"Zephyr" - Finally, the "v4" Fechter/Goodrum/Hecker BMS...

[fechter]

I have a question about the resistors on the back of the board. I've got them to squeeze into the box nice and snuggly, so that the resistors make good contact. However a few of them got the chalky layer scraped while getting them to go in (some were a tight fit) and I see traces of the bare metal of the resistors exposed in places.

Could this offer the danger of shorting with the box casing, or are the resistor cases fully insulated from whatever it is inside that does the resisting?

Yes. You want to make sure there is no electrical connection between the case and any of the resistors. You can measure with an ohmmeter from the case to each tap wire and make sure there is no continuity. I have not tried scraping off the outer layer to see what lurks underneath, but if you see metal, that would usually be a problem.

If any of the resistors are shorting to the case, the solution would be to add a layer of insulation between them and the case. Kapton tape is the preferred stuff. It can be stuck to the inside of the case where the resistors touch. Silicone heat sink pad material will also work. Ghetto approach is to lay down a layer of silicone glue over the resistors and use a flat piece of plastic or metal that's been greased as a mold to make it flat. Once the glue hardens, remove the mold. Push the resistors all the way down to the board before gluing to leave some clearance for an insulating layer.

Hopefully none of your resistors are actually shorting to the case and none of this will be necessary.

I guess it would be a good idea to round off the sharp edge of the case before sliding the board in and bend the resistors down enough so they aren't too tight going in.

 

[Zenid]

Yes. You want to make sure there is no electrical connection between the case and any of the resistors. You can measure with an ohmmeter from the case to each tap wire and make sure there is no continuity. I have not tried scraping off the outer layer to see what lurks underneath, but if you see metal, that would usually be a problem.

Thanks for the advice. I checked for continuity between the case and the cell tap points as you advised and none was evident. However I then realised that the case is coated in black paint which - intentionally or otherwise - does not seem to conduct (if I use probes on the case I only get continuity where there is exposed metal in the grooves).

However my concern is that vibration might erode this paint where the resistors touch, and if one is touching metal-to-case then there will be a problem like you say. This is unlikely, as the board is lodged pretty solidly in there and nothing is liable to rub against anything else, but perhaps as a precaution I should get the tape like you say...

I noticed today in another test that when I looked at voltages that they seemed to be 'wobbling' between 3.55-3.6V when measured with the probes positive way round, and -3.6 - -3.65V when measured the other way round. It seemed to do this whether the LEDs were "dialed up" or dimmed. Is this normal? I got a steady voltage with my light bulb scenario...

 

[alternety]

A thought about the resistors. Rather than sliding them in and hoping everything is OK; an alternative just off the top of my head.

Short concept. If you really want to dissipate heat and not damage resistors. Make a clamped heatsink that clamps around the resistors and then attach that to the case.

Make a heatsink using a metal bar for the top of the resistors (side exposed to case contact), and some equivalent material for under the resistors. It can be metal if there are no traces under the resistors. Obviously you don't want to touch metal with a metal bar. Use a few small bolts (#4, #6) to evenly clamp the two pieces together with the resistors in place. Ends and center are probably enough.You should tap (thread) the holes in the plate next to the PC board for easier disassembly and reassembly. You need to use a bolt that has a head that you can use with a countersunk hole so it sits flush with the surface of the top bar. Cut the bolts to proper length to clamp properly but not cause trouble with the PC board. For final assembly put some heatsink compound between the resistors and top plate (you could do the bottom too). You probably don't want to put on the heatsink compound until you are finished with most of the mechanical stuff and everything fits. You need to use a material that pretty much stays in place when it is hot. Use sparingly. I would recommend the material be something specifically designed as heat transfer material. Not a tube of silicon rubber. There is a material made for heatsinks that is a thin silicone with adhesive on both sides. On final assembly, put a THIN coat of heatsink material between the top bar and the case.

