"Zephyr" - Finally, the "v4" Fechter/Goodrum/Hecker BMS...
- Thread starter GGoodrum
- Start date
GGoodrum
1 MW
chxs said:
The board has places for two of the FETs but that's really only if you are going to be having charge currents of 20A, or higher. Below that, one of these FETs should be enough. They have a pretty low Rdson resistance, even lower that the 4110.
For the diodes, two are required. I will fix that.
-- Gary
AndyH
10 kW
You'd walk funny, need extra glasses, and need to stock up on motrin.Gregb said:
I'm only allowed to solder and break things - you wouldn't want me designing circuits. :lol: I have replaced all the caps (a few times over) but haven't gone outside my current stock/batch/lots. Time for another Mouser order. Thanks very much!Gregb said:Andy I expect you don't need help trouble shooting but have you for example replaced all the capacitors. In my experience, 30+ years, an out of spec capacitor can cause all sorts of problems. Once chased a problem for over a week because had been assured these new fangled tantalums never broke down........
won't go into my Valve stories. and the Miller effect.
Gary sent his V4.4 board and I've tested it - and it works perfectly. That proves the circuit and the PCB layout.
The only differences between the board he sent and my three boards are the choices for Q3 FET, different types of MLCC caps, and that my 1M resistors are not 1%.
I'll order resistors and caps and see what happens.
I think I figured out what the problem was. It should be very easy to fix and won't change the BOM.
One of the comparators was in a state where both inputs were essentially grounded (equal), making the output depend on minor component variations. By rerouting one resistor, this state can be avoided and should make it work the way it was intended. I'll have to wait for Andy or Gary to test this before declaring success.
One of the comparators was in a state where both inputs were essentially grounded (equal), making the output depend on minor component variations. By rerouting one resistor, this state can be avoided and should make it work the way it was intended. I'll have to wait for Andy or Gary to test this before declaring success.
Zenid
100 W
Is the 1% issue likely to make a difference? I have a whole load of resistors already, and was going to knock the ones I already have off my mouser list to save a few bucks. The ones I have here seem to be to within 2%...AndyH said:
GGoodrum
1 MW
The change is pretty trivial, but I sent Andy my board, so we'll have to wait for him to test it.
GGoodrum
1 MW
Zenid said:
That should be fine. I only used 1% ones because I like the colors to match.
AndyH
10 kW
I've done the mod on a 4.3.6 and a 4.4 board and it works perfectly on both. As the brains have already outlined it's not really about the components.
My 21S V4.4 board with the mod is cooking away balancing my bike battery tonight - it's been HVCing itself for the past six hours and it's going strong.
Very good progress!
[edit] After posting this last night, I walked into the garage to find a fully charged and balanced battery. The PakTrakr on the battery draws from cells 1-3, 9-11, and 17-19 and results in a significant imbalance with every charge. This is the first time I've been able to leave a charger connected to this pack and come back to a fully balanced pack without significant heat and without any cell voltage rising above 3.7V.
RIchard and Gary have done it - they've improved on the V2.x boards and created an excellent BMS!
My 21S V4.4 board with the mod is cooking away balancing my bike battery tonight - it's been HVCing itself for the past six hours and it's going strong.
Very good progress!
[edit] After posting this last night, I walked into the garage to find a fully charged and balanced battery. The PakTrakr on the battery draws from cells 1-3, 9-11, and 17-19 and results in a significant imbalance with every charge. This is the first time I've been able to leave a charger connected to this pack and come back to a fully balanced pack without significant heat and without any cell voltage rising above 3.7V.
RIchard and Gary have done it - they've improved on the V2.x boards and created an excellent BMS!
999zip999
100 TW
What is the price of the BMS ?
GGoodrum
1 MW
999zip999 said:
Well, that is something like asking, "What is the price of a Honda?"
It depends on the number of channels. I only sell the PCBs, but this includes a detailed set of build/test instructions and a Bill of Material (BOM) file that lets you order the parts somewhat automatically, using their handy "BOM Import Tool".Supported configurations include 16, 24 and 32-channels. Other configurations can be handled by not populating some channels.Eventually, Andy, and maybe Richard, will start offering assembled/tested units, but labor costs, so don't expect cheap prices, like the typical Chinese BMS. You get what you pay for.
You have to remember, this is still basically a DIY project. It is meant to allow a user, with decent soldering skills, to build a top-notch BMS unit, for a fraction of the cost of a higher cost commercial unit. The cost savings comes from you providing your own labor. Andy's labor charges are extremely reasonable, in my opinion. Frankly, I think he should charge more.
-- Gary
GGoodrum
1 MW
geoff57 said:
Well, for one thing, it probably won't fit in the box as easily. A single 5W resistor is quite a bit thicker. The other reason I like using multiple smaller resistors is that there is more surface area touching the heat pads, and/or the case or some other heatsink.
-- Gary
999zip999
100 TW
Sorry for the noob question. I got my bike stolen from san juan cap. last week (so be careful) and thinking lipo with a 16 channel bms. about a price ? Thanks
GGoodrum said:
From an electronic standpoint, it will make no difference.
Using series/parallel combinations of smaller ones may be more cost effective. This seems to be especially true of the surface mount parts. The thermal and packaging considerations are also important.
