Lead acid batt analyser

[tvtech]

A short short story.

Once I had an idea .
Then I talked to ppl who generally reacted like:
Which made me feel:
But the idea wouldn't go away..
So I did R&D and got some results:
Then some folks doubted me:
I felt:
Others encouraged:
Others advised:
I felt

I have come too far to stop now, ...how will the story end?

Either with success of failure. Nothing in between. And that's it.
Persevere....and nothing less. Don't give up. You have your heart and lot's of your own money invested in this...DONT GIVE UP.

Stick to your guns. Dammit. If you believe see it through. To the bitter end.

Otherwise...a few Years from now...someone will steal your idea...and claim it to be their own....finish the business. NOW.

tvtech

 

[MikeMl]

Finding the bad one is not the problem. Fixing it so it plays well with others seems to be the problem.
Have you tried this?
**broken link removed**
I think this is what Mike & others are talking about.

My experience is that any benefit that seems to derive from "pulsing" a flooded-cell lead-acid automotive starting battery can be achieved more simply by applying a periodic "equalizing charge" as described **broken link removed**Pulsing is just a way of limiting the heating of the battery while overcharging it. Simply using a DC constant-current supply set to deliver ~3A for a 65AH battery prevents self-heating and drives the chemistry to balance state-of-charge in the cells and dissolves sulphates. By necessity, it consumes some of the electrolyte by producing O and H, so can only be done to flooded batteries with caps.

 

[NorthGuy]

3 feet of #12 wire (18" + 18") is 5 mOhm and will drop 3.2 volts at 640A dissipating 2kW of heat!

Have you tried a diode accross the battery to dump the inductive current?

 

[Mosaic]

Actually I use a pair of 12AWG per terminal so we're looking at 2.5 mohm and the pulse duty cycle is around .02. S0 that's 2kW/2 * .02 = 20W. The wires do get up to around 40C.
A diode across the battery only dumps the battery's kickback not the kickback from the 3' of cabling.

MikeMl is reasonably correct, but you cannot push 3A of current with a limited voltage charger (<18V) thru the kinds of heavily sulphated batteries I encounter. That approach is limited to batteries that are in passable shape. Heavily sulphated batteries are more like leaky super capacitors, they respond better to pulses.

Also part of the pulsing technique involves reflex charging benefits (depolarization) not encountered in DC equalizing.

The challenge I am addressing is to reverse a sulphating process of several months in the space of several hours, while not warping the battery's plates due to heating or evaporating the electrolyte due to arcing or reversing the battery cells polarity and causing paste lead shedding. No chemical additives.

I have recovered VRLA batteries as well, not only wet cells. It's a matter of voltage, temperature and recombinant battery chemistry. Gel cells are not recommended due to gas release problems.

 

[NorthGuy]

A diode across the battery only dumps the battery's kickback not the kickback from the 3' of cabling.

You should connect it to the "device" side of lead wires, not to the "battery" side. I think it should alleviate the reverse voltage as well as voltage ringing.

 

[ronv]

The one advantage I could see with pulsing is that the voltage across the battery is higher than with just a 15.5 volt equalization charge. This might serve better to break down the coating.

 

[MikeMl]

The one advantage I could see with pulsing is that the voltage across the battery is higher than with just a 15.5 volt equalization charge. This might serve better to break down the coating.

The constant-current 3A supply I use for equalization is initially set to an open-circuit voltage of 30V. If I connect that to a still-good battery that I think will benefit from equalization, connecting the battery will initially draw the full 3A and it will instantly pull the supply voltage down to ~14V, which then rises over the next few hours north of 16V...

If I connect the supply to a junker, totally dead battery, the voltage may jump all the way to 30V (whatever the supply is set to), and initially the current may be only a few tens of mA. As the oxides break down as the S.G. of the acid increases, the charging current will steadily increase. As the current climbs, the supply comes out of constant-voltage mode and automatically becomes current-limited at 3A, at which point the voltage is dropping fast. If the battery is recoverable, the final voltage will be similar (~16V) to the case described above...

If I see the equalizing current increase as above, there is some hope that the battery is still useable... It takes about 1-2h to find out...

 

[Mosaic]

In essence Mike, I am accomplishing the same thing, hopefuly faster. Think of my approach as a Voltage/current & temperature (cell level) controlled smart charger that uses PWM to achieve its targets as opposed to a linear supply.
A low duty cycle PWM approach permits reflex charging and permits time for the battery chemistry ( ion migration) to react to the energy input. It is related to charge acceptance, a sulphated battery has problems with charge acceptance because of the limited active plate area.

It is likely that your 30V current limited supply will obtain results, how fast though? Can it do a 40-50Ah battery in a day?

 

[MikeMl]

..