You need to get some short screws through the case and into the top bar. With the board in it's final position, clamp the top bar and drill a couple of holes through the case and the top bar. It should help yoy to get alignment if you Insert (don't solder) a few resistors in the board and put them in the clamp you have made. Use whatever method feels good. A little extra bar length would give you a safe place for the end holes. Drill together to make sure you have proper alignment of the holes in the case and the bar. After those holes are established, measure for a hole in the center of the bar in a safe position and drill that one. Again you need to tap the holes in the bar and you should counter sink the holes on the outside of the case. If you try to just use nuts you won't be able to attach or remove it from the case. If you cut the bolts short enough so they don't go all the way through the top bar, placement of the attachment holes is less important. Personally I would use careful placement.

Put all the resistors into the proper holes on the PC board - Don't solder. Loosely clamp on the bars. Push the bar up against the PC board. Make sure all of the resistors are aligned the way you want them. Tighten bar bolts.

Put assembly into the box with the upper lid removed. Use the slot above where you would normally slide the board. Adjust the top bar (it would be on the bottom now) until it is in full contact with the case. Align the mounting holes you put in the case and attach the top bar to the case. Adjust board so resistors are straight. You could probably solder a few resistors at each end and remove the assembly for easier soldering. Solder all of the resistors. Now you can finish the board and put it back in when you are ready.

Just an idea. It involves some work but I think the end result will be worth the effort because all resistors are in firm contact with the sink, you avoid resistor damage, and it looks more "professional".

 

[Zenid]

Just an idea. It involves some work but I think the end result will be worth the effort because all resistors are in firm contact with the sink, you avoid resistor damage, and it looks more "professional".

That sounds like a great idea, but I just don't have the time to put together something like that. I've already spent an enormous amount of time on this and need to get things up and running ASAP.

I'm going to have to just go for some insulating material like the Kapton tape like Fechter suggested. Thanks anyway!

 

[fechter]

I originally designed it to use an aluminum bar as a heat sink. The idea was to put a thick bead of silicone over the resistors, then place the bar over the silicone. The whole mess can then be slid into the case and left to cure. Once cured, the board can be removed from the case without risk of damaging the resistors. A thin film of heatsink grease can be applied to the bar when it is put in.

The way the resistors are mounted, there isn't much clearance for a thick bar, but possilby a thin piece of aluminum sheet metal would work. As long as the resistors and sheet metal are glued together, there is not much chance of vibration causing damage later.

The alternate design was to mount all the resistors on the top side of the board and use a thick bar between the bottom and the case.

 

[alternety]

From what you are saying, you use regular old silicone rubber. I am not sure the heat transfer of that material is suitable for the application. I don't know; just feels wrong. There are thermally conductive adhesives. There is the potential for better heat conductivity with something like silicon rubber because the contact area can be greater than the essentially point contact you get using heat sink material.

If you move the board to a higher groove in the case you should have enough space to do as I suggest. I did not see anything that fills the space between the board and the top of the case. Unless you use a vertical connector. The internal connector does not seem real useful to me. If I were building it I would use wire directly to the board and a connector(s) either mounted on the case end or externally with a cable going to it. A hole in the case with a grommet and external connectors seems simpler. There are many sources for wrapping that makes a neat cable out of the bunches of wires leaving the box. Rather than individual wires, I would use a 2 conductor (red/black) stranded cable. Then I could easily label the cables as to what cell they go to, hence minimizing confusion while wiring.

For heat conduction is is important that the contact pressure of thermal junctions be positive and consistent. I believe my suggested approach will do a better job of that. It is more work.

 

[fechter]

Silicone glue is an amazingly good heat conductor. It seems to work well enough with just the point contact between the resistor and the case, so silicone will just increase the contact area. The resistors are not running that hot.

I agree about bundling the wires. A gromet and a U shaped cutout on the end panel would allow removing the guts a lot easier. You could solder the wires directly to the board and have a connector outside. If you pay attention to the connector wiring, there should be no confusion when connecting or disconnecting the plug.

