Powershield Specifications

[martyvis]

I decided to connect up one of the two Powershields I received last week to one of the R4-Moteinos. I was keep to use a single AA, and I tested the Powershield unloaded and was getting just over 5V so all looked good. It ran fine with the "node" code, and was communicating well with a R4-Moteino running the "gateway code".

I thought I would do something a little more serious and so loaded the "Struct_send" code on the battery powered one (and "Struct_recv" on the USB  connected Moteino). The only mod I had made was to turn on radio.setHighPower() (as I waned to a range test with RFM69HW). While it worked, for a few seconds it would then stop. Turning it off and on had the same response. I figured the AVR was browning out and sure enough it it was only receiving a bit above 3V from Powershield. I tried a full charged NiMh (instead of the Alkaline) but this didn't improve things. Both were providing close to 1.4V. I measured the battery current at around 40mA so it was a lot higher than I expected.

If I loaded the standard Ardiono "Blink" sketch the Moteino ran fine, so there was clearly a load related issue.

I removed the Powershield from the Moteino, and though I would do some raw load tests. I ended up trying a few different battery combinations. I have recorded the results here. I also repeated the tests with my 2nd Powershield which I hadn't used at all and have pretty much the same results. I guess my major concern is that there seems no way that the Powershield is going to provide anywhere near the 200mA mentioned. It clearly does provide a voltage boost, but won't provide the rated voltage ( I wanted to run at 5V) at any reasonable load. I'm wondering whether I got a bad bunch, or is there a design issue? I am happy to do some more testing. (BTW I know I don't have a true RMS meter, but I also did some testing by adding a small (10uF) capacitor in sereis to smooth the output and this didn't improve the measurements. The load below are just 1/4W resistors. Batteries were a NiMh AA, 2 x NiMh AAA, and a small LiPo (from a toy helicopter).

VBAT (unloaded)	Vout(unl)	Load	VBAT (loaded)	Vout (loaded)	Ibat
1.40V (1xAA)	5.11V	100ohm	1.38V	2.75V	148mA
1.40V	5.11V	360ohm	1.39V	3.11V	180mA
1.40V	5.11V	1Kohm	1.39V	3.37V	18mA
2.64V(2xAAA)	5.12V	1Kohm	1.62V	3.60V	8.8mA
3.74V(LiPo)	5.12V	1Kohm	3.7V	3.74V	N/A

Oh and as an addendum, after showing my son the LiPo charger, and the LED would light for a minute or so until it detected a full charge, when I then turned on the switch, the blue smoke escaped from the charger chip :-( .The battery is now showing 4.14V open circuit.

Any ideas on what might be going on?

[kobuki]

A quick glance at Figure 1. Maximum Output Current versus Input Voltage in the datasheet reveals that with an input voltage of 1.4V and output voltage of 5V, you can only get about 65-70 mA of max. output current, assuming the TPS61222 part. At an input voltage of 3.7V however, it should provide 200 mA output current, given that an adequate circuit design is implemented and you don't have a dodgy battery. Your last 3 voltage values are off by far, considering the actual, small linear load. Something's definitely out of order here.

[martyvis]

So Felix, any idea about the low output voltage I'm getting even with very light loads? Really wondering if there is a design issue, or I just got two from a bad batch?

[Felix]

I did some tests and found mixed results myself. My mailbox notifier now runs on a PowerShield and has been doing so for the last few months, no issues. The 400mAh lipo that powers it is still at 3.95V after at least a month (it seemed like it was suffering from the cold in february and discharging more rapidly).

On the other hand I've seen the low voltage during some constant load tests.
I'm not sure if the constant load is fair for the booster but I have to say I'm not very happy either.
I've spent some time to redesign the shield and hopefully at least get some answers. Otherwise the fate of TPS61222 is hanging in the balance. I can't find a real reason why the booster would fail to function according to spec. If it means using another one then be it...
But for now I'm waiting for some new PCBs to try out.

[kalbanowski]

This is only half on topic, but I was wondering if you'd had plans for running the R4 directly off the 3.3V boost? I've dithered between power options for low-powered nodes, and used a 3.3V booster myself (SparkFun's TPS61200-based LiPower, using 2xAA primary cells) before the PowerShield showed up.

I still don't have a good sense for how much power is wasted by back-powering an R4 over the 3.3V line without removing the regulator, and the PowerShield clearly isn't set up to supply power to the 3.3V pin; though that 3.3V selector is in there.

[Felix]

The main reason for the PowerShield was to be able to provide 5V where needed (for the HCSR04 and PIR sensors, and possibly others that need 5V).
If 3.3V is needed then using a Lipo or 9V or 3xAAA to the VIN is fine, you can get the 3.3V from the Moteino regulator (for most things that are not power hungry). With low power friendly code, such solutions will still yield many months of service.

Unfortunately there seems to be an issue with the PowerShield, as noted in the posts above. I really can't tell if it's my design (actually used the recommended parts in the DS) or the chip itself is not really up to the job of providing constant voltage at constant load. I'm trying to find a solution but the process is slow. Perhaps the charging chip is causing some weird feedback of some kind and sucks the voltage out of the booster, could not verify that theory yet.

I would not recommend powering the Moteino over the 3.3V although it's possible. The reason is that the regulator is directly connected to the 3.3V rail. So you're pushing current back into the regulator. Things won't blow up if you know what you're doing, but you might waste a little power doing so. Also the RFM radios are directly powered from the 3.3V rail, so any spikes above 3.6V might damage them. The regulator is really the gate keeper when it comes to that safety.