Hello all,

First time poster and occasional lurker here, I have a question about developing a 24V battery charger. While I know buying a battery charger is cheaper than building one, unfortunately that is not an option. Here is what I have so far, I want to charge 2x12V 35Ah batteries that are connected together in series. These batteries are used to power an ATV remote control rover I am building for use in desert conditions. Here is an attachment of the battery manufacturers data sheet. http://www.staabbattery.com/mm5/graphics/00000001/pdf/UB12350-V.pdf Also I have read the information from http://personal.atl.bellsouth.net/t/h/thib9564/GellCell_Battery_Charger.htm and http://www.batteryuniversity.com/partone-13.htm Knowing that are three stages to charging a lead battery I have developed a bulk charger that operates at 28.8V. I would like to make the charger intelligent enough to have temperature compensation and float charging, but I can't use an all in one IC that has these features built in. Here is a schematic of the circuit I have built thus far. Note that the diode bridge is just a generic rectifier and not exactly the one pictured here. Also I am using an LM338 voltage regulator as it has a 5 amp output. Thanks for any suggestions you all can provide.

 

Can you use a uC? Obviously you are going to have to use a thermometer to gauge temperature, then adjust the offset. This can be done via a uC with a lookup table, or algorithm.

Also, the charging time for the batteries will be considerable with < 5A available to the batteries.

If charging time isn't an issue for you, then there is no problem. But it is something you should consider up front.

Also, what is your location? You may want to put it in your profile so we know what parts may or may not be available to you.

 

Hello,

Thanks for the reply. I updated my location in my profile. I can use a micro-controller and I have programmed uCs in the past, but I was hoping to use a mostly analogue solution as I would have to get a programming board. Also I chose the Lm338k because it had the highest amperage output for a voltage regulator that I could find. Any alternatives are welcome. I would like to minimize the charging time as much as possible, however I have only been able to obtain a 120/24V transformer with a 4A output thus far. (Weight may become an issue if the xfmr was too large for example). Once the rover is complete it will be used by others who won't have any idea about how it works, they would just plug it in to charge the batteries so I am worried about using strictly a bulk charging method as they may inadvertently overcharge the batteries. While temperature compensation may not strictly be necessary it is a welcome addition based on the outdoor nature of the project. Thanks for all suggestions.

 

You don't want to use a single 5A regulator, as you won't be able to get the heat away from it fast enough, particularly in El Paso.

You could use an LM317 regulator driving multiple 2N3055 transistors in a TO-3 case, which have much better thermal transfer characteristics than the TO220 package.

Your temps in the summer might average around 110°F, which is around 43°C. An LM338 in a TO220 package has a thermal resistance of 4°C per Watt of power dissipation. The power dissipation will be at it's highest when the batteries are most heavily discharged at high temperatures.

Having the 4A max capacity is actually a good thing, as it will reduce the thermal rise of the battery mass during charging. The higher the charge current, the more quickly the battery temperature will rise. It takes a lead-acid battery a LONG time to cool down once heated.

You could use an LM34 or LM35 with the body thermally connected to the battery + terminal to monitor the internal temperature.

Lead acid batteries have a negative temperature coefficient of roughly 3mV*#cells/°C; in the case of your situation, you will have two 6-cell batteries, for a total of 12 cells.
That means -36mV per 1°C.

Clear as mud?

 

Thanks for the great info. I Think using k prefix for the lm338 was incorrect in my case. I have a NTE 935 which I believe is equivalent to the lm338 and is in a to-3 package

 

While I know buying a battery charger is cheaper than building one, unfortunately that is not an option.

Why is it not an option to buy one?

 

No, you were correct - if it's an LM338K, then it is in a TO-3 package.
I simply assumed you were talking about an LM338T, which is in a TO-220 package. My bad; I didn't bother to look at your schematic before I started rambling.

Good for you for attaching the datasheets. It would have helped a lot if I'd actually looked at the battery sheet before I started posting. My bad again.

Note that they do show the battery tempco for cycle and float; but they're only giving constant voltage charge voltages. I'd charge them at a constant current until they hit towards the high side of the constant voltage charge (adjusted for temp, of course) and then float them at the middle of the standby range.

 

Why is it not an option to buy one?

It's a requirement as other subsystems have purchased parts such as sensors: gps, digital compass, etc. Sorry my posts are getting shorter as I am on my phone, but thanks for all the great advice.

 

The 1N5402 diodes you've used in your bridge are rated for 3A continuous. While you might say that since they'll only be conducting less than 50% of the time for a 4A output transformer, therefore they should be fine.... it could be better.

When your peak current is over the average power rating, the Vf of the diode is rather high; in this case around 0.9v. Since you're using a FW bridge configuration, you lose a total of 1.8v across the two diodes; around 7.2W total at 4A.

What is the AC output voltage of the transformer when there is no load on it?

 

The 1N5402 diodes you've used in your bridge are rated for 3A continuous. While you might say that since they'll only be conducting less than 50% of the time for a 4A output transformer, therefore they should be fine.... it could be better.

When your peak current is over the average power rating, the Vf of the diode is rather high; in this case around 0.9v. Since you're using a FW bridge configuration, you lose a total of 1.8v across the two diodes; around 7.2W total at 4A.

What is the AC output voltage of the transformer when there is no load on it?

Actually I'm not using 1N5402 Diodes mentioned, I am using the RadioShack 8A/400V Full-Wave Bridge Rectifier Model:276-1181. I just happened to have one lying around so I used it, but I didn't see a model for this part in my part database so I just made up a rectifier using the 1N5402.

For AC output voltage do you want the voltage directly from the transformer output or after it passes the rectifier? Directly from the transformer I was getting about 24.3V with 116.6 Vac input. I noticed that over the course of an hour as I was testing the transformer and working on a breadboard that the mains ac voltage fluctuated from 116.6 to 117.3 Vac with a corresponding change in the transformer output.

Again Thanks for all the great info.

 

OK about the rectifier.

I wanted the no-load voltage, either peak-peak or RMS value, at the transformer itself.

What voltage are you reading at the filter cap(s), both with and without a load?

You DID say that the transformer was rated for 4A @ 24v, right?

 

Just FYI, here is a table of cycle (bulk charge) and float voltage for a single battery over a range of temperatures. This is according to the values in the datasheet for your specific battery.