Hi all,

I have 2 problems,

1) how to set up the simulation correctly

2) why the prototype (and the simulation) of the charger does not go into trickle mode.

I used this SgtWookie's batt_SLA model, modified for my battery:

http://forum.allaboutcircuits.com/attachment.php?attachmentid=24235&d=1288998265

I took this circuit of a a battery charger from an old EDN article about February 4, 198X.
It was advertised as:

"Battery charger snaps from full to trickle" by Kennan C. Herrick ESI Electronics Corp San Francisco, CA

I first built it and I was not satisfied that it did not go into trickle mode at the end.

I then decided to simulate it as a ltspice model to try to understand what was the problem.

1) Problem:
I am not sure how to set the .pulse or tran or anything else that could:

-charge the battery from 6V to 7.25V to see why the Q2 transistor does not switch off and stay off until some lower voltage.
Q2 should give hysteresis to the whole process.

Q1 is there to limit the charging current to 130mA for the fist part of the charge and that works fine on the prototype.

On the prototype the charge reaches 6.85V at the battery with I=15mA and Q2 Vbe of 0.6V min. then it stays there forever.

As I understood Q2 should switch off and a little trickle current go mainly through R7.

The lead acid battery is a Gel 6V 1.2A.

ILOAD is there with the idea of discharging the battery first before observing how the charge would go.

On the model the max V battery is 6.28V with max I charge of 90mA.

Q2 Vbe is min 0.9V, max 1.06V.

There are some differences with the prototype but the main thing is Q2 does not go off as in the prototype.

2) Problem:
Why Q2 does not switch off ?

Thanks for all help.

 

Post deleted.

 

The model clamps the battery voltage at 2.133 volts times the number of cells. You can change this in the sub by modifying VCELL to a higher voltage.

 

Of the 130mA, only ~30mA goes through R7 but ~100mA is flowing through the Q2 base-emitter junction! That can't be right?
Replace the battery model with a 100mF (that's 'm', not 'u') cap and run the sim for 10 sec (not 80000) and you'll see that Q2 does switch off, with hysteresis.

 

Well, I thought that VCELL was V Float = 90% of Vbattery which is 2.4V per Cell,

Is that what is supposed to be set to ?

 

Of the 130mA, only ~30mA goes through R7 but ~100mA is flowing through the Q2 base-emitter junction! That can't be right?
Replace the battery model with a 100mF (that's 'm', not 'u') cap and run the sim for 10 sec (not 80000) and you'll see that Q2 does switch off, with hysteresis.

According to the original article, 100mA on the base is right with TIP32 .

I tried your suggestion and went bigger with C2 Cap value and timing.
I added a further Pulse step to see the hysteresis.

Can you confirm it is the delay from 20k to 24k on the Ib(Q2) line after the 20 mA load step ?

V(out)-V(vreg) is Vbe(Q2) light blue trace.

 

Yes, the 20-24k section is where Q2 turns off.

 

Looking at your Vbatt I would try 2.66 volts.

 

Yes, the 20-24k section is where Q2 turns off.

I assume you mean it was off from 11k and turns on at 24k after a delay from 20K ?

I now would like to test it with the battery model.

Any suggestions ?

Here I tried with some differences:
Vin = 10.5V
Vout = 8.6V
R7 33 ohm
R4 2k
R5 4k7

VCELL = 2.4V

I can't get Q2 to go off nor any hysteresis.

Is the batt_SLA.sub a linear model ?

I don't understand the model at all.

Would anyone give an explanation of the syntax of the section called:

* The charge model

inside batt_SLA.sub ?

Or at least a link to a manual that describe that kind of working ?

http://www.4shared.com/download/j3cdxWrk/Picture_3.png?tsid=20131016-142620-745d81f1
http://www.4shared.com/download/VoF-wNa6/Picture_4.png?tsid=20131016-142646-8d5fadbd

 

Looking at your Vbatt I would try 2.66 volts.

Ok, with:

VCELL = 2.66V

it's a bit better, what is your reasoning ?

http://www.4shared.com/download/j3cdxWrk/Picture_3.png?tsid=20131016-142620-745d81f1

 

"Analog Behavioral Modeling", we'll see...

maybe it will be useful for other seekers.

http://www.ecircuitcenter.com/Circuits/ABM1/ABM1.htm

 

VCELL = 2.66V

it's a bit better, what is your reasoning ?

The model clamps the votage at 3XVCELL. so if that is lower than your applied voltage the current will not drop enough to kick you into trickle mode,

 

The model clamps the votage at 3XVCELL. so if that is lower than your applied voltage the current will not drop enough to kick you into trickle mode,

here are the batt_sla.sub parameters as I set them:

.PARAM CELLS=3
.PARAM VCELL=2.3
.PARAM CAPAH=1.2
.PARAM R_SER=50m
.PARAM SOC=.9
.PARAM CHEFF=0.75
.PARAM SELFDC=0.000028
.PARAM TEMPC=25

is VCELL the nominal Lead Batt Voltage per Cell ?

should that not be 2.4V ?


Thanks

 

I don't know exactly how the model works either. Maybe SgtWookie can explain it better. I just had the same problem you had where the charge current wouldn't go down low enough because of the clamp voltage. I suppose you could set it higher, I just don't know if there are any side effects.