Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Lead acid batt analyser

Status
Not open for further replies.
Tony I note you got good results with fast rise times, which is probably only possible with the circuit in very close proximity to the battery posts. Otherwise I may need AL or copper bar stock to bring the inductance down over the 1' x 2 cabling distance.
I can look at relocating the power pulse board as a remote mount on top of the battery to cancel the parasitic cabling 3.6uH inductance i have to deal with.
 
yes right over top or side of battery , in a small box, self contained powered by battery ( ~1W avg) activated by alternator.
 
Hi,

This is a good project because if you can prove to the world that this works (or does not work) that would be very worthwhile. Could be one of the most worthwhile projects on the internet.
 
hmm, how about twisting the pulse cables together to reduce net inductance?
The best solution here is to use Litz wire where each strand reduces the inductance L/n in parallel for n strands.

Then twisting the Litz wire to reduce EMI with maybe 6 turns per foot without adding too much capacitance, but effectively reducing CM radiation significantly by reducing the area of the loop.
Characterisitic impedance might be reduced < 200 Ohms which is still much higher than the source and load ESR, but lowers the Q of the switched current spike burst ringing.

The ideal pulse is spread spectrum with harmonics of the repetition rate with a null spectrum = 1/PW50 of pulse width then harmonics of all previous spread spectrum spikes. THen hoping that this covers the resonance of the lead-sulphate crystals. A network analyzer scaled from 50 Ohms to 50 mOHms might display the spectrum of the crystal resonance frequencies among all the lower physical plate resonances and cable resonance up to 50MHz.

Below , the f "Magnitude" humps show the spread spectral harmonics of the pulse rep rate and the nulls correspond to PW50 and then harmonics beyond that. The number of spikes between Frep rate and f null is the ratio of PW50 and Tr rise time.

upload_2015-10-14_12-17-55.png
 
Last edited:
it depends how much IR drop you want and length of wire. but n=280 strands is far better than 1 for both inductance and Skin effect losses.

Rolled Flat braid used as ground wire in big power transformers is good for skin effects and low inductance due to geometry ratio of Length/width ratio, but not as low as Litz wire since each strand is insulated.

In the end one calculates ESR & ESL and Zo to determine best transmission line for a pulse relative to RdsOn and ESR of battery.
Ideally cable and driver ESR and (Xf)=ESL*2pi*L /Tr for rise time Tr should be as low as battery ESR. Usually source and cable is << 10% of load ESR for maximum voltage regulation and matched impedance for maximum power transfer but at 50% loss.
 
Hmm, when we speak of frequency and skin effect in this context; my understanding is that we are referring to the frequency based on the desired Tr (rise time) of the pulse as opposed to a low duty cycle pulse. Thus a 1 µS rise time constitutes a 1Mhz frq (for the litz skin benefits) even if the pulses are 1Hz. Is this a correct position?
 
Yes rise time is the key for triggering harmonics.
e.g. until you determine the desired resonant frequency, you are hammering all of them, then they attenuated in power -3dB power octave ( per harmonic) for 1st order, and twice for 2nd order etc.
upload_2015-10-14_12-44-55.png


I forgot what your pulse looks like but with sharply integrated pulses ( LPF effect and skin effect) So the harmonics now say at 150Mhz are down ~51dB or 20 dB lower in this arbitrary pulse , so you need 20dB more power. (100x)
upload_2015-10-14_12-47-57.png
 
In that case, perhaps this flat braid (with heat shrink applied) satisfies the requirement in reducing inductance and and DC resistance to a minimum.
**broken link removed**

EDIT :
Then again:
This severely contradicts using woven braid for fast pulse currents:
https://www.edaboard.com/thread309730.html
 
Last edited:
The author was comparing 15 to 19AWG coax braid (not all the same) with AWG8 or 10 solid.

Of course this is even better. than both his examples.

"polished large-bore copper tubing, electroplated with high purity electrical grade silver to one or two skin depths, and coated with two-pack epoxy paint for robust environmental protection"

I was thinking something the diameter of welder's cable flattened. when I said "flat braid" your 100A? pulses.

50 Ohm traces have a width/gap ratio around 2:1 for a differential pair as a reference.
200 Ohm twisted pair is typical low current PVC 10 turns per ft. ( +/-50% depending on insulation gap and AWG.
You want much lower transmission line Z for >> 10MHz so width to gap between pairs must be larger. and as large a width/length if not otherwise. In power distribution transformers they use 3ft wide sheet copper for high current windings ~1~2mm th.) To give you an example of width.
 
Ok, I did a few measurements....with a MASTECH 5308, LCR meter on battery pwr.
With no DUT and the probes shorted @ 100Khz we have .04uH read.

With 14 AWG stranded x 1' we have about 0.46uH self inductance @ 100Khz
With 1' of 1/2" x 1/16" thick (wall) AL tubing we have 0.23uH.
With a 1.25" wide s/s ruler @ 1' we have 0.2uH.
With 22" (nearly 2') of dual 14AWG including the battery clip we have 0.57uH
With the same 22" but with 4 x 14 AWG incl the battery clip we have 0.45uH - this is the actual batt connection cable in use.

