Lipo charging

[Nigel R]

Further to the above.

A quick google suggests the balance resistors are 120ohm, 6 in parallel (I think this was dug out for the last thread too).

Balance current is therefore: 4.2 * 6 / 120 ~= 0.2A

Balance current of 0.2A would suggest (c/10) a 2000mAh pack is around the intended max size.

For that 2000mAh pack, it would also suggest that, given that a full charge has 500mV difference (3.7 to 4.2) it could alter any given cell voltage by around 50mV / hour. Very broad brush, but there it is.

Stick on a 5000mAh pack and things slow down, of course...

[Frank Skilbeck]

Posted by Nigel R on 11/09/2018 13:00:21:

I realise I'm ping in the breeze here. But.

1) Frank - it makes no odds (practically) to the balance time to reduce the charge rate during the initial charge. The rate of balance improvement is fixed by the balance circuitry. You might as well charge as fast as you can during the initial period.

 

Nigel, I'm not so sure, the current reduction around the highest voltage cell is fixed and on at least one of my chargers the balancing seems to start from the initial start, so if you are charging a 2200 mah cell at 2.2A and the current reduction on the highest voltage cell is only 100mA then it will still charge up pretty quick and not allow the other cells to catch up, but if you set the charge rate at say 0.7A then the low voltage cells have more time to catch up. What I have found is that if the highest voltage cell gets to 4.2v then the charger can't quite bypass all the current through this cell and so it carries on charging, at a very low rate, and you can end up with it going slightly over charged, but if the cells all balance before they get to 4.2v then it seems to balance easier.

I maybe wrong but it works for me

Edited By Frank Skilbeck on 11/09/2018 16:53:46

[Martin Harris - Moderator]

Posted by Nigel R on 11/09/2018 15:37:27:

Stick on a 5000mAh pack and things slow down, of course...

That last point may have a great deal of relevance - my experience is largely with packs around 2 Ah although the couple of 4 Ah packs I use regularly have stayed nicely in balance throughout their life (several years now).

What size packs (and levels of imbalance) have you been testing BEB?

[Dave Hess]

Posted by PatMc on 09/09/2018 21:19:34:

How does the CellPro charge through balance leads that are not plugged in ?

I'm not to sure whether I understand your question correctly. All lipo chargers require you to connect the balance leads for a safe charge. Some guys use massive packs of lipos for their Ebikes, like 16S 20Ah, and they do what they call a bulk charge. That means charging through the power leads without the massive array of balance leads connected. They do it because 16S chargers aren't very common and connection sequences are complicated to brake it down to multiple chargers, though they have to do a balance charge say every three charges or so and at least check the cell voltages every charge to confirm that nothing is going out of the safe zone.

As I said earlier, chargers like the BC168 charge through the balance leads without the power wires connected. They treat each cell as a separate channel and charge it to whatever you set, so it's like charging a 6S battery with 6 separate chargers.

The iMax B6 charges work by overcharging every cell a bit through the power leads, then bleeding down the higher ones through the balance leads. It's an iterative process and the control system doesn't seem very fantastic, so it takes a long time. The BC168 and similar chargers don't do any bleeding. Instead, they charge each cell to the target voltage directly and then stop.

I have drawers full of B6-type chargers, but I don't use them anymore. As well as the time to balance, some are not very accurate in the way that they measure cell voltages. You only have to measure with an independent device to see that. It's not helped by all the fakes that are floating around. Some of them look the same, but are terrible.

Edited By Dave Hess on 11/09/2018 20:40:17

[Biggles' Elder Brother - Moderator]

That's interesting Dave, that BC168 is the first intelligent individual cell charger that I have seen at a reasonable price - they are usually well over £200-250 minimum.

I also love a quote from their webpage:

"Its intelligent design does not work like those old style balance chargers (eg. iMax B6) which always discharge and recharge the battery in order to make all the cells balanced (this is just very stupid and wasting time)."

A sentiment I wholeheartly agree with!

BEB

Edited By Biggles' Elder Brother - Moderator on 12/09/2018 13:17:49

[Nigel R]

Shame they can't get it right, BEB, but I suppose they have to sell their over-engineered sledgehammer somehow.

