The Zener Diode Resistive Balancer

[safe]

The Zener Diode Resistive Balancer

I did a search using "Zener Diode Balancer" and did not come up with any topics about it. This guy has built a simple Zener Diode Balancer for lead acid batteries.

http://a4x4kiwi.blogspot.com/2008/06/battery-balancing-cheap-way.html

Basically this seems to fit into the category of "power wasting" balancers because the stronger cells are simply being made to bleed off their excess in the form of resistive heat.

This might be useful in that it will signal the rider that the "runt" cell is becoming empty because all the stronger cells will be dragged down to the level of the weak one. (you will feel the sag earlier) So it's negative in the power wasting sense, however, if you did this from the beginning it's likely that the wear rates would be at least a little better than naked cells and the price is certainly attractive.

I turn to the experts for advice...

Is this something worth bothering with or a waste of time?

(it's so simple and cheap it's worth at least talking about)

 

[olaf-lampe]

I'm no expert, but I'd use shrinktube over the Zeners to prevent vibration failure.
For lead batteries its not a big problem with overvoltage, but I woudn't use it for Li-Ion cells because of the smaller over voltage area. Temperature drift and voltage tolerances are quite high with zener diodes.

Be careful
Olaf

 

[mikereidis]

I'm not an "expert" yet (give me five years ), but I now have ample experience in destroying/damaging LiMn cells with my manual $5 charger when I get distracted and don't notice cells going over 4.2v


I've been thinking about zener diodes, or something similar for balancing. Esp. now that I have damaged cells that REALLY need the balancing, lest they get worse.

Main issues I saw with zeners are:

(1) limited selection of voltages

(2) no adjustability of voltage to match your current thoughts on max safe/allowed voltage. I'd like to try less than 4.2v since last few 10ths of a volt don't help much, and lower voltage is supposed to increase cycle count.

(3) less expensive zeners are only good for a watt or so, say 250mA at 4v. So you can't charge your pack over a quarter amp when in the CV phase. May be OK for well balanced packs (that don't need zener etc. balancing anyway), but slightly damaged packs could take much longer to charge at 1/4 amp.

Also been thinking of LEDs with diodes in series to select max voltage, but LEDs etc. don't have the sharp curve that zeners do. I'm thinking a zener, a resistor and a 3 legged power semiconductor of some sort might be the minimum needed to do this at decent currents.

Also been thinking of a per cell, PWM like solution that wouldn't waste energy. 1 FET or whatever turns battery connection off, and another FET bypasses that cell so rest of pack gets current. This happens when cell goes over say 4.18v, and goes back to normal when cell goes under 4.15v. Might be some cross oscillation in pack, but I think should be OK.

Been thinking that monitoring and shunting of all cells is not needed if you have one bad-ish "sacrificial cell" that can be monitored and shunted. Works OK if that cell is always the one to reach 4.2v first.

 

[wrobinson0413]

.

 

[olaf-lampe]

@mikereidis
Maybe I misunderstood your bypass_FET_BMS but wouldn't it be charging the remaining cells with the same voltage?

Lets say a 16s pack has 8 cells fully charged ( and each bypassed ), the other 8 cells would be charged with like 60 Volt...

-Olaf

 

[mikereidis]

@mikereidis
Maybe I misunderstood your bypass_FET_BMS but wouldn't it be charging the remaining cells with the same voltage?

Lets say a 16s pack has 8 cells fully charged ( and each bypassed ), the other 8 cells would be charged with like 60 Volt...

-Olaf

Yes, it wouldn't work unless all of the cells have the same bypass circuitry. Because my self-built "$40 power resistor (heater) and bridge rectifier charger" wastes excess power in the form of heat for a "pseudo CC" source, a bypass would increase the current in all the other cells. When that caused one/all of them to raise voltage beyond 4.2v, their individual bypasser circuits would trigger and cause more power to be dissipated in the series wasting/current limiting resistor. Thus my comment about oscillations; which would be a bit weird, but I think relatively stable and of low frequency with sufficient hysterisis between charge and bypass modes.

Monitoring a single "sacrificial cell" would work, but only if it turned power on and off to the whole series pack, which would tend to never fully charge the pack, and take way too long to top-up as much as it can. A small bypass might work though, with the danger of other cells exceeding 4.2 volts during bypass.


I saw the earlier thread about using current limited, voltage adjustable DC-DC converters for charging, and I think that has some merit.

In the end though, I'm thinking a single zener solution is not practical. Perhaps what would be sufficient is some simple voltage regulator part, like a 7805 with adjustment on the ground leg (?), or an LM317T set for max cell voltage for up to 1.5 amps.

So, for example, you have a single 5 cell battery of LiMn you want to charge at up to 1.5A and 4.2v max. I think 5 LM317Ts in series (for input connections) and outputs going to the cells would work. You'd need a few other components like 2 voltage sense resistors.

There are some switching voltage regulator modules and parts that could also be used, and more efficiently, but more complex, more parts, higher cost, etc. When I'm charging I don't really care too much if 50% of power goes to heat (esp, in winter in my garage ), and under 1-2 amps or so, it is still quite common to see linear voltage regulators being used.

