Lead acid desulfator circuit

[MrAl]

Hi,

Anyone have a good desulfator circuit handy or a good description of the theory of operation?
I'd like to look into this.
I've worked with chargers for years but not any desulfators and wonder about their effectiveness
as well as operating frequency, current levels, voltages, etc., to get one to work with a
6v LA battery.

[dobrien]

Home Power magazine has an excellent article on a desulfator circuit.

[dacflyer]

is that article inthis month publication ?? i haven't picked mine up yet, i should buy a subscription. buying it at the B&N is getting expencive.
i need to buy me a new battery bank also.. my donated set has finally went out on me.

will desulfaters work on AGM or gel-cells also ?
not sure, but i think they are for wet acid types only...or ?

[dobrien]

The article was in the Issue #77 of the magazine.
This a link to the information in the article.
http://www.alton-moore.net/graphics/desulfator.pdf
Since the crystals are formed from sulphur and lead,
I would suspect AGM and Gell batteries are effected.
I have seen them mentioned for AGM, Gell and SLA
batteries.

[kheston]

I guess my HomePower subscription is too new, I'm only able to download past issues 110-current. So extra thanks for the link, dobrien.

The thing is so simple, one wonders whether it will actually work. I've got a bunch of batteries around I'm going to try it on (standard lead-acid and sealed-lead-acid). I read somewhere that SLA batteries don't respond to such measures, but they're junk anyway so I'll try it.

[MrAl]

Hello,


Thanks much for the link. It looks like an interesting circuit.

I was hoping for some theory to go with it too though, because the question always comes up
as to whether or not these circuits actually work, and there can be two reasons why it
doesnt work if it doesnt work when you try it:

1. The circuit itself doesnt work right.
2. The idea (theory) doesnt work, either on all batteries or just not on one or more types.

The theory behind the desulfator could help us debunk ahead of time, if necessary.

To start i decided to simulate the circuit and see what would happen. If the circuit worked as
the article description then i would proceed.
Sure enough, the circuit did not work as the article explained. The current level did not reach
anywhere near 6 amps, especially if the internal resistance of the 'battery' was increased to
a high enough level to simulate a real LA battery after it refuses to take a charge.
What's more, a 220uH choke can not reach a level of 6 amps in only 50us time (the time
the article quotes) and the actual circuit time is even less than that, nearer to 40us.
This tells me that something surely is wrong. Perhaps the value is incorrect? There apparently
was already one mistake in the cap value for C2, which should be 0.0022uf not 0.022uf.
This is why theory is better than a circuit, because at least if you know the theory you
can design a circuit that will do exactly that.

Anyway, my main question on the theory that might get this to work is just what polarity
'pulse' is required to get the battery to respond correctly. In other words, do we apply
a 'charging' pulse or a 'discharging' pulse?
A charging pulse would take the form of a higher than usual voltage pulse which causes
a higher than normal charge current to flow, and a discharging pulse would take the form
of a low resistance tied across the battery for a very short period of time applied
periodically. Using the theory of electron current flow (normally used for physical processes
like this one) the charging pulse would be a current flow into the negative terminal of the
battery, and a discharging pulse would be a current flow out of the negative terminal of
the battery. The charging pulse would make the battery terminal voltage rise slightly,
while a discharging pulse would make the battery terminal voltage drop slightly.

The other reason i bring this up is because it looks like there are much simpler ways to
get the same result, as well as use the more common N-MOSFET without changing
anything else.

Also, this thing has uC written all over it (chuckle) and of course i am thinking of
going that route.

[psycho]

I looked into this before. I looked at a few designs and decided against it. Although people say that it does work, I got the impression that it takes a lot longer than I thought it would. If I remember right, the guy had a battery on it for 8 weeks and it regained 70% of it's capacity. It would not make sense to me for the $15 battery that I wanted to use it on...

Psycho

[Janitor Tzap]

I agree that it would not be worth messing around with a Lead Acid Battery that was nearly completely used up.

But for say, keeping new batteries in top condition.
Extending the life of the cells. This would be worth at least trying.

As for your tests that you tried on a battery MrAl....

