EPsolar Charge controller Teardown

In this page, we will have a look at:
LS2024B 20A PWM controller
Tracer4215BN 40A MPPT controller
Tracer1210A 10A Epever MPPT controller

       Up for teardown is an EPsolar LS2024B charge controller.

       I have been using this model for several months in three separate installations and is working great.

       It can work alone but you'll definitely need the MT50 display unit to acces the advanced features of the controller. By default, it is set for a 200Ah sealed battery, light operated mode.

       With the MT50 display unit, you can tweak the various charging voltages, protection cutoffs, timer control, display the reatime clock and it even has data logging for energy both daily, monthly and total usage from PV and to load. As mentioned, I have three of these in three different locations but I only bought one MT50 display unit and just take it out and connect it whenever I need to tweak something. It also prevents unwanted tampering of the settings although it also has password protection of the settings.

Here is the charge controller. I forgot to take a pic of the box but it was a plain brown carton.

Charge controller heatsink at the back. The fins are correctly oriented (vertically).

Removing the four screws pops the top cover off.

The board appears to be conformal coated for moisture protection.

Main MCU. STM32F051K86

Interestingly, I once disconnected this controller from the battery for a few minutes for maintenance and when I reconnected it, the real time clock and settings are all still correct.

The power devices underneath the board are held in place using the steel bar and plastic thingy that protrudes through the holes in the PCB.

Underneath the board, this brass thing is only for mechanical support. The current measuring shunts for both input and output can also be seen.

I'm not sure what the two 8pin devices are for. But they look like MOSFETs.

A side view of the brass support thing.

SOIC14 device is an LM2901 quad comparator.

SOIC8 device is a TC4427 dual mosfet driver

Input protection is a small TVS diode(not shown), a 40A fuse directly soldered to the PCB for main battery protection and an inductive kick back diode protection on the output.

Voltage regulator and RS485 communication interface chip?

PV input switches. IRFB3207
VDSS 75V RDS(on)typ.3.6m max 4.5m ID 180A

Internal temperature sensor under the board.

As with most charge controllers, it is common positive. The PV and load negative side are switched.

Heatsink is cleanly machined and well finished.

Terminal blocks are rising clamp style that can take up to 8ga wire.

I did manage to bend it when I overtightened the screw and the terminal block twisted but it did not break.

Input for the external temperature sensor.

I tested and was able to find out that a 47K thermistor works very well and is accurate if you don't want to or can't buy the external sensor in your place.

       So far, I find this pretty decent built, a lot of useful features and looks like it could last a long time.

Back to top of page

       10 Jan 2015:

       So, I bought a new Tracer4215BN 40A MPPT controller as an upgrade to the home offgrid system and what do you know, there is no warranty sticker! Let's take a peek inside!

Package as received.

The controller is pretty big compared to PWM types.

Here it is on the operating table.


With the back cover off, we are greeted with the PCB.

And some yellow caution tape!

I believe what is under the potting are two large inductors in series.

The wires are nice soft silicone insulated type similar to those used in RCs. Two wires are in parallel so there is just one big inductor in the circuit.

Same as the controller above, the power devices are held in by mounting bars and plastic pins.

I guess most of the parts are on the other side. Here is an IR2110 mosfet driver chip.

Some SMD fets.

Sorry, I decided not to take it apart for now to see what is on the other side. Maybe next time.

The middle of the heatsink houses some large electrolytics.

They did not use the usual plastic housed terminal blocks this time. This looked like wire clamps and some plate soldered to the PCB. This makes a simple, solid, high current connection to the PCB. Much better than terminal blocks that twist and break if you overtighten.

       06 Sept 2015:

       With a new controller installed, now I have time to take this one apart.

Here is the main board removed from the heatsink.

They used a different kind of silpad that appears to not have the reinforcing mesh inside.

The small patch on the upper left is for the heatsink to make thermal contact with a thermistor on the PCB. The square on the lower right is for the shunt resistors and SMD mosfets.

Overview of the main board.

Here is the thermistor mentioned before.

The terminals are nice beefy ones that can accept up to #4 wire.