With these numbers idk if trying for lower inductance with flats or litz will make a lot of difference.
 
Last edited:
This is an example of a commercial strap, but only 1/2 inductance of round cable of same guage.
**broken link removed**

I don't know if you have any transformer plants near you that refurbish, but they recycle a lot of sheet copper and heavy gauge. You might be able to get some scraps.
e,g, "
transformer plant in St Ann, Jamaica
 
I can get sheet brass (26 gauge) and thinner gauge copper sheet from a nearby radiator fab shop.
I just installed a new DS2032, 300Mhz scope and checked the rise time on the pulse amplitude....it's about 80 to 90 nS with a 5V amplitude.
 
Got some interesting recovery results, perhaps even remarkable.:happy:
A 12V 36Ah OEM battery at a rest V of 0.3V, below LED potential, recovered to a battery capable of 380 hot cranking amps with about 15Ah capacity in one day. Useful in the tropics in 1.5L sized vehicles, as we don't suffer cold temp. capacity loss! That batt had been sitting in the weather for a long time.
I suspect it had reversed polarized cells to get the standing V to go that low = discharge abuse. It took about 2.5KWh of energy to be recovered.
 
Got some interesting recovery results, perhaps even remarkable.:happy:
A 12V 36Ah OEM battery at a rest V of 0.3V, below LED potential, recovered to a battery capable of 380 hot cranking amps with about 15Ah capacity in one day. Useful in the tropics in 1.5L sized vehicles, as we don't suffer cold temp. capacity loss! That batt had been sitting in the weather for a long time.
I suspect it had reversed polarized cells to get the standing V to go that low = discharge abuse. It took about 2.5KWh of energy to be recovered.


Half discharge cycles are known in many chemistries to yield more kWh cumulative energy delivery.

So one full discharge cycle at 1C rate is 15Ah*12V=180Wh or 7.2% of energy used to reclaim battery operation.

I wonder how efficacy and recovery power levels are related and how many 100~50% discharge life cycles are left. Or 90~10% or whatever.
I also wonder relationship of how current slew rate (A/s) and rise time (ns) affects energy required to restore s.g. etc.

I estimate hot battery environment >+40'C gives 25~50 more CCA than 0'C but at <<25%~50% of expected life time from faster aging rates.
Arhennius Law affects MTBF above room temp, as I recall of 50% drop in MTBF per 10'C rise.
 
Half discharge cycles are known in many chemistries to yield more kWh cumulative energy delivery.

So one full discharge cycle at 1C rate is 15Ah*12V=180Wh or 7.2% of energy used to reclaim battery operation.
.
Well, I no longer do the 100% discharge cycle as it degrades SLI batteries significantly.
I wonder how efficacy and recovery power levels are related and how many 100~50% discharge life cycles are left. Or 90~10% or whatever.
I also wonder relationship of how current slew rate (A/s) and rise time (ns) affects energy required to restore s.g. etc.
Hopefully my Mphil/PhD efforts will answer those questions.

I estimate hot battery environment >+40'C gives 25~50 more CCA than 0'C but at <<25%~50% of expected life time from faster aging rates.
Arhennius Law affects MTBF above room temp, as I recall of 50% drop in MTBF per 10'C rise.

It's actually 50% per 8.4°C, which means a 2°C global warming = 12% decreased life globally with regional 'el nino' hotspots making it worse.

What's more the US EPA air quality standards have cut the permissible airborne lead from 1.5mg/ cubic meter to 0.15mg, which makes recycling more costly. A lot of recycling from the Caribbean takes place in Puerto Rico which will no doubt abide by the US EPA.
 
A Snubber cap (500uF NP) across battery reduces input pulsewidth by abt 5% for the same avg current into the battery. I wonder why?
no cap:cap-out.png
with cap.cap-in.png
 
If you take the Norton equivalent of your source, the output voltage is always a product of the load impedance and the source.

The pulse ESR ratio affects initial pulse height, RC slope to 100% and thus energy transferred before next transition. Pulse shape is also affected by SRF affects on Fourier and time domain effects.
 
Last edited:
I did an experiment with 2 new marine starter batteries. ACDM27C.
**broken link removed**

Force discharged them to 1.7V...flatlined. This was rest voltage after a week with no load.

Attempted recharge on a modern smart charger, Schumacher 1200CA....failed.
Processed them on the regenerator to obtain OEM 95% CCA and 50% Ah on both after about 24.5 hours - autonomous regeneration with auto stop.

So flat lining a new starter batt effectively sheds 50% of the active plate material.
So if you accidentally do that, you can still have a serviceable batt to regen. but with wear and tear.

Previous experiments showed that forced discharge to 10.5V with a starter batt erodes 10% of its Ah capacity.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top