[solved]WARNING. here be dragons. Stress testing a single cell with a DIY transformer

[brandnewb]

WARNING. here be dragons. Playing with mains voltage is not recommended and could be fatal if done wrong!

I really could use some pointers here reaching my main goal of getting 3.6VDC at 140 amps.

Background is that I have a few 3.2V 280Ah Lifepo4 prismatic cells and before I get more of those I'd like to see if they can really charge at .5C i.e. 140 amps as their specs state.

The reason I'd like to see that is because I am planning to build such an overkill in power generating stuff around the house that my (future) battery bank should never become a bottleneck. It would be a shame if I purchase many more of those cells only to find out later down the line that the seller has been untruthful regarding their capabilities.
Another reason is simply to see if the seller is to be trusted so I can also purchase other items from them.

I'd like to start testing a single cell so that I do not damage all of them if something goes wrong.

I have considered using a variac to transform my mains 230VAC to 3.6VAC and then using a bridge rectifier to convert it to DC. The problem with that is that the variac's that I can afford are well below the 140 amps output goal.

I have considered an arc welding transformer but I was informed that those drop the amps once a bridge has been formed or something to that extend. I did not really understand what was told to me.

As I did not find any other things to try I landed on a DIY solution.

My first attempt was a misguided one when I still had even less knowledge about the subject than I have now. It was a rather small 3d printed transformer core using ferromagnetic filament.


On the primary side there are 63 winds of 1mm enameled copper wire. On the secondary side a single wind of 8mm (50mm2) wire.

This does not work and leads to the circuit breaker tripping. The material used for the core is not magnetically permeable enough I guess and the winds on the primary side are far too few.

So now I'll try again with a rather large (is it or should I go larger? Soon enough I can print really big objects with a huge diameter) torodial core using an iron filled resin I am going to make.
below is one half of the mold to hold the resin.


The resin will be heat resistant and loaded with iron dust, but not so much so that it becomes electrically conductive. It will be close to the magnetic permeability as regular iron would be I recon although I am nto sure how to test that yet. If anyone knows I would love to learn that.

Then the primary coil will consist of 630 winds of 1mm enameled copper wire and the secondary coil will consist of 10 turns of 8mm wire.

Shall I simply hook it up to the mains power and see what happens or should I first get a variac and see what happens when I apply a much lower voltage to the primary coil first and then gradually ramp up the voltage if things seem to work.

Does anyone have any other suggestions on how to get 3.6VDC at 140 amps?

 

[brandnewb]

I forgot to mention that I am not a math wizard and am a bad reader of technical documentation so I would like to avoid diving into theory.

I am hoping that oversizing things can lead to victory rather than calculating things ;(

 

[brandnewb]

also I am not sure how to limit the amps to max 140 as, if my calculations are correct, I will have too much if the transformer works.
If we ignore the losses then 230VAC at 16 amps transformed to 3.6VAC will lead to 1022.22 amps.

I guess I can't avoid doing at least some calculations. If only the formula is correct

 

[DJSmiley]

3.6V AC is 3.6x SQRT(2)) = About 5V top-top when rectified. Also, highly recommend to do some smoothing at least by adding a capacitor. General rule is 1000uF / A, so that 'requires' some decent capacitors.

The 280Ah cells should be able to handle 140A charge fine, but you need to use proper wire gauge cables, otherwise the wires and lugs will act as resistor and limit the current, even while the cells could handle more.

Lastly, test at 20-25c ambient temperature. On lower temps, cells can't handle these high charging currents. For most LFP cells, at 5 celcius, its only like 0.15C (thus 40A or so), and 0.3C at 10 celcius (85A).

0.5C is generally not within specs when cell temperature is <15 celcius or so. (Check datasheet for exact details)

Inrush current of huge transformers can be big, that's why there are multiple options to limit that. Easiest is using a NTC in series with the primary coil.
If you have a variac you can use that to slowy increase the amps, magnetizing the core so the current remains limited while ramping up

 

[brandnewb]

thank you for the heads up

 

[brandnewb]

so you see no reason not to try my plan?

