Effecient Battery charging using Linear Transformer

Hi, i know for max efficiency we go for SMPS. but that's not my field. i have been developing battery chargers for my project based purposes but this time i need to save power.

I would be using SLA battery for my 24/7 powered on NAS.

With linear components like LM317 with bypass transistor set to 15v and 10amps for battery charging i have to provide 18v input. this give 150 watts on output and 30W wastage. Actually it will be more than 30W because there would be ripples also.

My NAS consumes 35W average so the whole system will go more that 75W. Not good.

How about i control the transformer primary winding with a triac, secondary winding goes to a rectifier and Cap for DC. that DC voltage goes to comparator and fed back to triac(to turn it on and off), so we will get 15v stable. as the transformer resistance is low so all voltage drop will be across battery and rectifier, ideally 1.4v. but that will lead to high current sink in the battery and may heat up the transformer and damage the battery. so another current feedback parameter must also be included.

I hope you guys get the picture so i dont have to draw it .

Need your suggestions on this. thanks.

On-off control would be only acceptable if the maximum on-state current doesn't exceed the battery, transformer, rectifier etc. ratings. Otherwise phase angle control is necessary. It must be used we care on the primary side of a transformer. Otherwise DC voltage might be fed to the transformer, causing saturation and high currents.

FvM said:

It must be used we care on the primary side of a transformer. Otherwise DC voltage might be fed to the transformer, causing saturation and high currents.

Click to expand...

I didn't get it. please explain a little.

It means carefully designed phase angle control to give pure AC voltage to the transformer primary.
You'll face the details during design and test process.

In fact, the battery charger, included in my 12V inverter, uses a triac (actually 2 or 3, to cover a wide input voltage range) at the primary side (when mains power is on). I sold about 2000 sets during last year. I mean it is a practical idea.

The first important note is to add a choke (not less than 100uH inductance at 50Hz, with air gap).
Obviously, I use an MCU (actually AT89C2051) to regulate the charging current (actually about 20A) and limit the battery voltage (actually at 14.5 V).

After reaching the voltage upper limit, the charging current is allowed to decrease only, till the battery voltage reaches a lower limit (actually 13.5 V).
In case of a fresh/good battery, this ending current decreases gradually till it becomes equal to the very low internal leakage one of the battery while its voltage stays in the range 13.5-14.5 V.

To monitor the charging current, I use the shunt of an ampere meter (actually, rated at 100A maximum, to also display the inverter current).

Unfortunately, the stability and reliability of this sort of charger depends mostly on the software (MCU firmware).
But perhaps you can find a solution using hardware only. I usually have to include an MCU in every product I design and sell to avoid its exact imitation

Please be free asking any question (and as many as you like), as long it is not a question about the whole design

Kerim

KerimF said:

The first important note is to add a choke (not less than 100uH inductance at 50Hz, with air gap).

Kerim

Click to expand...

You added choke at the primary side to protect transformer from inrush current? i have toroidal cores and lcr meter, i can make one to check.

What triac you used? i have BT136 and BTA16 right now?

Sorry, I forgot mentioning that the choke is in series with the battery so that it smooths the charging current and act as a limiter as well. For the same average charging current, this lets the triac be on for a wider phase with a lower current peak.
I use BTA16 though the amplitude of the primary current pulses may not exceed 6A for an average current of 20A (battery 12V). So BTA12 could be used too.
For instance, I use triac having a higher rated current than necessary so that it can be protected by a fuse in case an error occurs (though this may fail sometimes).

i have setup small transformer, choke not used yet and load it 10ohm power resistor. the LM358 which is acting as comparator switches the triac when capacitor voltage is lower than v-reference. the thing is that as FVM said about high pulsating current, the transformer make a lot of noise and due to high current the transformer gets hot. did you face the same issue?

I didn't face the same issue for two main reasons:
1) the choke at the secondary side increases the conductive phase of the triac hence it lowers considerably the current peaks that occur when only the internal resistance of the battery and transformer limit the charging current.
2) The transformer is designed first for the 1KVA inverter (about 100A DC) so when it is used for charging at 20A DC, the transformer wires don't get real hot.

I forgot adding that also the choke helps in reducing the transformer noise which may be arised by relatively narrow high current pulses.

And since our mains voltage may vary from 170 to 250 V (50Hz) and the battery voltage may be in the range from 9 to 14.5 V , I started my project using 4 triacs (connected to four input terminals). Only one triac is made active at a time (selected by the MCU). For economical reason, I reduce them to 3 then 2 triacs only But for the 24V inverters I had to use 3 triacs as minimum.

For instance, when I started my project, I analysed this charger using LTspice to get an idea on how its currents and voltages are. If you like I will try searching the files in my archive for you.