LM2679 weird behaviour

[akis]

Has anyone used a LM2679 to help me? I have constructed two circuits so far and am observing a strange behaviour. I can pull about 10-15mA before the output voltage drops as if it's being current limited (but it is not current limited). Maybe I am missing something very obvious. My components are OK, for example my coil is 22uH/4.6A ! Maybe the diodes cannot cope with 260KHz? I have used the SB1100 and SB5100 with exactly the same results.

[c4757p]

Can you probe the switch output pin and see what it's doing? And do you have a part number on that inductor?

[akis]

The switch is displaying a complex 100-260KHz wave when I am pulling very little current (less than 10mA). When I pull more (so that the output voltage drops) it becomes a complex squarish wave.

The coil is a panasonic http://uk.farnell.com/panasonic/elc16b220l/choke-22uh-4-6a-0r031/dp/1749170?ref=lookahead
I just noticed the coil has a resonant frequency at 10KHz. Maybe this is something of importance ?

The diode is an SB1100 and later an SB5100 (no difference made).

[c4757p]

The coil's the problem. First thing - that 10kHz "resonant frequency" is a misprint by Farnell, that is just the test frequency (the frequency used by the inductance meter used to measure the coil).

It's hard to comment much on the coil itself, because that datasheet lists screw-all data, but a "choke" coil is not meant for energy storage, it's meant for filtering. Your inductor is probably saturating. It's tricky to directly measure the inductor current in a buck configuration, but look for evidence of a fast, huge current spike right before the 2679 switches off (distortion before the falling edge of the waveform, a drop in the input power supply voltage measured on the V_IN pin also right before the edge, etc).

[akis]

How can I be saturating the coil at 15mA ? I have lots of coil cores I can use to make my own coil, I will wind on a much larger (heavier) core to see what happens.

[c4757p]

The pulse currents are much higher than 15mA. Be careful - heavy doesn't imply a high saturation current. You can air gap it - cut out a small sliver, or break it and glue it back together tightly.

[akis]

I just replaced the coil with a huge one: RM12 core N87 at 27uH. Result : exactly same behaviour, voltage output drops drastically as soon as I try to pull anything over 10mA.  I also air gapped it, same as before, no change.

I have used the RM12 with N87/N97 to pass about 5W at 200KHz in another application, and it's been perfectly fine.

I cannot understand what is happening here.

[c4757p]

Huh. Double check that you've calculated the current limit resistor properly. And post more info! All the component values, an actual scope photo/screen capture, maybe a picture of the circuit...

[akis]

Input cap is 100mF/63V, ESR=123mOhms. Later added 470uF/63V, ESR=15-20mOhms (no difference).

Output cap is 100uF/35V, ESR=150mOhms, later added 470uF/35V, ESR=30mOhms (no difference).

Coil is 22uH/4.6A - later changed for 27uH on RM12/N8 core - no difference.

Softstart cap is 10nF, boost cap is 10nF.

Feedback potentiometer is 1K to ground (over which the 1.21V is being developed) and a 10K on the top leg, so we get 10 * 1.21 + 1 * 1.21 = 13.3V. I do get this voltage, and this voltage remains stable when I vary the input voltage from 15V to 30V, so something is working right.

Diode (my main suspect) is SB1100 and SB5100 (no difference between the two). Maybe the diodes are to blame? I have a box of other diodes, not sure what parameter is so crucial to this.

The current limiting resistor seems to have an effect. On paper it should be about 6K for 6A which is vastly more than I am trying to use. I have used values as low as 100R, and this allows a bit more current to be pulled from the output without the voltage dropping so badly, so it may be that the current limiting mechanism is at fault ?

The schematic is the standard out of the spec sheet. I will take some pics out of the switch pin on the scope and will post.

[akis]

Here are screenouts of the switch pin under loads, input voltage is 21V, output voltage is 13.3V, loads are 30R, 100R and 470R and as you can see the straight line is what we get at the output (more or less)

[wraper]

How much voltage drops under load? For example, how much voltage remains under 100mA load. Did you try to change boost cap? Are you sure it's right value?

[Circuitous]

I have used this before, about a year ago.  These are layout sensitive, see the guidance in the documentation.
Also, are you building this on a breadboard? 

I got it to just barely work on a breadboard, I had to build it on a PCB to get it to work under even modest load.

[akis]

Here is a better scope output: input 21V, output nominal 13.2V but drops to less than 3V on modest load (less than 20mA).

It is built twice, once with point to point soldering and the other on a prototype board (soldered) using a new LM2679.

I would expect non-optimal components to reduce the efficiency or produce some ripple at the output, but not to drop the voltage from 13.2V to 2.8V !

[Ton]

Something is triggering the current limit, the on time is way to short, it should be around 2.2usec when in continues mode with a bit of load on it, according to scope pictures it is not much longer than 0.5usec at 30ohm load,this load is also approximately where this converter should go into continuous mode.

When the output voltage is below target, then it is only a over current situation that can turnoff the switch. (Or reach of max duty cycle)

It is clearly stated in the data sheet, that in continued over current situations the controller will do frequency fold back, ,typical to belov 100khz, this corresponds with you scope pictures.

Your coil might not be saturating, But some thing is triggering the over current protection - could be noise, since you have build it on a protoboard, your ground might not be good enough and causing spikes to trigger over current protection.

According to datasheet your output capacitor should be close to 47uf when having a output voltage bigger than approximately 12volt or a load step could trigger an over current situation.

So try to improve the current paths, add wire or cobber foil, and reduce the output capacitor back to 100uf or less,  and maybe use a Rset of 4k7, smaller than this doesn't make sense according to the data sheet.

Regarding to your diode both should work, sb1100 is a bit small with 1A current rating but this is not the reason for The behaviour we see here.

At work I have a similar converter running at 370khz with a 5A schotky without any issues
16 to 24 volt in, 12v out at Max 4A controller is LM22677 very similar to lm2679

Hope you get it is to behave soon   

[akis]

What you say sounds spot on, I also suspect something wrong with the current limit circuitry. But when I am pulling 15mA, enough to cause the drop the voltage, the total consumption would not be more than 20mA from the source, and there are simply not enough currents anywhere to cause spikes and significant voltage drops over thin wires.

But I will try to improve the ground and see what happens. Maybe I will be obliged to build it on a proper PCB after all.

[akis]

OK, found it. It was a bad solder connection on the current limit resistor. I am so sorry to have wasted everyone's time.

[Ton]

  no worries
it was fun remote debugging a circuit