 

[efficientPV]

I have an adjustable 7V 150A power supply with current limiting from the old days of 5V TTL electronics. I remember when give these away for shipping. Everything comes around eventually.

 

[brandnewb]

I have an adjustable 7V 150A power supply with current limiting from the old days of 5V TTL electronics. I remember when give these away for shipping. Everything comes around eventually.

I am not sure what you meant with " I remember when give these away for shipping. Everything comes around eventually."
can you please elaborate?

 

[efficientPV]

I think faster than I type, should say....... I remember when you couldn't give these away for shipping. Nobody wanted these. Two years ago I bought a $2500 Xantrex 7.5V 80A for $40. Nothing better than Xantrex power supplies.

 

[brandnewb]

Easiest is using a NTC in series with the primary coil.

Are you refering to an NTC thermistor? I could not find anything else using that abbreviation.

 

[Hedges]

I don't have actual experience with LiFePO4 cells, just contribute my electrical understanding to threads here. Others are familiar with their behavior.
I was going to caution that 0.5C charge rate was only allowed around nominal 25 degrees C, but Smiley already brought that up.

Transformers, variacs, and arc welders - there I can help.

To charge a LiFePO4 battery, it would be best to have electronic regulation of voltage. Without that, the cell will hold voltage down while accepting current, and watching it charge is like watching grass grow. But suddenly at the end voltage shoots up past target 3.65V to 4.5V and the cell is wrecked. At the very least, you could have a circuit that keeps supply connected to cell by holding a relay closed, and lets relay open when a target voltage is crossed. Then stays open, doesn't automatically close again.

As Efficient said, you can buy a used supply.

I tried using my arc welder as a current transformer for a circuit-breaker tester. Electric heaters as resistors in AC line to set a current flow, current transformer to boost current, drop voltage, with secondary shorted by circuit breaker under test. Arc welder has 3:1 windings. Problem was, lots of leakage inductance (that's how it avoids tripping breaker when you strike welding rod to workpiece), which dropped voltage across heaters in half and limited current. I instead used toroid transformers. I was able to slow-trip 20A breakers, also deliver > 100A for fast-trip.

Inrush current of huge transformers can be big, that's why there are multiple options to limit that. Easiest is using a NTC in series with the primary coil.
If you have a variac you can use that to slowy increase the amps, magnetizing the core so the current remains limited while ramping up

That would be if magnetization of the core is what causes the inrush. In Newb's case, I think it is primarily lack of a core.
"What a terrible thing it is to lose one's core, or not to have a core" (At least that's what I remember Dan Quayle saying)

A coil of wire in the air has inductance. To the extent windings are close together, L is proportional to N^2. If coils are widely spaced so not interacting, more like proportional to N.
With windings around a core, inductance is multiplied by relative permeability of core material. The core will also couple windings that are separated from each other, by making a circuit that carries the magnetism (which is why I suggested pole pieces for your generator.)
One problem is, above some magnetic field, magnetic domains of core are all fully aligned and it can't respond to stronger field. Above that, the coil only feels the effect of an air core.

You need enough inductance to limit current so magnetic field doesn't reach saturation. For the transformers I've worked with, saturation occurs at about 1/100th the current the transformer (or common-mode choke) is designed for. You want to make a 140A transformer (secondary winding), so you need a source for about 1.4A, which the variac should do. (Above that current, the second winding of the transformer carries current in opposite direction, cancelling about 99% of the magnetic field.)

To test an inductor/choke/transformer, ideally you have a oscilloscope with at least 2 channels. And, an AC source of variable amplitude. We're playing with 50/60 Hz here, so variac is handy. If RF components, then a higher frequency signal generator. If no variac, use step-down transformer and resistance in series to drop the voltage. If using a variac, note that it is an auto-transformer, NOT isolated from line voltage.
You will want a source of AC voltage as high as the voltage you plan to use the transformer with. But start lower, especially since you know you're tripping breakers.