 

[Zenid]

Well I finished wiring up the cell taps on the board and have finally gotten round to a proper test.

I attached the main pack terminals first, then plugged in the connector for the tap wires. Finally, I switched on the charger and the main indicator light came on green, but flashing intermittently, the ammeter reading on the charger showing bursts of current every second or so. The already charged packs showed as glowing LEDs, and over the course of a couple of minutes another came on as presumably another cell reached 'full'. I noticed the resistors on the first two banks were getting hot, so I unplugged everything, and the board it into its case so that resistors were touching the case, then ran a continuity test to make sure their were no shorts, and ran it again for another ten minutes or so. Another LED came on, as the charger did its job.

However, I've noticed something that doesn't seem quite right: Though the first two banks are getting quite hot, the third one remains stone cold, even though all the banks are jumpered. I've noticed that the third bank almost looks 'perforated', like it's almost an optional extra part of the board. Is charger current perhaps not making it to the third bank or something, or is it just that the banks only get hot if at least one cell is at full capacity.

Here's a vid showing how it's behaving. It's dark and there's not much detail, but it shows the basic behaviour of the LEDs and charger current reading.
http://www.youtube.com/watch?v=PUvn4ig8XIk

Do you think everything is OK, or should I check something?

 

[Gregb]

I would measure the voltage on the channels of the third bank that isn't getting hot and see if they are charging

eg put your voltmeter across one and turn the charger of and on and see if it changes. or if the voltage from the cells is there first. This is where the cell log is invaluable

 

[Zenid]

False alarm! The LEDs on the third bank started to come on a little later. So far 19 of my 24 cells are charged, and I'm waiting for the remainder to play catch up. Also, does the fact that five of my cells seem to be lagging quite far behind indicate a problem, or is this common during a "breaking in" period? They all measured identical voltages when they arrived.

I've also noticed that the whole thing is running pretty hot, so I switch off at intervals to let things cool down again. Could this be to do with the calibration of my power supply? I left it so the LEDs are just lit ever so lightly on the calibration test (running the charger across the cell banks). Should I maybe dial it down a bit so that the LEDs are barely lit at all?

A couple more videos of my test:
http://www.youtube.com/watch?v=QN6tYkzuMNM
http://www.youtube.com/watch?v=DG7luWwZccQ

 

[Gregb]

Running "hot" is often a matter of interpretation. Put an ammeter in series with one of the charging channels and measure the current or alternatively measure the PD across the parallel network and work it out. You can then see what power is being dropped across them.
you could also put the board in front of a fan...... :wink:

 

[Zenid]

Well the good news is, as you might have seen from my videos, it all seems to be basically working. It charged and balanced all the cells in my pack nicely, though did seem to get quite hot.

The problem now it with the EOC adjustment - it just doesn't seem to behave as it says it should in the documentation.

I dial down the EOC resistor to minimum, but as soon as the charger is switched on, it goes to green (after the orange LEDs light briefly) and doesn't charge. It only actually started charging when I put the jumper in place, but the LED is still green. On one occasion it kept going though regardless of whether the jumper was removed or stayed on, but now it's not doing that, and needs EOC jumper on again.

[EDIT] I can't make any sense out of how this thing is supposed to behave when EOC jumper is off. Could I have some help here please?

 

[fechter]

In the video it looks to be working properly and in the balancing mode. During this phase, the current comes on until the highest cell hits HVC, then the current drops until that cell bleeds down by the shunt, then the cycle repeats. On the first charge cycle, it may take a while to get all the cells balanced. After that, it should not take nearly as long. During balancing, the thing will get pretty hot, but this is normal. You could try tweaking the charger voltage down a little to reduce heating.

With the EOC disable jumper in place, the EOC will never trigger.
Once you get the cells balanced enough so the current reading while charging is under 1A or so, then it should kick in with the jumper removed. I suspect you just need to let it cook longer to get all the cells balanced before it will work. You can adjust the pot higher to make it cut off sooner. With the pot all the way CCW, it may never turn off. Lowering the supply voltage a little will help too.