It is likely that your 30V current limited supply will obtain results, how fast though? Can it do a 40-50Ah battery in a day?

Mostly, I apply equalization to batteries that are not being used daily. For example, I have a motor home which is not driven in the winter. The batteries (three of them) are constantly maintained with a constant-voltage float charger. This prevents self-discharge. However, since the batteries are stationary, the acid will stratify, and the cells capacity becomes unbalanced. The equalizing charge is aggressive enough that the cells out gas. The weakest cells begin gassing first. The process is terminated when the SG of the strongest cells reaches "full charge", or when they begin gassing, too. Usually takes two to six hours. The main benefit of equalization is twofold: first it stirs the acid (eliminates the stratification) due to the bubbles rising in the acid, and it balances capacity of the cells.

I rarely attempt to "recover" abused batteries because I take care of them as I go. However, I have occasionally had to recover a battery after someone left a car parked with headlights left on, which will not accept charge from a regular charger. The 3A constant-current method will usually takes several hours (up to 10h) to get significant current flowing into the battery, and then about 120% of Ah/3 to finish the job. A 70Ah battery might take 36h...

 

[Mosaic]

I am interested in doing the comparison.
I will build a 0.5A to 3A charger with a 36V limit and do some tests. I have a 338K in a TO3 can around that I can stick onto a 2lb, copper/fan heatsink off a Xeon processor which should do the job.

i'll have to keep track of the temps tho....sometimes batts need a long time at C/40 before ramping current or weaker cells boil.

 

[NorthGuy]

It would be also interesting to look at higher current. In your graph you use 640A. If you use 2% duty, it gives you about 12A average current, which is higher than 3A. Constant current should produce the same charge pressure, but less heat.

 

[Mosaic]

Not so much current NG...the peak current is 640A but it is on a Sharktooth wave profile not a square. The average is substantially less!

On the matter of heat.... pulsing will cause heat spikes which then result in more rapid thermal conduction/convection due to a higher pulsed Delta temp. Thus heat losses to ambient take place faster even though the net heat input is equivalent to the average IxIxR . I have not done the maths to calc the benefit but I believe that's the process.

EDIT: on thinking about that some more...that raises the possibility that the heat pulsing causes pulsed expansion/contraction in the lead grid which can cause cracking in any crystalline lead sulphate revealing good plate material. I'd need to set up a test rig to evaluate that one day.....some microscopy would be called for.

 

[NorthGuy]

Sharktooth should be similar to square wave - if you fold a little piece from the right into the left it'll fill in the gap. Probably won't be full 12A, but definitely higher than 3A.

Heat. I x V. If you have 600A and it rises voltage to 40V, it is 24kW or 480W at 2% duty cycle. Constant 12A with 16V would be only 192W. So, the same average current produces 2.5 times more heat than when pulsing. This is, of course, for square wave, but the effect from sharktooth should be similar. You can do exact calculations with your numbers from the graph.

Not all the current goes to heat. Some goes to charging. This may make the difference even more dramatic. If you assume 100W goes to charging, then numbers change. Now you get 380W left for heating with pulsing and 90W without it. It's 4 times difference in heat.

 

[Mosaic]

One of the values of the pulsing is that it generates hotspots wherever there are weak spots in the cells as opposed to average current charging. Hotspots are early signs of grid corrosion, busbar corrosion or soft shorts.

I have seen a battery look promising at 200A pulsing...and start to arc internally at 600A because the electrolyte was bridging & concealing a crack in the grid. That type of failure never shows up with average current charging.

Looking at the scope trace in post #7.
I count the grid area under the pulse current curve to be 2/3 of a square wave. 16 grid squares rather than 24. Pulsewidth is 40uSec with a 2000uS period = .02DC.
Looking at the voltage differential (yellow) over the pulse the average appears to be 19V. Thus we have .02 * 640 * 2/3 * 19 = 162W delivered.
At the average of about 8.4A supplied (ammeter measured), this time 100% DC => 19 x 8.4 = 159.6W. I'd call that a dead heat given the eyeball errors reading from that chart.

Then there is the matter of pulse charging being more efficient for charge acceptance based on the paper ronv linked to as well, possibly yielding more power to charging rather than heating.

Thus with the pulsed heat pattern the net heating of the battery ought to be less due to extra heat loss to ambient as a result of the DELTA Temp spikes conducting more rapidly. I suppose I could verify that easily enough. Use the same battery with pulsed current and then linear and measure the delta over a fixed time frame in a large water bath of known volume & temp.

 

[NorthGuy]

Looking at the scope trace in post #7.
I count the grid area under the pulse current curve to be 2/3 of a square wave. 16 grid squares rather than 24. Pulsewidth is 40uSec with a 2000uS period = .02DC.
Looking at the voltage differential (yellow) over the pulse the average appears to be 19V. Thus we have .02 * 640 * 2/3 * 19 = 162W delivered.
At the average of about 8.4A supplied (ammeter measured), this time 100% DC => 19 x 8.4 = 159.6W. I'd call that a dead heat given the eyeball errors reading from that chart.