The B6 doesn't inherently discharge and charge cells. Balance current is entirely shunted by the balance resistor applied in parallel to each cell. As the top cell gets to "full", it switches to balance mode. Top cell does not go up or down at that point, as the balance current is completely shunted. Low cells are brought closer to the top cell, as fast as the shunt resistors allow.

There is plenty of scope for cheap copies to get something wrong, the tolerance on the shunts, or the balance current, or the algorithm controlling of all this, or whatever, which prevents them from working as they should. However there is nothing inherently wrong with the B6 design. It is a very simple, elegant, cost effective solution.

Apparently it's just very hard to understand that it won't correct a bad imbalance quickly. But if I'm honest, I believe that is generally a problem with the user, not the charger. Balance charge the packs every charge, and the balance time will never be an issue, as there will only be minor imbalances to correct.

I think I must agree to disagree with a few posters on this thread - as we're just going in circles.

[Dave Hess]

Posted by Nigel R on 12/09/2018 14:21:46:

The B6 doesn't inherently discharge and charge cells. Balance current is entirely shunted by the balance resistor applied in parallel to each cell. As the top cell gets to "full", it switches to balance mode. Top cell does not go up or down at that point, as the balance current is completely shunted. Low cells are brought closer to the top cell, as fast as the shunt resistors allow.

There is plenty of scope for cheap copies to get something wrong, the tolerance on the shunts, or the balance current, or the algorithm controlling of all this, or whatever, which prevents them from working as they should. However there is nothing inherently wrong with the B6 design. It is a very simple, elegant, cost effective solution.

Apparently it's just very hard to understand that it won't correct a bad imbalance quickly. But if I'm honest, I believe that is generally a problem with the user, not the charger. Balance charge the packs every charge, and the balance time will never be an issue, as there will only be minor imbalances to correct.

I think I must agree to disagree with a few posters on this thread - as we're just going in circles.

Ok, I hope i don't come across as argumentative, but that's not my understanding of how the B6 works. I just opened up mine to confirm. I've got a feeling that you mean bleed resistors rather than shunts. I've taken some photos, which will follow this post.

Here's how I believe it works. Values are approximate because they change too quickly to measure.

1. The voltage in each of the 6 channels is measured. Each channel is connected to each cell through the balance leads

2. If all 6 voltages are below 4.2v, a mosfet opens to allow power through the power leads from the charger into the cell-pack. It will stay open until any channel reaches 4.25v. The power from the power leads charges each cell equally because the current goes through them in series.

3. There is an array of bleed resistors on each channel and a transistor in series with them to create a path to ground through them when it opens, which drains that channel/cell at a rate of something like 100mA. The transistor opens when any cell is greater than 4.2v.

4. When the no channels are over 4.2v, the charge mosfet opens again and the procedure is repeated.

That's it in simple terms, but there are other factors involved, like time.Basically, there are two things happening independently. The charging is allowed if no cell is as high as 4.25v, and the discharging happens whenever a cell is over 4.2v. The charging is at a much higher current than the discharging, so it switches on and off repeatedly at the end of the charge.

The power is controlled by PWM and the voltage is most likely instantaneously higher than what you see. If you watch individual cell voltages during the balance phase at the end of the charge, you will see the voltages jumping up, then draining down every few seconds.

You can see the bleed resistors on the LHS. The transistors are just under the end of the LCD. You can see them in the second photo. The Mosfet is on the other side of the PCB along with the discharge mosfet. The two large resistors are connected to the discharge mosfets, so they must

[Martin McIntosh]

Simply buy a Battery Doctor and it will save all of this fuss. Will balance to 0.01V. I do not even bother to connect the so called balance lead on the B6 types because it is just a waste of time.

[PatMc]

Posted by Dave Hess on 11/09/2018 20:37:47:

Posted by PatMc on 09/09/2018 21:19:34:

How does the CellPro charge through balance leads that are not plugged in ?

I'm not to sure whether I understand your question correctly.

Perhaps you should read Stephen's answer, he understood even though it was his leg that was being pulled.

But first you should read his post that the question was directed at.

[Nigel R]

Not at all Dave, we're having a detailed dig into what we think is going on in these units.