Your transformer is meant to take in 220Vrms 50 Hz and put out something like 3.6Vrms 140A, 504W. So primary only carries 2.3A; a variac could drive the primary.

I'm presently testing and writing code to build models of chokes at work, so while I'm not the best Academic, I have very practical experience in this.

See the hysteresis loop of B-H field:

Hysteresis - Wikipedia

en.wikipedia.org

That's in terms of magnetic parameters most of us aren't familiar with. But I get the same shape from measured voltage and current with a scope. I'm using a LeCroy Waverunner 8208HD with CP031 current probe. If you've got $55,000 to spare, makes a good setup.

Teledyne LeCroy - WaveRunner 8000HD 8 Channel High Definition Oscilloscopes

Teledyne LeCroy is a leading provider of oscilloscopes, protocol analyzers and related test and measurement solutions that enable companies across a wide range of industries to design and test electronic devices of all types.

teledynelecroy.com

Teledyne LeCroy - CP031

Teledyne LeCroy is a leading provider of oscilloscopes, protocol analyzers and related test and measurement solutions that enable companies across a wide range of industries to design and test electronic devices of all types.

teledynelecroy.com

To do the work with less money, any 2-channel oscilloscope probe and a sense resistor for current measurement. A differential probe could deal with ground-referenced signals, but I suggest using an isolation transformer on output of variac; that way scope ground clamp doesn't connect to grid neutral (or hot!) and it is safer for the operator as well. I used a step-down transformer, 120V to 9V, so a larger range of the variac could be used for the voltage range of interest. My DUT is a common-mode choke not power transformer, so rather than line voltage across each winding, I find about 1.5V (around 1% of line-to-line voltage) is what the coil can hold off before going into saturation.

Observe the red trace, coil current, which starts out as a low sine wave but shoots up near the peak. That is what happens as core starts to saturate (approaches tails of hysteresis curve.)

https://www.allaboutcircuits.com/textbook/alternating-current/chpt-9/mutual-inductance-and-basic-operation/

The magnetic flux (which you can't measure without other instruments) can be computed as integral of coil voltage. If an inductor, you only have one coil and there is some V = IR drop. I was measuring chokes/transformers, so I put scope probe on an undriven winding. Because the LeCroy scope had waveform math, I was able set F1 = integral(channel 1) and see the result (unit Webers) on the scope. I was also able to pop up another window and graph Webers vs. Amps, which showed the hysteresis curve.

From BH curve, you can determine Hc, Br, Bs (see curve in link below), which are the core properties and allow modeling. Except, you have electrical measurements and have to scale them to core properties considering turn count, cross-sectional area of core, and magnetic path length (centerline) of core.

L. Inductor

If you can download voltage and current traces, you can perform integration and graph X vs. Y with any math program. I used Matlab. You could use Excel. At first, my integral had a slope to it and B-H curve looked like a spirograph. I was able to de-trend the data and it improved considerably.

Traditionally, this measurement is performed with an analog scope. Integration is accomplished with a resistor and capacitor circuit, which of course requires values scaled to the size signal you're dealing with.


In other words, if you put your wind turbine project on the shelf for about year and dive in, you might be able to wrap your head around magnetic core modeling. At least, use a scope or DMM to observe current waveform or ratio of current to applied voltage, make sure you will be able to operate with margin above intended voltage without saturation (the point where current shoots for the moon.) Or, just buy a used transformer for this application. (better yet a power supply with CV/CC operation)

The following eevblog was my starting point. It has link to MIT video of hysteresis curve on an oscilloscope. I tweeked the LTSpice model to operate at 60 Hz rather than 10 kHz. Most of my effort has been on Matlab code to process noisy data captured by scope and extract Hc, Br, Bs. They are easy to determine graphically, but doing that with software, making it work regardless of where the traces begin on the curves, is more difficult.