I though it is 10V per square and the voltage is around 40V. So, it's only 19V. It is surprising that voltage increases only that little with 640A current. Have you tried pulses longer than 40us?

At any rate, if you charge with 8A current, the voltage will not go as high as with 640A, will be perhaps 14V or so. 14 x 8.4 = 118W.

Thus with the pulsed heat pattern the net heating of the battery ought to be less due to extra heat loss to ambient as a result of the DELTA Temp spikes conducting more rapidly. I suppose I could verify that easily enough. Use the same battery with pulsed current and then linear and measure the delta over a fixed time frame in a large water bath of known volume & temp.

That would be a good idea. You can also test to see if there are any differences if you keep the ambient temperature at different levels.

 

[Mosaic]

I started recovery of another 55B24L, an SLA version this time, last night.
So far it's looking promising. Was 6.84V @ 7.8 ohms resistance standing. Now it's standing at 12.36 @ 8 milliohms after some processing at 2A avg. I just bumped the avg charge rate to C/10 and cell temperatures are 2C over ambient, voltage is at about 14. That indicates good charge acceptance is happening. The pulses are about 500Amps, 22uS.

 

[Mosaic]

Ok, the SLA battery is back in action at 34 Ah (about 80%) and 700 cranking amps....recovered....even the built in 'eye' shows good SG.

I am beginning to think that the pulse charging is only required at the onset of recovery and once that is done 'regular' equalizing can do the rest. The hi current pulse charging can also validate the battery interconnects in a more rigorous manner where lower fixed currents cannot. Also, it can overcome sulfated cell resistance more effectively than fixed (low) current charging.

Here is a 'scientific' paper ' done by purported 'researchers' with high end equipment debunking pulse charging.....
https://www.batteryvitamin.net/sulfation_pulse_treatment_surprise

I quote a sentence:
Sharp nanosecond/ microsecond pulses produced "pings" or damped sinusoidal oscillations. We examined these oscillations closely. The current wave was 90 degrees out of phase with the voltage wave, due to the resonance of the circuit inductance and the circuit capacitance, showing that virtually all of the pulsing energy ends up being wastefully dissipated in the wiring and circuit resistance, instead of going into the battery.

Well...what a freaking lie!
Kirchoff must be rolling in his grave.
If current is flowing in a circuit in which there is a battery...the same current MUST flow thru the battery.
An optimally built pulser has very low inductance and resistance in the delivery system otherwise much energy turns to heat 'outside' of the battery. My unit certainly heats the battery up quite a bit and I certainly data log AH going into the battery and I certainly see the SG rising. I am also USING submillisecond pulses to do all this.....
These guys focused on the 'ringing' and not on the primary energy carrying pulse it seems.....looks like they missed the forest while looking for the trees.

Then they say the opposite!
We are not suggesting pulsing does not work at all. The average energy in the pulses of an electrically powered pulsing unit can definitely charge a battery. The energy content in inductive and capacitive discharge pulses automatically adjusts to battery volts and amps and that part is a good idea. Pulsing works very well on batteries that have become "tired" and have acquired a relatively mild form of sulfation diffused within their plates, and it works on "open circuit" batteries.

Well...they have a curious definition of 'open circuit'.

It's almost impossible to get reliable data on the internet on this subject. Everybody has an angle and a pitch. Doublespeak!

 

[ronv]

There are people that claim the ringing, like what you see at the end of your waveform, is what breaks down the sulfation. As a matter of fact there are a lot of people selling them. They provide no charge current, only the ringing. I think that is what they are debunking.

https://www.instructables.com/id/Desulfator-for-12V-Car-Batteries-in-an-Altoids-Ti/

 

[nsaspook]

Most solar PWM controllers are effective pulse chargers when out of bulk mode with sufficiently powerful arrays. Most of the scientific testing has been mixed on the end results but my own personal experience has been positive with extending the life of tired batteries.
https://flic.kr/p/a7YLSY

https://www.cleanenergy.com.ph/projects/CBRED/TA RE Manufacturers Sub-Contract/Compendium of References/Solar References/Collection of Solar Standards and Articles/C22 Capacity loss in PV batteries and recovery procedures.pdf

 

[NorthGuy]

Well...what a freaking lie!

They're merely saying that high frequencies (MHz - GHz range) will never make it into the battery becuase the battery along with the wiring form an LC (RC?) filter, so you need lower frequencies (Hz and kHz). Sounds reasonable.

They're not scientific though. They sell battery vitamins.