 

[alternety]

These are naive questions, but I am just learning.

I see a number of scenarios for charging.

1 - smart charger with balancing (may or may not need bulk supply for power. That makes sense to me.
2 - bulk charger with the BMS board. Using EOC function to turn off bulk charger. That one is a little trickier for me to understand.

As I understand it, the charging protocol starts in constant voltage mode, looks for a particular transition in the voltage of the pack/cells, and then goes to constant current. Or vice-0versa - not clear.

The bulk supplies seem to be CV devices.

Is it sufficient (and wise) to use a bulk CV power supply (e.g., Mean Well) and the BMS board as the sole charging system? Or have I completely missed the point?

I also keep seeing references to the Hobby King voltage logger. I am not identifying the actual product name on HK. Can someone provide a bit more specifics to allow me to find it.

 

[Zenid]

With the EOC disable jumper in place, the EOC will never trigger.
Once you get the cells balanced enough so the current reading while charging is under 1A or so, then it should kick in with the jumper removed. I suspect you just need to let it cook longer to get all the cells balanced before it will work. You can adjust the pot higher to make it cut off sooner. With the pot all the way CCW, it may never turn off. Lowering the supply voltage a little will help too.

Thanks for the feedback!

Unfortunately I've been having another problem which is a bit of a show stopper. The BMS seems to just randomly fail to work at all It happened once first thing today, but then righted itself. I put it down to a loose connector I'm using on the board, but now it's happened again and I can't figure it out.

When I connect the charger (pack all connected up to the BMS) the BMS light comes on green, but only dimly. When I measure the voltage (on the board), it fluctuates rapidly, but this is a normal reading for chargers that are waiting for a load. Putting the EOC jumper on or off makes no difference, neither does starting it with the jumper connected or not.

I think there's some kind of intermittent connection issue, but I don't know where to start in tracking it down. I noticed earlier when I was charging that it all seemed to be a bit dimmer, and moving it would make the LEDs flicker or dim, like something wasn't properly connected. Could you help me diagnose it, as nothing works at all at the moment...

Thanks.

[Edit] I just did some tests and my resistance across the first two cell taps is 2.8K rather than the 168K or so I'm seeing across the others, could this mean a dead capacitor? If so, is this a separate issue or could it be related to the failure?

 

[Gregb]

I also keep seeing references to the Hobby King voltage logger. I am not identifying the actual product name on HK. Can someone provide a bit more specifics to allow me to find it.

Bad news. It seems they may have dropped it. Just checked their site for "Celllog-8M/10952 Cell-Log 8M " which is what i ordered and nothing came up. They don't answer emails anymore so you will have to register with them to ask if they are going to restock.
It is made by shenzhen junsi electronics

 

[fechter]

Not sure what's going on there.

Any problem with the cell circuits would only affect the control circuit if HVC was coming on. You can measure the HVC line. If you use the light bulb setup, you can double check the shunt voltage settings on the cell circuits. An abnormally low resistance measurement usually indicates something is shorted.

It could be that your charger just won't start when the pack is too full. That would be a function of the charger.

On the control circuit, try measuring the voltage on the 12v supply when the charger is connected. You should have at least 10v there.
U1 could be susceptible to static electricity damage from the EOC disable jumper pins, so that would be one suspect if the 12v supply is OK.

The board is designed to work with a 'dumb' CC-CV charger. At the beginning of the charge cycle, the supply will be in CC mode until the pack reaches the CV setting, then the current will taper off and the voltage will remain steady. Once the charging current tapers off to a fairly low value, this signals end of charge and triggers the board to terminate the charge.

There are several voltage monitors, but the one referred to in this thread is the CellLog8.
http://epbuddy.com/index.php?main_page=product_info&products_id=27
Progressive RC was the other place that had them.

 

[Gregb]

hat's $5 more than HK but it still good value, especially as HK don't appear to have them anymore.