Schematic for a Turnigy Accucel 6 was posted here

https://www.modelflying.co.uk/forums/postings.asp?th=136503&p=1

Manual for one of my chargers is here

https://hobbyking.com/media/file/298640740X176152X1.pdf

which gives exit conditions from the bulk phase (page 10) and the balance current (page 1).

So, what you've posted differs a bit from the manuals I've seen.

Specifically,

The charger does a bulk charge first:

* During this phase it measures total pack voltage.

* Charging current is limited by max current (as set by user), and pack voltage. In effect this means a constant current charge until part way through when it switches to a constant voltage charge.

* Exit condition is usually something like "current has dropped to 10% of user's programmed start value".

All completely standard stuff so far. Prone to a bit of mis-measurement with respect to total pack voltage - bit of a weak point in the design (which could easily be worked around using a better algorithm - the necessary hardware is on the unit). The charger has an internal "guess" as to losses in the leads and internal circuitry and if it gets that wrong the constant voltage can be off, which I think is why you are picking up that packs are overcharged - the charger isn't deliberately overcharging.

Then comes the balancing phase:

* Sit for a few seconds, look at cell voltages.

* Charge current set to whatever the shunt resistors can deal with, unless top cell has drifted over 4.2 in which case no charge current.

* High cell has shunt applied.

* Anything within a few mV of high cell also has shunt applied.

* Sit for half a minute (ish)

* Exit if total spread across all cells has dropped to a few mV, or, timeout has occurred.

* otherwise, rinse and repeat balance

During balancing, you can watch my charger (maybe not all) as it sits idle for a few seconds, while it allows cells to relax and it obtains a reasonably accurate reading of individual cell voltages, then it applies the appropriate shunts and charges for 20 seconds or so (or, does not charge, but drains). Low cells can be observed to jump up when current is applied and top cells come down a fraction. Then a short relax period where they go back to their actual levels.

The current is driven by PWM, you're correct, it just chucks that PWM through an inductor to smooth it out somewhat. The result is averaged out and used to generate an error signal to feed into a diff amp on the input of the PWM circuitry. The lipos don't seem to care much if the current turns up like this in small but intense chunks, its all just coloumbs being poured into a pot

[Dave Hess]

I don't think we're a million miles away. I did mention that there are time factors involved in the algorithms. My comments are only in relation to the balancing phase. Regarding the terminology, I'd call them bleed resistors, not shunts.

I can't see any inductors to smooth the PWM. I can't see a reason to do that anyway because charging lipos doesn't need a smooth voltage. The inductor at the other end of the PCB looks like it's part of the buck converter circuit to raise the 12V high enough to charge a 6S battery - probably something like 25V.

[Nigel R]

Yes. I would say the inductor is doing double duty. The PWM bit, the buck converter is being driven by the desired charge current.

Also your statement 'charge current is much higher than discharge' was what I principally disagree with, desired charge current is set equal to what the shunt can cope with. As it gets delivered via the PWM it will not be a smooth delivery of charge, but on average it will be right.

[Dave Hess]

Posted by Nigel R on 13/09/2018 14:41:15:

Also your statement 'charge current is much higher than discharge' was what I principally disagree with, desired charge current is set equal to what the shunt can cope with. As it gets delivered via the PWM it will not be a smooth delivery of charge, but on average it will be right.

I'm only talking about the discharge during balancing that goes through the bleed resistors. The current will be around 100mA. When you set the charger in discharge mode, it it discharges through those two 0.5 Ohm resistors. They're in series, so that makes a 1 Ohm 5W load, which is confirmed by the specification. It must use PWM to adjust the current through them. It doesn't use those resistors during balancing because they're connected to the two main battery terminals and switched by a mosfet on the back. They can only drain the whole pack voltage, not any individual cell like the bleed resistors can.

[Nigel R]

I'm also only talking about the shunt (or bleed) resistors. Bulk charge and/or discharge is a bit irrelevant to the discussion?

Taking the Accucel 6 (as per that schematic in the other thread), the balance current is 200mA.

If the charger is pumping 200mA into the pack as a whole, then each shunt resistor (which is 6 parallel 120Ohm in the Accucel) can bypass a cell completely and exactly.

That means you can either;

Charge low cells up, by switching in the bleed or shunt resistors around any high cells and set overall charge current to 200mA, thus putting 200mA into low cells and nothing into the high cells.

or

Bring high cells down, by switching in the shunt/bleed resistors around any high cells and applying no charge current, thus draining high cells at 200mA.