Heat from the regulators can be managed with heat sinks, ventilation and fans.

 

[safe]

Charging Only?

I just realized something. (or I suspect something)

This doesn't balance during runtime does it?

The idea is to achieve balanced charging by setting an upper limit for the voltage and when that gets triggered the excess bleeds it's way out.

In my case I already balance during charging (I plug in the wires as parallel and ride in series) so this doesn't seem to be of any value to me at all. :?

Did I get this right?

 

[mikereidis]

Charging Only?

I just realized something. (or I suspect something)

This doesn't balance during runtime does it?

The idea is to achieve balanced charging by setting an upper limit for the voltage and when that gets triggered the excess bleeds it's way out.

In my case I already balance during charging (I plug in the wires as parallel and ride in series) so this doesn't seem to be of any value to me at all. :?

Did I get this right?

Yes, I think you've got this right.

Balancing only makes sense (for the most part) on charging only. When discharging, the only thing you need is to trigger LVC when the lowest cell hits its' minimum voltage: 3.0v or 2.5v or whatever.

I say for the most part, because it would be possible to have a "Uber Controller" that uses relays or somesuch to constantly rearrange cells in series or parallel, in any combination, switching low batteries out of the pack as needed. Not really practical (yet).

 

[Link]

Actually, I doubt rearranging the cells would be a help. The amount of capacity difference should be so low on properly balanced cells that rearranging circuitry would take more energy to run than it would be able to squeeze out of the cells.

 

[mikereidis]

Actually, I doubt rearranging the cells would be a help. The amount of capacity difference should be so low on properly balanced cells that rearranging circuitry would take more energy to run than it would be able to squeeze out of the cells.

But I have too many unbalanced cells from unfortunate charging experiments.

No matter what, in the real world, cells will eventually degrade and some more than others. I'm not going to throw cells or batts in garbage or transition them to other uses just because they have diverged. And yes, it would be nice to charge each cell to 4.15v or whatever I choose precisely to maximize my bike range.

 

[mikereidis]

Charging Only?

I just realized something. (or I suspect something)

This doesn't balance during runtime does it?

The idea is to achieve balanced charging by setting an upper limit for the voltage and when that gets triggered the excess bleeds it's way out.

In my case I already balance during charging (I plug in the wires as parallel and ride in series) so this doesn't seem to be of any value to me at all. :?

Did I get this right?

Yes, I think you've got this right.

Balancing only makes sense (for the most part) on charging only. When discharging, the only thing you need is to trigger LVC when the lowest cell hits its' minimum voltage: 3.0v or 2.5v or whatever.

I say for the most part, because it would be possible to have a "Uber Controller" that uses relays or somesuch to constantly rearrange cells in series or parallel, in any combination, switching low batteries out of the pack as needed. Not really practical (yet).

Wait a minute. Unless you parallel ALL of your cells to charge up to 4 volts or whatever, I presume you are talking about batteries and not cells. Are you concerned about your cells being out of balance ? I am.

What kind of batteries/cells are you running ?

I've been further motivated to build a safe charger by an unfortunate "forgot it was running" manual charger accident tonight that destroyed 480 watt-hours of LiMn. Snap, crackle, pop, but no fire or explosion... :/

I'm currently charging one of my remaining decent (but not perfect) cells with a 7805 5v regulator through a pair of schottky diodes with a resistor on output to pull voltage drop on diodes up enough to have 4.3v (still too much) open circuit. Only 0.5 amps with a cell at 3.6v OC, and 0.3 amps at 4.0v, but it's just an experiment.

I'm going to continue down this linear voltage regulator path to build a safe, balancing charger. Lithium batts are too expensive to charge manually.

 

[safe]

"Runt" Cell Problem

The big problem with any set of cells is that some have more capacity than others. The "runt" cell is the weakest cell (lowest capacity) and even though you might get all the cells equalized as far as charging (which you can do easiest by just charging SLA's in parallel) you still can't prevent the weakest "runt" cell from ending lower than the others.

For chemistries like NiCad or NiMh you have to charge in series anyway and this battery type is less damaged by low charge. (in fact NiCad's even like it)

For SLA when you have a "runt" it just dies so much quicker than the others that if there was a way to equalize on runtime discharge (even if it means resistive wasteful release of energy) you would at least have more notice that it's time to stop. The problem is that the bike "feels" strong when the strong cells are working well even though the "runt" is failing. If all the cells could be dragged down at least then it's easier to diagnose your state of charge by feel.

Anyway... the Zener Diode Resistive method has no value for anything but charge balancing, so it's of no value to me.

I like how cheap and easy it is though.

 

[mikereidis]

Anyway... the Zener Diode Resistive method has no value for anything but charge balancing, so it's of no value to me.

I like how cheap and easy it is though. [/color]

I'd say "how cheap and easy it seems to be".

Build one and get it working well and then we'll see how easy/hard it REALLY is.

I'm going to try something like LM317 linear voltage regulators; that should be do-able with 1 LM317, 1 fixed resistor and 1 trim pot per cell.