 

[Zenid]

Not sure what's going on there.
Any problem with the cell circuits would only affect the control circuit if HVC was coming on. You can measure the HVC line. If you use the light bulb setup, you can double check the shunt voltage settings on the cell circuits. An abnormally low resistance measurement usually indicates something is shorted.

I'll do this, but first thing is to get any life at all out of the control circuit. When I take a resistance reading of that first cell bank, I don't get the steadily rising reading to 168K Ohm or so. As I understand it, that rise is the capacitor charging. If there's no rise, doesn't that indicate there's no capacitor working?

It could be that your charger just won't start when the pack is too full. That would be a function of the charger.

I checked the charger by connecting it directly to the terminals of the pack. It comes on fine when I do that.

For now, I've gone back to my 'dumb' charger that I used to test it initially.

On the control circuit, try measuring the voltage on the 12v supply when the charger is connected. You should have at least 10v there.
U1 could be susceptible to static electricity damage from the EOC disable jumper pins, so that would be one suspect if the 12v supply is OK.

OK. From the circuit diagram, it looks to me like I should be getting 12V across pin 3 of Q1 and earth. Is that right? If not, then where specifically should I measure? I'm afraid I don't understand much about how it works.

I'm not getting any readings to speak, and the whole thing is just dead. I've plugged my 'dumb charger' in and I'm getting around 80V into the charger terminals on the board but nothing is happening at all. The main LED is dead, and I'm only measuring tiny voltages on the control citrcuit.

Could that 'resettable fuse' be at issue? There was an episode where I thought something might have gone wrong with it a while back but I thought it was OK. Now I'm not so sure. How would I test for this?

In conclusion. 80V is going into the board. Nothing is happening.

 

[fechter]

You want to poke around Q1 to check the voltage regulator first.
You should be getting full charger voltage on both sides of the fuse.
You should have 15v on the gate pin, and about 12v on the source pin.

 

[Zenid]

You want to poke around Q1 to check the voltage regulator first.
You should be getting full charger voltage on both sides of the fuse.
You should have 15v on the gate pin, and about 12v on the source pin.

Okay. 80V makes it through the fuse and that diode on the other side (D3) to the middle pin ("D") of Q1. But I see no voltage at all on the top (G) pin or the bottom (S) one.

Thanks for your continuing help! Perhaps this workthrough could be the basis of a troubleshooting section in the manual.

 

[fechter]

It's wearing me down though...

Anyway, no voltage on the other pins is a problem. Most likely D1, the 15v zener diode that goes to the gate.
Find D1, using diode check mode or ohmmeter, measure across D1 with no power applied. It should measure like a diode, open one direction, continuity in the other direction. If it reads near zero both directions, it's shorted. This could possibly happen if the drain pin was accidentally touched to the gate pin. That's a cheap part at least.

 

[Zenid]

It's wearing me down though...

Me too. I'm exhausted. I was hoping to get this all finished while I had some time off work. I was absolutely gutted when it just stopped working... This is so depressing.

I'm back on 12 hour shifts this week so you'll get some peace.

D1 diode checks out - same readings as its counterpart across on the other side. All the other diodes look fine too.

I also re-ran the light-bulb voltage tests. The first bank looks to have the same voltage as the rest, but - like last time - the voltages aren't steady and seem to wobble up and down quite a bit.

P.S: Thanks for the cell log 8 link!

 

[MrEco]

Hi,

We are testing the new BMS right now and need more assembled BMS in near future.

Does anyone know where A. Hecker is? Tried to e-mail his company, but all mails returnes.

Help please.

Best Regards

Rickard

Sweden

 

[fechter]

Hi,

We are testing the new BMS right now and need more assembled BMS in near future.

Does anyone know where A. Hecker is? Tried to e-mail his company, but all mails returnes.

Help please.

Best Regards

Rickard

Sweden

I've sent him a couple of messages, but no reply. I hope he is OK.