One side of the TO-220 devices has two MBR30200CT diodes.

The other four are FDP075N15A mosfets.

The other side has five devices. Two FDP075N15A mosfets, two MBR30200CT diodes and a 1BV3P18 mosfet.

Main processor is a STM32F103CB.

This chip is marked RS485. SO I guess it has something to do with the RS485 communication to te remote display.

A 555 IC! and a TC4420V mosfet driver chip.

The row of devices on the top are the shunt resistors. The one on the left are SMD mosfets. The heatsink goop is put there for thermal contact with the heatsink.

This section is an auxiliary powersupply for the internal circuitry.

A TVS diode can be seen just below the LED. It is for the PV input.

The controller is internally fused with two 40A fuses.

Input side mosfets. Notice the labeled test points.

Some power conditioning circuits. I'm assuming these are for the auxiliary supplies for the mosfet drivers.

More test points.

These SMD mosfets are in series with the TO-220 devices on the other side. Gate drive is paralleled.

My unit has a PCB version 7.

The main bulk caps. HV side uses Nippon chemi con caps. The battery side uses Jamicon caps.

Back to top of page

       07 Oct 2015:

       I found out about the new Epever line by EPsolar so I wanted to take a look at the changes from the old. So, I got myself a Tracer1210A 10A MPPT controller on the bench ready for us to have a peek inside.

The box as received.

It is cute. Although it can only do 10A max, I rarely see MPPT controllers this small.

Ruler is 6 inches.

The sticker label.

Back heatsink is cleanly machined.

Removing four screws that hold the plastic cover off. I see fine thread screws.

Popped the cover off.

Metal inserts on the plastic cover. NICE! None of that self tapping crap.

Here is the main PCB.

Looking from the front. I'm not liking the way the inductor wires run around the PCB. This makes a large loop area which can radiate EMI all over the place.

The inductors are potted in their own plastic cups with hard epoxy.

The inductor cups are screwed from the inside. Then potting is poured over so this is mounted permanently.

Another view from the side.

Battery is fused at 25A.

The two green rectangles are shunt resistors.

They skimped on the caps. They used Jamicon caps now.

The design looks reliable though. There are only three electrolytic caps and the whole PCB is conformal coated.

Input side.

There are three TO-220 devices underneath and a complex mounting arrangement of the clamping bar. There are springs and pegs that poke through the PCB to push the power devices to the heatsink.

The LCD is mounted on a sub PCB. This PCB does not have much components, just a Holtek driver chip.

Inductor wires are high temp #14 silicone wires.

The LCD is translucent. I can see a LCD backlight mod in the future.

Behind the LCD PCB is the uC. I'm betting it is an STM device.

Even though the manual and the MT50 manual shows different pinouts for the connections, they are compatible.

It is different from the older Tracer I used. With no PV input, there is no residual voltage on the PV terminals.

This is also different from the usual MPPT controllers as this is common positive. Most MPPT controllers are common negative.

I moved the inductor wires to reduce EMI.

Input cap is 100V. The controller maximum input is also 100V so this is running close to the cap limits.

Output caps are suitably rated at 35V.

It is now connected to my portable power bank for testing.

I have a 40W PV module on the roof connected to this.

       13 Oct 2015:

Modification to add a backlight commences.

Starting with this tiny PCB which contains 2 mosfets, 2 diodes a a few passives.

This controls the backlight so it turns on everytime any button is pressed then stays on for 14 seconds and the light fades out afterwards.

Back side.

Schematic can be found here.

one LED was used for testing.

DIY diffuser using acrylic sheet cut to size then sanded to scatter the light.

It is sanded vertically (in the pic) with medium grit paper. The two notches are where the LEDs will be placed.

The back side is covered in alu tape as a reflector.

2 PLCC LEDs are used on one side to illuminate.

It is spaced a few mm behind the LCD since the guide prongs hit the diffuser panel.

It is hot glued to the LCD pins for attachment.

Here is the finished LCD with backlight.

The hotspots on the right are not that bad with the naked eye.
The sanded acrylic worked well as a diffuser.

Page created and copyright R.Quan © 09 Nov 2014.