Battery switches off

[dave-at-axon]

I have an A20 with a LiPo battery connected. The battery is 6600mAh rating.

I have a +12V power input to the board for normal use.

I have an external power supply on my system that is used to power sensors. This is +15V output and is generated from the +5V output from the A20 via a step up converter (LM2700)

If I switch this +15V supply on with external power applied, the system works and I get +15V.

If I switch off the external supply and run from the LiPo and now try to switch on the +15V supply, the system powers off and if I check the LiPo output it is ZERO. If I check the LiPo prior to switching on the +15V supply, the battery is reading 4.1V (it's 95% charged during this testing)

If I unplug the LiPo and plug it back in the system powers up from the battery. If I leave the LiPo plugged in and re-apply external power, the system powers up and the LiPo recovers and I can power off the external supply and it runs on battery.

I can't figure out why the battery is going into this off state. I know that they have a protection circuit that triggers if there is a short on the output but I can't see how switching on a +15V supply causes the battery to do this. I have no load on the +15V and the max it would ever have would be 80mA (at +15V) and that would mean all the sensors are at max input, which is very unlikely. In most cases, I would never be drawing more than 40mA or so with the intended design use.

This exact same circuit was used in a previous battery design and it worked fine. In fact, the IC used is designed for battery powered systems that need to generate higher voltages from the battery so it should not be an issue.

I am wondering if this is something in the AXP209 settings that is causing this shutdown. As the manual for this device is all in Chinese it's impossible to know.

By the way, this happens with the use of the Olimex supplied battery or another 2 that I have that are from other sources. It almost leads to the AXP209 operation.

Has anyone seen this before or similar?

[MBR]

Even if the step up converter draws only 40mA, this is current in steady state, the peak inrush current during power up (to charge all the capacitors, especialy the otput one) may be many times larger (but hard to determine withouth memory osciloscope or something similar). Have you tried to connect it with large ferite choke? I read about something similar with USB powered gismos, which turned up overcurrent circuits when pluged in on battery.

[dave-at-axon]

Thanks.

I put a scope on the 3.3V rail and it drops to 2.9 when the +15V supply powered up.

One thing I noticed is that Olimex are using the SY8008C which is a buck or step down regulator. This is fine as long as the input is higher than the output voltage required.

I think the reason my old design worked was because I was using a buck/boost regulator which could handle voltages lower or higher than the input. The A20 design does not lend itself to the battery voltage dropping below 3.3V in this situation.

Not a big deal as I made changes and power the +15V supply from the 3.3V rail instead. I sat with 80mA on this output for 6 hours last night and there was no drop in the 3.3V rail or an increase in the regulator temperature so it should be fine. The final design is actually only going to be 40mA max and most of the time will be around half that. It should be fine.

[dave-at-axon]

I've made a few changes to the design. I've changed out the output capacitor on the switcher for a lower value that's still able to support the current I need.

I also added a high value capacitor to the 5V rail into the FET switch with an inductor between it and the 5V from the A20.

Now the system does not switch off but the LCD touch panel does not like the voltage drop that still appears on the 3.3V rail. The rest of the system seems to be fine with it.

I still feel it has to do with the buck regulator used to generate the 3.3V

I am monitoring the IPSOUT from the AXP209 and on battery this tracks the battery voltage. On external DC or USB this is 5V as expected.

I have posted on the TI tech forums to see if there is a way to have the LM2700 soft start. Some of the other switching regulators have this feature. I may have to change this part of the design with another device that supports this soft start.

[jess]

I have an external power supply on my system that is used to power sensors. This is +15V output and is generated from the +5V output from the A20 via a step up converter (LM2700)

Can you point on the schematic where exactly do you take this +5V from? 
Is it the same as +5V that "powers" the USB socket?

[dave-at-axon]

Yes, it's the same 5V.

I have it working for now in testing by maintaining DC input. I am waiting on a small test board to arrive that has a different power supply IC that has a soft start to reduce the inrush current.

The present issue is that the 3.3V buck regulator does not like the battery dropping to below the 3.3V although the device is capable of working to 100% dropout. The issue for me is that the LCD touch controller does not like the drop as it's min is 3.15V and it drops below this. Everything else seems to work fine.

I'll see how the soft start board works and this should allow my design to work as intended.

[jess]

The issue for me is that the LCD touch controller does not like the drop as it's min is 3.15V and it drops below this. ...

That's the LCD newhavendisplay - correct?

Did you try to obtain your +15V by boosting it directly from LiPo?
That way you can bypass the unreliable olimex dc-dc circuit completely.

[dave-at-axon]

Hi Jess,

The LCD is the 7" Touch Revolution one. I only knew about the Newhaven one yesterday from an email from them.

It wouldn't make any difference coming direct from the LiPo as it would still cause the LiPo to dip and the 3.3V would again go low. I also need to be able to use external power to drive it and going direct to the LiPo would stop this working. I think the softstart IC will work fine. I know next week some time when the board arrives for testing.