 feet away...

I cut and drilled all the bus bars today. This ended up being easier than expected. I purchased 3/4" x 1/8" bar stock from Midwest Steel and Aluminum. I should note that I got the Aluminum compression plates - cut to size - from these guys as well, and was very happy with the quality of the material/cuts. 80" of bar stock was $44 and the eight aluminum plates were $42. With shipping and tax, the whole order came out just under $100.

I cut the short bars to 91mm, and the longer 'bridge' bars to 106mm. I used my 10" miter saw to cut them to length. I wasn't sure this was going to work (safely...), but with everything clamped down, and applying minimal pressure, the carbide blade ate right through them. The cuts were super clean. Drilling went fine as well. After marking and punching the hole locations (72.5mm spacing for the short bars, 87.25mm for the long bars), I went directly to a 1/4" bit (no step-wise drilling necessary in copper). The 1/4" holes fit the 6mm with just the right amount of play (IMO). Still need to clean them up and add heat shrink.

View attachment 35704 View attachment 35705 View attachment 35703

I've been wondering why people don't use steel strapping. Be sure and let us know how it works.


I am going to use that in a setup very similar to yours with stainless steel strapping and buckles.
 
I've been wondering why people don't use steel strapping. Be sure and let us know how it works.


I am going to use that in a setup very similar to yours with stainless steel strapping and buckles.
I agree. The stainless strapping is a really nice way to avoid 'bulking up' the bank. That tool would certainly make things easy, and for a great price.

For reference, I had the 1/4" plates cut to 6.875" x 8.000". The width (6.875) is pretty much perfect. The plates overhang the cells by about 2mm on each side, so the stainless straps don't touch the cells at all. In hindsight, I should have cut the plates about 1" taller so my plexiglass lid can sit right on top of them. As they are now, I'll need to use some 'risers' to get the to sit above the top of the studs.
 
Last edited:
Looks great! Just keep in mind that the tightness of the fixture varies with SoC. I recommend discharging to 30% SoC or less and then checking to see how tight everything is (or isn't).
Once I get everything back together, I'll take things down to a low SoC again and see what the guitar strings sound like...

On an unrelated note, I stumbled upon your build thread and realized that I've seen quite a few of your videos! I'm about 5 pages into your thread. A ton of useful info on there. This may already be answered in your thread, but did the Orion BMS suffer from the same unstable current readings that I'm seeing? I considered Orion, but the price wasn't quite justifiable for me.
 
please consider the weight of the whole battery box. One cell is about 5kg. All 16 Cells have about 80kg + the transport case. Two persons are able to carry 40kg. So i would suggest to use two battery boxes.
 
please consider the weight of the whole battery box. One cell is about 5kg. All 16 Cells have about 80kg + the transport case. Two persons are able to carry 40kg. So i would suggest to use two battery boxes.
Thanks for the feedback FrankB! I don't ever plan to move the assembled case (this is going into a building, not a vehicle). On the rare occasion that I need to move the cells, I will remove the 'bridge' bus bars and pull the cells out of the box in blocks of 4. Again, that's the big advantage of this arrangement (no back injuries!).
 
Once I get everything back together, I'll take things down to a low SoC again and see what the guitar strings sound like...

On an unrelated note, I stumbled upon your build thread and realized that I've seen quite a few of your videos! I'm about 5 pages into your thread. A ton of useful info on there. This may already be answered in your thread, but did the Orion BMS suffer from the same unstable current readings that I'm seeing? I considered Orion, but the price wasn't quite justifiable for me.

Yes, it's all over the map, too. Although it provides an instant current and an average current value - so the average current is better (but still moves around a lot). But in real life, I won't be looking at battery current. I'll be paying attention to the inverter watts and my SoC and my solar watts. I won't care about the seemingly random DC amperage readings which won't be particularly prominent on my primary display anyway (below).

The no-load voltage readings of the cells is amazingly accurate, however. Every cell matches my DMM. Balancing works amazingly well.

1612368232121.png
 
Battery Case Update: I picked up the welder's box from Grainger yesterday and got to work on the insulation. I'm lining the inside with 1" XPS (Foamular 150), which has an R-value of 5. Not great, but all I need to do is keep these things above freezing during the daytime, right? Overall, a pretty easy process, and it came out pretty nice. Still need to do the door.

As you'll see from the pictures below, the fit is excellent. I was a bit concerned that I didn't have enough room side-to-side, but it came out pretty much exactly where I wanted it. There's just enough room between cells to slide them in and out. Plenty of room in the front and top to fit the BMS, wiring, and possibly even my main breakers. For anyone running these cells in a vehicle, this is an extremely rugged setup. Cells are almost completely immobile and the case is pretty much waterproof (if you put a gasket on it). $135 for the case and $20 for the insulation. I think it's worth it for the protection of 14-gauge steel. Perhaps the most impressive part of all this: My pingpong table didn't collapse after adding the 50lbs case to the 200 lbs of batteries that were on top of it.

