Guru
Joined: 13/10/2014
Location: AustraliaPosts: 2193
| Posted: 01:36am 18 Dec 2023 |
I grabbed a 16S 200A BMS and an Active equalizer, I took 16 cells from one bank (4) batteries and managed to get all cell sense and balance leads correct first go for 54V out. I have not top balanced each cell yet. I decided to charge the 54v battery at 25A and slow the rate of charge if the Delta was more than 300mv, and just let the active balancer do its thing for now.
Every cell appears to be fine, I will do a discharge / charge cycle and plot the real capacity when I get a chance – I have a tester for that.
At the moment the 16 cell battery is running the home and workshop with the cell voltage variation across 16 two year old cells at 42mV.
The second bank was in almost the same sorry state, they used the correct size Copper links this time, but again, some terminal holes were not deep enough with screws bottoming out without fully clamping the links. Lot of loss in the Link to terminal connections. I think I caught these just in time.
BTW I did get a free replacement for the faulty dead unit. Apparently one cell had failed?
I used a JUNCTEK Battery monitor to make an automated top balancer with relay cut out at a precise cell voltage, works brilliantly with a modified 40A 5V Adjustable PSU set to 3.6v and using the Terminal voltage sense lead and Over Voltage and over current relay cut-out from the Junctek.
I’m also using a JUNCTEK BM for quick visual SOC and to protect the Batteries from a rare but not unheard of failure of the Solar Regulators going short from Input to output, (yes I know some expensive units have isolated outputs) the Charge input is isolated from the batteries via a Contactor with incorporated economizer, works brilliantly.
I found two cases where the BMS fused the charge FETS from over voltage input due to the mentioned Solar reg short – results – 16 fried LiFePO4 cells. So I don’t fully trust the BMS isolation FETS, yes I know there are other BMS units that use a real relay or contactor, or have the control interface for them, but it is what it is, and this is simple and low cost for me to add, BTW my inverter also incorporates fully automated cap bank inrush pre-charge using the same big HV contactor, the inverter has timed mains transfer along with automated start stop or shutdown if selected.
The 48V BMS was DELGREEN 200A 16S and their 16S Active balancer. Again there are haters of this brands, whatever, not interested in hearing it.
I will agree that their WEB site sucks, they have limitations depending on your expectations, but they were at a reduced cost and up to half the price of the same unit sold in AU. I have a second unit coming to convert the final bank.
The PC Software is ok, but it’s a kludge in having options for all models and features that don’t apply to most of the units I see people using.
The Active balancer works well with the PC SW and the Phone app, however the Phone APP had the same Kludge of being very amateurish with every needless or unused tab available for every unit made, but nothing tailored to the BMS you are using.
Once I had a handle on the way the PC and Phone APP worked with the Balancer and BMS (independently) I was able to work through all settings and makes sense of every unused option. In the end the SW is fine and not the horror story I was led to believe. It could be really good if they would listen to the complaints.
I have seen posts of users complaining that the readings jump around and they are dismissed by others as not happening to them at all. Well - the PC app voltages and current jump around due to the way the SW is written, it appears to NOT return discharge and charge voltages in the same packet?? The SW shows charge and discharge current on the same Meter, but on each alternate read update. - So each data update is time shared if current is being drawn WTF? If no current is being drawn it look fine, if current is being drawn and you are not charging it look fine. Do both and it alternates I/O current on each update.
Another complaint is the apparent inaccurate voltages between cells when the Balancer is running.
Now one of the options from DALY is a combined package of BMS and external Balancer, and there is another that appears to have the Cell connection Loom spliced and split for the BMS and Balancer connectors. This will make the problem worse and the expiation is simple.
The Balancer is reading the cell voltage at the end of the longish wires running from each cell terminal, if you are drawing 1A in balance current, you will be measuring the voltage drop across the cables, they will appear to be inaccurate, if you splice into the BMS loom, you will also cause the BMS cell voltage indication to sag and dance about for the same reason.
As the cell voltage sense wires connect to the Positive terminals only, reading cell voltage across these terminals will also include any voltage drop due to bad inter cell connection of the Link Bars, obviously only under load, but everything needs to be checked under load to minimize these problems or cell balance won't work correctly. Currently every connection is less then .07mv @ 45A. I still need to go through the Links though.
BTW I was two short in the build and ended up making two from some copper sheet I have, made the same way as the new EVE link bars with two thinner bars together which are more flexible, and on those two links I read 0mv @ 45A.
I found the Balancer readouts to be very accurate, the BMS was about 14mV out, not bad.
FYI - I have an accurate voltage source to calibrate my meters, I find most of the meters and readouts available on line should be correctly renamed to - Random Number Generators.
This is the load, 43A @ 53v and 42A @ 12V using a combination of series or parallel or single connection. That’s a couple of KW at 50 volts.
.
Edited 2023-12-18 11:46 by KeepIS