Hi guys,

I'm designing a combination circuit. One are the brains with the microcontroller and the other part is it's battery charging circuit. For the battery charging circuit, a 18650 Li-ion battery is supposed to be charged by a voltage power of 5V or USB. Charging current is limited to 600mA. The battery charging technique is using the buck converter.

Here is what a came out for the battery charging circuit.

With all the available electronic parts, I end up having to use diodes in parallel to keep voltage drop minimal for higher charging voltage for the battery and also for higher battery voltage range for the device to work during normal use.

Low dropout 3V voltage regulators are also used for their low dropout for the same purpose and two are arranged in parallel for higher current capacity. The other requirement was to use through hole components.

I think it works but multiple and parallel arrangement of these quite don't look good but does it?

*Device start-up is by pulling low the gate of the mosfet before the voltage regulators and kept low by turning on microcontroller pin label BATT_POW.
*For battery charging this pin is pulled low or when microcontroller pin label USB_POW_DETECT is high.

thanks.

 

Diodes in parallel do not share well. That's why they are seldom used.
The diode with the slightly lower forward voltage will carry more current than the others.
This will cause it to heat up more, which further reduces its relative forward drop (forward drop is inversely proportional to temperature), causing to hog even more of the current.
It's somewhat of a runaway condition.

Much better to use a single diode rated for the current you need.

 

For example:

 

For the three diode in parallel configuration, I don't know what to say more. I expected 200mA through each for a forward voltage drop of 0.250mV; according to curve in the datasheet. What causes this, heat?


edit: ok, this curve has indicated the signal is pulsed, not continuous.

 

What causes this, heat?

That graph is for a typical diode. You can't place an order for typical diodes from any supplier.

The fact that the 3 diodes aren't identical means that one will conduct more current then the others...

 

Power is volts times amps. Did you not know that?
Each diode dissipates its forward voltage drop times the current through it (in this case 0.2A * 0.25V = 50mW).

And here from the data sheet is the unit-to-unit difference I mentioned, that will cause a difference in the forward current between two or more in parallel:

 

I was too hasty and already ordered more of it.

@dl324 : You're correct, my 1N5817 was from ON Semiconductor. The curve shows at 600mA, forward voltage drop is 0.3V.

The PMEG3050P indeed would be best for my battery charging application, but I can't find an equivalent through hole part.

At 600mA, forward voltage by the 1N5817 (according to the datasheet) plus VDS by the IRF9540NS which is 0.2V is still higher 0.3V from battery's maximum voltage of 4.2V. There is room for PWM with 4.2/4.5*100=93.3% duty cycle to drop this excess.

But the simulation for a single 1N5817 diode was 380.3429mV!

 

Lithium batteries do nasty things if charged incorrectly. Does your 18650 battery have a built-in safety circuit to protect it against over-charging and over-discharging?

 

Overcharging is detected by measuring the battery voltage level and voltage across a sense resistor.

The plan is also to implement a temperature sensor and a 2 ampere fuse which is not shown in circuit I posted. I searched the expensive ones do have those built-in.

 

Fixed..slimmed down and added protection.

 

Take a look at some of the ideal diode controllers from Linear Technology/

 

One of those application is the LTC4410 - USB Power Manager in ThinSOT. It is a total solution for building exciting brands when battery charging. Thank you but I'm just not there yet to say a lot about the benefits. Hardly forgiving the basics here..hope it'll be smooth.

Maybe regarding my particular circuit, I could post stripboard prototyping development and diagnose..

 

If you are paralleling diodes to reduce voltage drop rather than to get around the current rating limitations then you should have no reliability issues as long as you obtain the voltage drop you desire.

 

Got it, getting the 3V 500mA ROHM BD30KA5FP.

 

BUCK_PWM is 3.3V microcontroller pin PWM with period 20us.
Here is what we got:

No obvious slew for charge and discharge..but some delay

Signal swing before the inductor is +3V and -7V.
Question: so what if there is a dip? And this dip is taking some time..

Next: try to get 600mA through and measure the how much volts will be left for the battery.

 

For the step-down stabilizer, there must necessarily be another diode (or transistor) connected to the ground. It's better if it's a Schottky. Without this diode, the transistor will break through (inductive ejection). I do not know how long the transistor will work in this case (it will be OK).

 

Yes it did happen, the transistor isn't working anymore, and it's not clear what happened to make counter measures.
You must mean to reduce the potential dip and rise effects of the signal before the inductor or isolate the transistor..?

 

My recommendation:

 

For the three diode in parallel configuration, I don't know what to say more. I expected 200mA through each for a forward voltage drop of 0.250mV; according to curve in the datasheet. What causes this, heat?

I assume you meant 0.250 V and not 0.250 mV.

Paralleling diodes to lessen the voltage drop is possible, but due to the exponential nature of the characteristic, it takes an order of magnitude increase in the number of diodes in order to reduce the voltage by 120 mV or so. Though in real diodes this exponential shape transitions to a more linear one as the effective series resistance starts to dominate.

 

Hi,

Thanks Mr. Bordodnynov. I must've troubled you I couldn't understand it the first time and you were still at me with the clear picture I need. That diode will help keep transistor emitter and collector voltage levels more upright.

On my query about diodes in parallel relating to needing increased voltage above the maximum battery levels; before I couldn't help thinking it would be best to have plenty of allowance so can charger voltage levels be determined by making adjustments to the PWM duty cycle on the fly.

However after having building the circuit, a realized some heat with the mosfet. The greater allowance translates to the greater voltage needed to be cut down for more excessive heat by forth effective series resistance. If the circuit with 100% PWM duty cycle could produce 600mA at a load that will drop exactly 4.2 volts or the battery voltage - that'd be perfect. It'd also be "protection built-in".



With my multimeter and 100% PWM duty cycle:
Inductor out was 4.62 volts.
Current was 0.62A.
Voltage across the diode was 0.32V.
Voltage across the mosfet was 0.19V.