Before you did your test.
Did you check the resistance of the battery first?
Or physically look inside the cells for lead material,
that has built up in the bottom of the cells, effectively shorting the cell?
Once a cell is shorted by this means.
The only thing to do is clean out the cells, and add fresh electrolyte.
But normally when a cell reaches this point, the plates are pretty much completely used up any way.
Thus, it doesn't make sense to mess with a battery that is in this condition.


Signed: Janitor Tzap

[kheston]

MrAl,

Once again I'm cherishing your presence here. You've saved me some time and hassle in your scientific approach to this.

I've got a dozen 12v ~100Ah LA batteries to experiment with and a 10A charger I can rewire into a pulser if you've got a circuit that needs proofing. It will also give me an excuse to buy that $50 load tester I've been coveting on the Harbor Freight site to check results.

[MrAl]

Hi again,


psycho:
Well, i just hated to throw out my nice 10 dollar LA battery because i hardly
used it and it's died already and im guessing it is because of sulfation.
I could be wrong, but then again i wanted to look into this kind of thing
eventually anyway.

Janitor:
The circuit was *simulated* on the computer, not built and tested.

kheston:
Well thank you very much!
I am still looking into the possible circuitry and found some new info too
which follows...

The new circuit i found simply drives a 12v LA with a 10us pulse from a
voltage source that is a voltage doubler with a 2200uf cap. The input is
18vac 60Hz, and that gets doubled and also rectified. The resulting
voltage is used to drive the battery through a P MOSFET. The pulse width
is only 10us wide however, and the frequency is 100Hz. That means one
10us pulse every 0.01 seconds.
Since this new info specifies the kind of pulse and how it is to be applied,
i think i will try this first. I can easily build a circuit to do this so
i'll probably try that first.
If you have a way to build a pulse generator like this you can do the
same, or wait until i can come up with a circuit to do it. Of course
a uC would simply output a 11us pulse to a MOSFET with the MOSFET
connected to a 50vdc power supply and that should do it. I dont know if
a steeper wavefront/wavetrail would help any.
Just to note, i did buy a cheaper load tester when it went on sale for
something like 13 dollars US, and it has a meter and everything on it.
You may find one cheaper if you check again, unless of course they sold
out. I think the normal selling price was only 20 dollars US.
Also just to note, i dont think you need a load tester if you find a resistor
or other load that can discharge the battery in a reasonable time. You can
then clock the process and calculate the Ah capacity and determine if
'desulfating' actual did anything at all or just wasted our time

[MrAl]

Hello again,


Some results now...

Setup:
Built the 11us pulser with a uC, set it for 12us instead so it's a hair longer than 10us.
It is driving a 0.038 ohm (Ron) N-MOSFET. Waveshapes look good.
The power supply is +18vdc.
The battery is a 6v sealed lead acid.

First results:

The most current i can get is 3 amps peak. I suspect that this is because the leads
are too long going to the battery, and the internal cap in the wall wart is not low esr enough.
I am going to have to tighten the loop a bit more before this will pump out 10 amps peak into
the battery. I dont think a low current like 3 amps is going to do anything.

After i shorten all the leads and connect a better cap i'll test again and post some more
results here.

For anyone interested, the program for the uC is very simple because it's just a loop
that turns the output ports on for 12us and then off for 10ms. I am going to use
3 outputs to drive the MOSFET directly through some low value resistors.

[Bigglez]

I looked into this before. I looked at a few designs and decided against it. Although people say that it does work, I got the impression that it takes a lot longer than I thought it would. If I remember right, the guy had a battery on it for 8 weeks and it regained 70% of it's capacity. It would not make sense to me for the $15 battery that I wanted to use it on

From what I've read the process is mechanical, not electrical
(or electronic). The victim battery is bashed by an electrical
current that dislodges the sulphur crystals from the plates.
This is not the same as debris collecting at the bottom of the
cells and causing shorts. The crystals grow everywhere and
effectively reduce the plate area, and therefore the capacity.

The key is to bump the plates by shockwave in a way that
causes them to ring mechanically. Any dampening makes the
process slower and much longer to see results. The design of
the plates, the frames, and the case affect the resonance and
that in turn makes one scheme better than another, or one
brand better than another.

