PCB Heaven

General Category => General Discussion => Topic started by: cheerio on January 20, 2013, 13:32:17 PM

Title: boost converter problem (TI LMR61428)
Post by: cheerio on January 20, 2013, 13:32:17 PM
Hi guys.
I put together a boost converter circuit using the TI LMR61428. When i use 3.7V input voltage i get 5V output voltage. But when i put a 20Ohm load on it the voltage drops to 3.05V. The 3.7 Power supply does not limit.
Here is what confuses me most. When i use 5V input voltage and apply a 20Ohm load the voltage drops to 4.4V. I measured the Voltage drop from input to output and got 0.12V. When the IC measures the output Voltage is 5V it should not switch anymore and the Voltage should not drop i guess...
 I attach the circuit. Maybe i messed smth up.
Can you imagine what is going wrong?

C4 X7R Ceramic
C3,C5 Ceramic
22uF caps are Tantalum low ESR

The Diode is this one: http://www.diodes.com/datasheets/ds30286.pdf (http://www.diodes.com/datasheets/ds30286.pdf)
The Inductor is the SDR1006-100ML: http://www.bourns.com/pdfs/SDR1006.pdf (http://www.bourns.com/pdfs/SDR1006.pdf)
The TI chip datasheet is here: http://www.ti.com/lit/ds/symlink/lmr61428.pdf (http://www.ti.com/lit/ds/symlink/lmr61428.pdf)

This drives me crazy :(
Title: Re: boost converter problem (TI LMR61428)
Post by: kam on January 20, 2013, 14:10:45 PM
Interesting chip. I'm currently finishing the writeup of a marvelous Microchip chip boost regulator for minimal applications (350mA 3.3/5V outputs) staring from a low 0.8V!

Anyway. My experience dictates that poor PCB design leads to funny results. Worst case is when you prototype the circuit on a breadboard - Breadboards have parasitic capacitance that can cause mayhem.  Where are you prototyping this circuit?

As general guidelines for such supplies:
- Never prototype on breadboards
- Use low switching frequency unless size matters
- Be very careful how you place the feedback wiring. If the inductor is big, then keep it away from the feedback
- If necessary, use a grounding layer around the inductor
- Prefer small and shielded inductors whenever possible
- Absolutely avoid having any sort of wiring underneath (or very close) to the inductor (this includes the bottom layer as well)

The parts you've used are ok according to the datasheet. I can only guess that the PCB violates some of these rules???

I experienced once a totally crazy situation when i was prototyping the A6210. The ON signal was passing close to the inductor. So, when the current was rising, the inductor's field cause severe interference to the ON signal transistor. The funny result was that, as long as the circuit was dimmed LOW, it was working very well. Then,  was dimming the circuit HIGH. Up until one level, the LEDs became brighter and brighter. And then, after about half way or more, the LEDs became dimmer and dimmer! Go figure! The problem was that the inductor provided a false feedback to the ON signal which was larger when the LED current was larger, so the more the current tried to go through the LEDs the larger the feedback signal, and the larger the fb signal caused the A6210 to shut off for longer period.

Title: Re: boost converter problem (TI LMR61428)
Post by: cheerio on January 20, 2013, 14:16:36 PM
Interesting. I prototyped on a breadboard first but that was a pain in the ass. so i soldered it on a card. Maybe i should make a pcb for further testing.
Title: Re: boost converter problem (TI LMR61428)
Post by: kam on January 20, 2013, 14:30:34 PM
Maybe i should make a pcb for further testing.

absolutely
Title: Re: boost converter problem (TI LMR61428)
Post by: George on January 21, 2013, 11:08:46 AM
Kam and Cherio, if you haven't already watched this, I would recommend it.

http://www.youtube.com/watch?feature=player_embedded&v=lrELRVFEx_s

Tutorial and Experiments on Energy Harvesting ICs by Shahriar
Title: Re: boost converter problem (TI LMR61428)
Post by: kam on January 21, 2013, 18:09:51 PM
Kam and Cherio, if you haven't already watched this, I would recommend it.

http://www.youtube.com/watch?feature=player_embedded&v=lrELRVFEx_s

Tutorial and Experiments on Energy Harvesting ICs by Shahriar


20 mV!!!! Are we serious??? Wow thanks George