Arbitrary (saturable) coupled inductors in LTSpice - Page 1

Arbitrary (saturable) coupled inductors in LTSpice - Page 1

eevblog.com

Fortunately for me, I'm bringing home a paycheck while I figure out what I need to know and how to measure it. A basic coupled inductor model of transformer wasn't giving me what I needed, so I had to determine the hysteresis curve (not provided on data sheet) and model it in SPICE to predict performance in system.

 

[brandnewb]

wow, thanks a lot for the in depth contribution. It will take me some time though to get through it

 

[brandnewb]

2kw, 8a 0-250VAC variac is on it's way to me as we speak. An isolation transformer I did not find yet. Can anyone please give an example of one so that i can search better?

 

[brandnewb]

as for oscilloscope I am considering the following;

ScopeFun - Open Source Oscilloscope

Open Source Oscilloscope, Waveform Generator, Spectrum Analyzer, Logic Analyzer, Pattern Generator. Cross Platform - Software runs on Windows, Linux and Mac.

www.scopefun.com

would that be any good?

 

[Hedges]

"Saleae" is a brand used by some "big" companies

Saleae Logic Analyzers

Effortlessly decode protocols like SPI, I2C, Serial, and many more. Leverage community created analyzers or build your own low-level or high-level protocol analyzer.

www.saleae.com

Have to stay within voltage limits. Get suitable probes to attenuate voltage.

 

[brandnewb]

sweet, found a 220VAC to 12VAC isolation transformer. it's on it's way.

 

[Hedges]

Here's one I got for a different purpose at work, then used for step-down and isolation between Variac and the chokes I was testing (to see B-H curve and hysteresis). It has dual primaries for 120/240V, and secondaries for 9/18V

https://www.hammfg.com/electronics/transformers/power/1182.pdf

Variac like this

VEVOR Transformer Variable 1000va Ac Voltage Regulator 110v 1kva 60hz Pro | VEVOR US

Discover VEVOR Transformer Variable 1000va Ac Voltage Regulator 110v 1kva 60hz Pro, Pure Copper Coil and 0 - 130V Adjustable at lowest price, 2days delivery, 30days returns.

www.vevor.com

We happen to have a very fancy and expensive digital scope with AC/DC clamp probe, but I figure with isolation transformer, a sense resistor would work to measure current. Also, resistor limits current, so when inductor/choke/transformer under test saturates, current doesn't shoot up so fast.

Obviously if your DUT has only one winding, the two scope probes need to have ground clamp on same node, connection between sense resistor and winding. Other end of resistor gives current, other end of winding gives voltage.

I've now got my Matlab code extracting core Hc, Br, Bs parameters from digitized voltage and current waveforms. The Bs parameter isn't working as well; multiple estimates from multiple cycles of input data give varying results. LTSpice simulation with the parameters reproduces Hc and Br within 10%, but Bs is off by 50%.

 

[brandnewb]

I am a bit confused ;( I am reading up on how isolation transformers work while waiting for it to arrive and for what I can tell the core I have constructed earlier (the photo I have in the OP) is already an isolation transformer core is it?

It consists of 2 coils that are electrically isolated from each other. Or am I missing some of the finer nuances?

 

[brandnewb]

Also can someone please confirm if I tested the iron powder for conductivity correctly.

I used the setting that beeps when I touch the probes together or put them on something electrically conductive.

When I place the probes on the iron powder really close together it does not beep still showing OL as if the iron powder is not conductive.

If that is correct then I will upsize my first core attempt and make it hollow and fill it to the brim with this iron powder.




The variac has arrived in the meantime. still in the box so I'll show it off once it is unpacked and ready to rock

 

[RV10flyer]

Or you could just let someone with one or two dc power supplies do the test safely. My two will only go up to 80A at 3.6VDC.