In our case, I recall the parallel pair of resistors winds up being cheaper, partly because you need twice as many and you can get a price break over a certain quantity.
Hifechter said:
So if I could find a 10w 3ohm resistor it would not cause any problems in the way the BMS works.
I want to fit ti to a alu heat sink before it goes in the case so the original fitting poison is out the window anyway so I decided if I was going to go with the heatsink I might as well get a resistor fitted that was designed to be attached to a heatsink .
I understand about the price this way is going to be more expensive, I have found somewhere local to me that can get 20w 3ohm resistors but only if I order 360 units
, they will do have the resistors in 2 ohm and 5 ohm where you only have to buy them one at a time, would either if these work as they would reduce the cost of making up one board dramatically,the alternative that you have listed brings the value down to 2.8 ohm so the 3 ohm should be fine would reducing the value down to 2ohms also work then if so it would be great not having to get the 360 minimum order saves me having to make up 15 of the 24 channel boards and find buyers for the spares, I just don't think there is the Market for that number in the UK.The main advantage of this resistor is it can attach to a heatsink that can be attached to the case and is low profile, I have searched high and low for it elsewhere in the UK and USA but found none nor was there a PDF I talked to the company by phone to get the info I have, 20w is major overkill compared to the 6w that are used as standard but I just think if you use a higher wattage resistor it will not get as hot which can only be a good thing.
Geoff
SlyCayer
100 W
Is the BOM and Instruction on TPPacks.com up to date with all the needed updates?
With all these changes, I am wondering when will be the good time for me to purchase the electronics and start assembling.
P.S: I will have a 20A+ charge current, what FETs do I need to buy and where to install them on the board?
With all these changes, I am wondering when will be the good time for me to purchase the electronics and start assembling.
P.S: I will have a 20A+ charge current, what FETs do I need to buy and where to install them on the board?
geoff57 said:
I think 2 ohms would be exeeding the limitations of the transistor. The BD136 is rated for 1.5A max.
Anything higher than 3 ohms should be fine. It will just run with a lower shunt current. If your cells are healthy and reasonably well matched, this should not be a problem.
With heat sinking on the transistors, you could probably go as low as 2.5 ohms.
damcard
100 W
How do I charge 12s? Can I leave 4 channels off and make a 12s charger? -Damcard
damcard said:
Yes. You can even saw off the part of the board you don't need, but it's a lot easier to just not install the parts. Even with the parts installed, you can leave the unused channels disconnected.
Zenid
100 W
Hi,
I've been updating my mouser parts list in the light of the IDC header change and the new parts list. However comparing v4.4 and v4.4.1 of these documents - for the 24s LiFePO4 section at least - it looks like the new part number for the angled header has been changed in the part number column, but the Mouser BOM number is still the same as the old one, so an import will pull the old header into the parts list instead:
P5, Ln 2:
9-pin MTA 156 IDC header 571-6403879 71-6404459 3 71-6404459|3
Have I got this right? Apologies in advance if I missed something...
I've been updating my mouser parts list in the light of the IDC header change and the new parts list. However comparing v4.4 and v4.4.1 of these documents - for the 24s LiFePO4 section at least - it looks like the new part number for the angled header has been changed in the part number column, but the Mouser BOM number is still the same as the old one, so an import will pull the old header into the parts list instead:
P5, Ln 2:
9-pin MTA 156 IDC header 571-6403879 71-6404459 3 71-6404459|3
Have I got this right? Apologies in advance if I missed something...
Nate
100 mW
FYI - I believe I have a typo on my board. It is on the 5th channel and shows the Rx08 Resistor as 470 instead of 210 like the other channels.
- Nate
- Nate
Nate
100 mW
The BOM 4.4.1 is showing a quantity of 16 for Rx08 and Rx09 in all versions.
- Nate
- Nate
Yes, you're absolutely right. That's a typo on the board, it should be 210.
Thanks for spotting that. We'll fix it on the next batch.
Thanks for spotting that. We'll fix it on the next batch.
GGoodrum
1 MW
You guys have some real eagle eyes. :wink:
The BOM file has been fixed/updated. In addition to the quantity corrections for Rx08/09 and the incorrect part number for the right-angle MTA156 connector, I've also added the TVS diode that Richard added in the latest variant, which is now v4.4.2. The TVS diode basically goes across the charge control FET (or FETs...), and is really there just to add an extra level of protection for higher voltage setups, maybe up to 150V, or so. For 24s, or less, setups, I'm not sure if it is needed. Anyway, that is the only change between v4.4.1 and v4.4.2.
The instruction and schematic files have also been updated.
-- Gary
The BOM file has been fixed/updated. In addition to the quantity corrections for Rx08/09 and the incorrect part number for the right-angle MTA156 connector, I've also added the TVS diode that Richard added in the latest variant, which is now v4.4.2. The TVS diode basically goes across the charge control FET (or FETs...), and is really there just to add an extra level of protection for higher voltage setups, maybe up to 150V, or so. For 24s, or less, setups, I'm not sure if it is needed. Anyway, that is the only change between v4.4.1 and v4.4.2.
The instruction and schematic files have also been updated.
-- Gary
