[SOLVED] Unexpected behaviour of LM2596

[zeeekay]

Hi all, I am using DC-DC converter LM2596 in buck mode to power up charging interface of a GSM module (SIM900D). The standard charging current of this module is around 650 mA . I am providing it 5.5 V (from LM2596) as Vchg. The regulator is perfectly providing required current when difference between input and output voltage is lower (upto Vin = 8 V). However, upon increasing input voltage , the o/p current starts decreasing (o/p voltage stays constant). unfortunately, I have not access to oscilloscope here, however it seems to me that greater input to output voltage difference increases ripple voltages. I am using following values for capacitors and inductors :

Cin = 220uF / 50V
L = 330uH
Cout = 330uF/25V


I have tried changing regulator and other components. Still, that did'nt bring any luck.
I will need to power up all circuity by 24 V in my application. Any help will be highly appreciated!

 

[alexan_e]

Your inductor value is extremely high, it is about 10 times the value that I see in the datasheet for 15v input, how have you selected this value?

- - - Updated - - -

What exactly is your input range?
You said 5.5v output and up to 8v in and then you mention 24v.

 

[zeeekay]

Your inductor value is extremely high, it is about 10 times the value that I see in the datasheet for 15v input, how have you selected this value?

- - - Updated - - -

What exactly is your input range?
You said 5.5v output and up to 8v in and then you mention 24v.

Thanks for quick reply. According to figure 12 in datasheet, for higher E*T ratio, we need higher inductance. The E*T ratio increases with increase in input/output ratio. I have chosen this value considering higher input-output voltage difference. Also, in reference design (SIM900 application note), they have chosen same inductance value for Vbat supply design.

updated :

The input range is 8V to about 26 Volts.

 

[alexan_e]

That didn't answer my question about the input range and the ET value is calculated based on that.

For example is the input is 24v and Vd=0.5v and Vout=5.5v then ET=29.7 , that gives a value of 100uH/68uH/47uH depending on the max output current

 

[zeeekay]

alexen, I am getting your point. The inductor value is about 3 times greater than what it should be. Unfortunately, I dont have any lesser value available right now.

The peculiar thing in all this is . according to the DC-DC converter equation, the higher inductor value will just decrease the ripple current. How does it relate to the decrement in output current with increase in input voltage. The most confusing thing is output voltage remains same. Just the output current decreases.

 

[mtwieg]

So the output voltage isn't changing, but the current delivered to the GSM module is dropping? That's strange, but it doesn't sound like a problem with the buck regulator. It's only supposed to regulate output voltage, not output current. It's certainly not a problem with the inductor value...

Are you sure the output voltage isn't changing at all? And 5.5V sounds like a strange voltage for a charger. Are you sure it's not supposed to be closer to 5V?

 

[zeeekay]

So the output voltage isn't changing, but the current delivered to the GSM module is dropping? That's strange, but it doesn't sound like a problem with the buck regulator. It's only supposed to regulate output voltage, not output current. It's certainly not a problem with the inductor value...

Are you sure the output voltage isn't changing at all? And 5.5V sounds like a strange voltage for a charger. Are you sure it's not supposed to be closer to 5V?

I have been thinking on same lines. Hence, I changed my load and connected four alphanumeric LCDs backlights in parallel. I connected it to my external power supply with 5.2 volts. It was consuming about 800 mA.

Next, I connected backlights with LM2596 board and set its voltage to 5.2 volts. When Vin of LM2596 is closer to 8V, the regulator is easily providing 800 mA. However, upon increasing the input voltage, output current starts decreasing. For 15.5 volts (upper limit of my external supply), LM2596 is providing about 300 mA on same load. its very peculiar for me. Above all, the output voltage stays same (5.2 Volts which I set).

- - - Updated - - -

That didn't answer my question about the input range and the ET value is calculated based on that.

For example is the input is 24v and Vd=0.5v and Vout=5.5v then ET=29.7 , that gives a value of 100uH/68uH/47uH depending on the max output current

I searched out a 100 uH inductor from some old circuit. I replaced 300 uH with 100 uH.. Still same results

 

[alexan_e]

The two things you mention contradict each other.
If you have a 800mA load @5v then to have 300mA the voltage must drop to less than 2v, there is no way to have a 5v/800mA circuit being fed with 5v (DC) and having less current, the voltage must drop.
The only justification I can give is wrong voltage measurement or wrong current measurement.

Have you used good quality capacitors proper for the HF switching operation?

Edit: actually what I described was for a resistive load but in any case the voltage must drop in a level that makes the load need less current.

 

[crutschow]

Sounds like you are measuring the input current, not the output current. The input current of a switching regulator will normally decrease with an increase in input voltage for a fixed load, keeping the input power relatively constant. That's the reason for using a switching regulator.

 

[mtwieg]

Yeah I'm also suspecting the OP might be looking at input current instead of output... not much other explanation.

 

[zeeekay]

Thank you very much friends... I myself figured out that it was only input current which was decreasing with increase in input voltage.
I just came on forum to inform others and found out that crutschow and mtwieg had already pointed out it...
I am wondering now.. How could I have missed that much important point. Anyway... Thanks again!

 

[crutschow]

...........................
I just came on forum to inform others and found out that crutschow and mtwieg had already pointed out it...
I am wondering now.. How could I have missed that much important point. Anyway... Thanks again!

I've done similar things in my various circuit designs. I will discover something that seems to defy the laws of electronics or physics and scratch my head, wondering how that could be. Invariably it's an error in my understanding or my measurements of the circuit. ;-)