I read a US patent on this topic a while ago but I don't have a
citation handy. It talked about mechanical resonance, that in
turn was initiated by dumping a pulse into the victim battery.
The rise and fall times of that puilse were the real secret to
success.

Perhaps this 'trick' applies to newly installed batteries, as a
means to extend their life. In this mode it may be a gradual
process, than acts to prevent build up (see you dentist for
similar recommendations with a toothbrush).

Also, the applied voltage is much higher than the nominal
battery voltage, as the pulse must punch through the
remaining (unaffected) plate area with enough current to
shake the plates.

I would expect the same results can be obtained by putting
the victim battery on a shaker table. (Meaning one that
moves slightly, not a table made by the Shakers who once
roamed around Sturbridge Village in New England).

[kheston]

I've got an IRF-530 and a BS2E to play with but I'm short a power supply. The closest I have to 50vdc is:

24vdc (2 amp)
30vdc (400ma)

I can parallel 4 of my FLAs but that will surely limit the length of my test before they die. I've got 6 or 7 ~13vdc wall warts that I could run in series and fry, too...

What sort of current will I need? Perhaps a trip to HALTED is in order.

[MrAl]

Hi again,


Well i might be able to answer some of those questions with my next set of results.

I made all the connections much tighter, with lead length to the battery only 6 inches
for the ground lead and 3 or 4 inches for the positive lead. I also connected a large
3300uf cap across the ends of the wires that come from the wall wart to reduce
the inductive drop from the wall wart leads.
BTW i am using a smallish 14vdc wall wart that is rated for only 350ma and puts out
18v unloaded, and since the 10us pulses are hardly a load at all (0.1 percent duty cycle)
the voltage stays around 17 or 18 volts even with high pulse heights.

So after all that i got up to about 10 amp peak pulse height, which is considerably better
than before. This tells me that the lead length and the extra cap are some key points to
get this thing to work at all. Without this attention to wiring detail i dont think it will
ever work even if left on for months, simply because the pulse can not get high enough.
If Bigglez's comments apply (the mechanical vibration aspect) then we really need to
get these kinds of high pulse currents flowing or else nothing measurable will happen.
Thanks to Bigglez for those insights.

Some details to note so far:

1. Low current wall wart with long leads (3 feet, 350ma) is ok as long as there is a nice
size cap at the ends of the wall wart leads. We'll call this cap C1. I used 3300uf and
got very good results.
2. With C1 connected, the leads from the battery (-) to the MOSFET (D) should be
6 inches or under, and from the battery (+) to C1 (+) should be under 6 inches too.
The source lead of the N MOSFET should connect to ground at the same point that
C1 (-) connects to ground. This makes the circular path between the battery, MOSFET,
and C1 equal to 12 inches or less. This ensures a high current pulse with decently
rising wavefront.
3. The drive to the MOSFET may have to be higher than i currently have. I intend to
use three outputs of the PIC chip through three 100 ohm resistors for the drive.
Right now i am only using one output, and getting some decent results, so three should
be even better.
4. Even at 10 amps pulse the MOSFET does not heat up. I am using a 0.038 Ron
N chan MOSFET (standard gate drive voltage type). It appears to be working
ok with only a 5v peak drive voltage, although i can not say if this will work
for all MOSFETs unless they are of the logic gate variety.

Since my last post i have made these changes and left the new circuit connected to the
cell and already i see a difference in the waveform across the battery. It went from
a somewhat slow rise rectangular pulse with a quick fall to a quick rise and a somewhat
ramping pulse and i think it got lower too (unfortunately didnt make that good of a
measurement at first) which indicates that already the internal resistance of the cell
has decreased, at least to some degree, which is what we are after here.

I am guessing now, that the pulse width is not that critical. The power dissipation
vs pulse width is probably the limiting factor. I would bet that 20us would do just
about the same thing.
I see some ringing which tells me that it is working to some degree, but the only
progress indicator i have so far is the change in pulse shape and it's height decrease.
Only charging and discharging will tell me for sure if this is working at all.

[kheston]

Are the batteries gassing at all? My FLAs are sealed (out of a UPS) and I haven't figured out a way to water them. If the pulsing causes lots of bubbles I may need to figure out a watering method pronto.