It's that simple.

[Dave Hess]

We seem to have the same understanding of how they work. It's only the details of the timing of events that's not clear. The main points seem to be:

1. All cells are charged together.

2. The bleed resistors bleed down the high cells at some point and stop bleeding at another.

From my observations, I can see the charging switching on and off during the balance phase. The cell voltages rise, then drop back down again. This is what I mean by an iterative process, i.e. it doesn't do it in a single step.

As I said earlier, all these type chargers work in a similar way, though some must have better components and algorithms to read the voltages more accurately and home in on the 4.2v. Whether it's the batteries that cause them to get stuck in a loop or something in the charger remains a mystery.

If you want a balance charger that is accurate and can't get stuck in that loop or take all that time to balance, get a BC168 charger. It just goes straight to 4.2v in each channel. It's a fantastic charger - pretty powerful and very quick. This answers some of the questions in that thread referenced above.

The 3-cell charger I got with my RTF Hobbyzone plane does the same, except there's no voltage readout and no adjustment. It is also very quick and seems to be at least as accurate as the iMax type. Not bad for £3.99. To charge from the mains, you need the additional mains to 12v cigar lighter socket converter that you get with their planes.

**LINK**

I haven't tried one myself, but I know several guys that use banks of these to charge their ebike lipos. It looks like they work the same:

**LINK**

There is one caveat though with these chargers that charge through the balance leads. Their accuracy is only as good as they can measure the cell voltages. I know that the BC168 has a separate measuring device for each cell, so any errors in them will give you the same errors in the cell voltages. If you accidentally short any of the pins in the 6S connector, it damages the measuring system so you get false reporting of the cell voltages, i.e. it shows all cells balanced at 4.2v, but one or more could be anything else depending on the amount of damage, so you should check independently from time to time, especially if you ever see sparks. I would add that it's not easy to short those pins unless you do stupid things like I do. I fix and build a lot of ebike batteries, so I often use the BC168 in ways that it wasn't designed for. I'm on my third one now, though the other two are still OK on channels 1 to 3.

[Mike Blandford]

Posted by Dave Hess on 13/09/2018 13:45:57:

I can't see any inductors to smooth the PWM. I can't see a reason to do that anyway because charging lipos doesn't need a smooth voltage. The inductor at the other end of the PCB looks like it's part of the buck converter circuit to raise the 12V high enough to charge a 6S battery - probably something like 25V.

The chargers I have use the inductor in both buck and boost modes, converting the supply voltage to that required to charge the battery. They operate using the charge current to regulate the output voltage that is controlled by the PWM mark-space ratio. The PWM frequency used is sufficiently high that the inductor, with the associated output capacitor, produces quite a constant voltage/current, in the same way any switch mode power supply works.

Mike

[Dave Hess]

Did you see this schematic. It seems to show the output mosfet connected directly to the battery? What's controlling the current if it's not the mosfet?

**LINK**

[Nigel R]

"From my observations, I can see the charging switching on and off during the balance phase. The cell voltages rise, then drop back down again. This is what I mean by an iterative process, i.e. it doesn't do it in a single step"

The charger has to let the cells relax after a short burst of charge/discharge before it can get good readings of the cell voltage. In short, it's meant to go on and off. They only need a few seconds of "off" to get a decent reading, before going again for 20 or 30 seconds with another bout of balancing.

[Nigel R]

"It seems to show the output mosfet connected directly to the battery"

There is a direct connection from the batt negative to negative power in, which I think you have missed? It is by R15, top left of schematic.

The big inductor is the only output connected to batt positive.

[Mike Blandford]

Dave, that is exactly the circuit I expected.VT13, VD4 and L1 form a buck (step down) switch mode regulator, while VT15, VD5 and L1 for a boost (step up) switch mode regulator.

Mike

[Nigel R]

.

Edited By Nigel R on 14/09/2018 17:49:24

[Denis Watkins]

Posted by Nigel R on 14/09/2018 17:48:51:
.

Edited By Nigel R on 14/09/2018 17:49:24

Exactly Nigel, a Testosterone battle excludes us all, he he