I'm running my first tests of the heating pads now. The pads will sit under the bottom compression plates as shown in the last picture. For this first test, I'm basically just making sure that the 25watt pads aren't able to heat bottom plates to a point that it could damage the cells. Still working out the control strategy, but I suppose that if these ever came on and the batteries were very cold, they could stay on for a while. After 10 minutes, the bottom plates got up to about 90F. That was higher than I expected, considering that the plates should be in direct contact with the bottom cell (a giant heat sink). I may need to add a 'buffer plate' to keep the bottom plates a bit cooler. On a positive note, that 10 min test was enough to raise the temperature of the top cell by almost 2 degrees, and that's with no insulation on the front of the box.

IMG_7584.JPG IMG_7587.JPG
IMG_7589.JPG IMG_7585.JPG
 
If you want to estimate the heating load needed to maintain a certain delta T, this calculator does a decent job.

That's a great tool (assuming it's reasonably accurate). Thanks for the link. I especially like it because it tells me that my 100W of total heating power is not too far off... According to their model, to maintain a delta T of 20C (ie, maintain temps above freezing down to -20C), I would need ~35 watts of heating power. I intend to keep the shop above freezing, but even if the heat was shut off for days on end, I highly doubt the interior would get that cold. Passive solar alone (6 south facing windows with unobstructed slab exposure) should keep things at least 10C above ambient at all times, I would think.

So, in my absolute worst case, my heating pads would need to be on ~1/3rd of the time to maintain temps above freezing. I would think that with the right control method, I shouldn't have to worry too much about the pads overheating the bottom cells. I'll probably still add some kind of a 'buffer'. Any ideas on a material that would provide good heat dissipation?

I think I might have components on hand to set up a relay to modulate current using an arduino. I'd like to see what frequency it takes to maintain a deltaT of 20C or so.
 
(assuming it's reasonably accurate).
I verified it with hand calcs a couple times, its fairly good. My exterior box insulated the same (8 cell pack) needs less than 65W down to 0F.

If you are actively charge/discharging you will need even less due to losses in the pack.

I wouldn't bother with PWM. Just on/off with a couple degree hysteresis works well.
 
Battery Case Update: I'm running my first real test of the battery box heater today. I'm using an Arduino MKR 1000 to monitor temperature and control a relay for switching the heating pads on and off. I don't have a 48v relay with 5v control voltage, so I'm using an optocoupler (4N25) to drive the solid state relay (Crydom ED10F5, 48vDC/48vDC). A wiring diagram is pasted below, in case anyone's interested in how I've got it set up. One question I had while setting this up is whether its better to use SSRs and/or Optocouplers on the high-side or low-side of the circuit. I went with high-side for both, and it seem to be working fine. The SSR is not getting hot at all driving the heating pads (2amps) at 33% duty cycle.

For this first test, I'm using a program of 1 sec on followed by 2 sec off (33% duty cycle). I'm starting here because 1) the calculator that Luthj linked to suggested i'll need about 35watts to maintain a delta of 20C (100 watts of heating pads * 33% duty = 33 watts), and 2) I still have the pads in direct contact with the bottom compression plate, and I'm trying to keep the bottom plate below 40C.

I have 4 10kOhm NTC thermistors wired to the Arduino and placed in the following positions:
-Temp 1: Directly between heating pad #1 and the bottom compression plate of cell block #1
-Temp 2: Dead Center of the full cell bank (side-to-side and front-to-back)
-Temp 3: Top of box, just above the top compression plate of cell block 4.
-Temp 4: Ambient temp, monitored about 2 feet away from the box

Starting temp of all thermistors was 14.0C. I'm now 2.5 hours into this test. Temp 1 quickly jumped up to about 35C, and has since slowly drifted upward and is now at 39.5C. Temp 2 is has only increased about 0.6 deg so far, while Temp 3 is up about 1.3 deg. Ambient temp (in my basement) has dropped to 13.2C (sun is out, so furnace hasn't been running). One thing I didn't really think through until now is the HUGE thermal mass of those batteries. If I assume they have a similar heat capacity as water, it's probably going to take a full day to raise the temp of the batteries 10C. Note that I had an error in my code that was causing duplicating the Temp 3 readings for Temp 4. That explains the strange profile of the ambient readings.

One other note on this: According to the spec sheet, these cells can be safely charged and discharged at up to 65C. That makes me feel a bit more comfortable letting that bottom compression plate get above 40C.
Screen Shot 2021-02-05 at 12.52.32 PM.png
IMG_7593.JPG IMG_7595.PNG
 
Last edited:
Back
Top