Pulse charging Lithiums

How do you store and manage your electricity?
Post Reply
doggy
Groupie
Posts: 80
Joined: Tue, 24 Jul 2018, 09:59
Real Name: David Kerr
Location: Sydney
Contact:

Pulse charging Lithiums

Post by doggy » Thu, 27 Sep 2018, 18:06

I am after some advice.

I purchased 8x180AH CALB cells from SuziAuto. I have built my own BMS and am building my own 24V battery charger which I am linking into my home solar system.

My EVSE is already linked in and tracks the excess solar output and feeds it to the car rather than the grid.

The battery charger will do the same. Now to the question. I have a couple of decent transformers (3,000VA) and am using a MOSFet bridge (utilising the wonderful LT4320 ideal diode bridge). This whacks out around 70amps and is microprocessor controlled by phase shifting the drive to a triac on the primary side.

At the moment, I am planning around 148,000uF of capacitors to reduce the ripple to the battery. But I am still going to have a few volts of ripple- depending upon the internal impedance of the battery. What I would like to know relates to the chemistry of the cells and their internal impedance. I am well aware of the cell charging limit (3.65V in the case of the Calbs) but am wondering if brief ( a few milliseconds) pulses every 10msecs and approx 300mV over the limit will do damage. In other words, at 100Hz (the ripple freq), does the cell chemistry "care" about the brief pulses or only about the average voltage? If no-one knows, I will play safe and keep below the peaks (I have fast 12bit A/Ds so this is quite easy and I'll do it on a per cell basis).

Many thanks,
Dave

T1 Terry
Senior Member
Posts: 730
Joined: Thu, 30 Sep 2010, 20:11
Real Name: Terry Covill
Location: Mannum SA

Re: Pulse charging Lithiums

Post by T1 Terry » Sun, 30 Sep 2018, 14:55

Just be make sure I have this right, you are talking about pluses of 3.95v per cell, 10milli sec on/10milli sec off, is that correct.
If I've got that part right, the 3.95v won't do any harm, as long as you monitor the cell temp and keep it below 50*C while charging. What they don't appreciate is AC ripple were the voltage could swing below 0v, even a 2.5vdc lower limit will age the cell prematurely.
The cell electrolyte will start to heat up once it can no longer accept any current causing the pulses to remain above 3.6v and the climb to over 4vdc can be very rapid at around 0.5CA or higher and 70 amps into a 180Ah cell is getting really close to the figure.
Pulse charging will be fast to the 99.5% SOC point and the Winston cells claim they can be pulsed at 3CA or higher, but I doubt I'd be game to pulse 540 amps into a 180Ah cell, but the rate you are planning to charge at is less than 0.5CA and I'd expect they will handle that without any issues. The cell temperature will tell you if they aren't happy

T1 Terry
Fixed my error with the 3CA and cell Ah capacity
Green but want to learn

doggy
Groupie
Posts: 80
Joined: Tue, 24 Jul 2018, 09:59
Real Name: David Kerr
Location: Sydney
Contact:

Re: Pulse charging Lithiums

Post by doggy » Tue, 02 Oct 2018, 10:08

Hi Terry,
Thanks for that very useful information.
Your assumptions are correct, though the 300mV over-voltage will last more like 3msecs than 10.
I am having to change my voltage regulation because phase shifting was causing too many higher harmonics which heated the transformer cores too much. So I am about to test Integral Cycle Control (i.e. zero current switching and cycle dropping). Given all of this is only likely to be problematic at high SOC where the average charge current is low I think the capacitance I have will be more than enough.
I'll post ongoing info to this thread.
I am proceeding with caution rather than "gung-ho" as I want to avoid any possibility of damage!
Cheers,
Dave

doggy
Groupie
Posts: 80
Joined: Tue, 24 Jul 2018, 09:59
Real Name: David Kerr
Location: Sydney
Contact:

Re: Pulse charging Lithiums

Post by doggy » Tue, 02 Oct 2018, 10:39

Hm, I might have problem with this. One cycle of the mains (ie 20msecs) of a notional 2100watts is 42Joules. Assuming that the cells can take very little due to high SOC that means I am going to need a lot more than 148,000uF of capacitance to keep the voltage peak to an acceptable level. I might have to look at switching in an alternative charging regime just for high SOC.....

User avatar
Richo
Senior Member
Posts: 3362
Joined: Mon, 16 Jun 2008, 00:19
Real Name: Richard
Location: Perth, WA

Re: Pulse charging Lithiums

Post by Richo » Tue, 02 Oct 2018, 12:52

Bridge Rectified mains is 100Hz - 10ms -> 21J.

So mostly these days its a PFC then flyback SMPS.
60kHz -> 16uS -> 0.035J
All the components become much smaller.
Help prevent road rage - get outta my way! Blasphemy is a swear word. Magnetic North is a south Pole.

doggy
Groupie
Posts: 80
Joined: Tue, 24 Jul 2018, 09:59
Real Name: David Kerr
Location: Sydney
Contact:

Re: Pulse charging Lithiums

Post by doggy » Tue, 02 Oct 2018, 19:11

It is 42J because I must have two half cycles not one or else I will have core saturation problems. So I drop or add cycles in 20msec blocks.

You are right about the 60kHz- a lot of things are easier. But some others are harder.

In another day or so I'll know whether I'm continuing on this current path or going the higher freq. route.

Cheers,
Dave

doggy
Groupie
Posts: 80
Joined: Tue, 24 Jul 2018, 09:59
Real Name: David Kerr
Location: Sydney
Contact:

Re: Pulse charging Lithiums

Post by doggy » Fri, 19 Oct 2018, 17:23

I gave up on the triac driven primary regulation.

Harmonics from the cycle dropping were causing too much core heating in the transformers. However, the idea did basically work- but not satisfactory for continuous operation.

The same scheme worked brilliantly for me in regulating a 2khz 3 phase 600volt alternator but there I was just going to 400dc then chopped via magic sines to 240VAC.

So, I ended up building a three phase synchronous buck regulator/charger. Not quite finished yet but testing has been okay.

The heavy duty 24V DC supply using Infineon's "Ideal Bridge Diode" controller works very nicely, delivering 80amps at 29VDC. The FETs barely get warm. However the 2oz, wide copper PCB tracks are not up to the task and I had to build up the source and drain connections with multiple copper wires soldered onto the tracks. Even then, they all get quite hot as do the FET leads. Even though some TO247 FETS have impressive current ratings, I would not like to put it through their leads continuously!

More when the project is complete.

Regards,
Dave

Post Reply