@sundberg84 Nice work with that sandwich!! 
Maybe you could move nrf24 to the other board to get away from ac-dc field and obtain better signal exposure.
My idea is to move this node to my home powerswitch and control all my home lights, but as size isn't a problem maybe i'll use this:
http://www.dx.com/p/8-channel-5v-solid-state-relay-module-blue-black-green-250v-2a-213880
Hello everyone,
Since we probably wanting this for multiple destinations, what if we used a centralized way at our home powerswitch.
We could stack multiple boards, and just use one arduino nano and one nRF on top of it. We just have to care about link each stack to a different arduino output.
Based in this idea we could create a second pcb for stack mounting, this board will just have the SSR and a deep switch for easy selectable arduino output. To make this ok we should connect each home powerswitch to each stack, another way exist if we use the main supply of the bottom board for all stacks, and place our circuit before our home power switch, but this is not recommendable.
For minimum stack size SSR should be horizontal mounted.
The advantages are obvious, just one arduino and nRF, one ac-dc and regulator for all home switches. Super low volume occupied and no more size constraints inside wall switches.
I don't have time to go forward and design this, so if someone wants it please go for it!! 
PS - Since we just use one arduino output per stack, we could avoid deep switch cost and configure by soldering selected track.
@axillent said:
tps61222
In the past i was thinking on TPS78233. The idea was solder it on a pro mini board, then i think it could be difficult to solder something too small and quit the idea 
@hek Thank you very much Henrik! I will try it!
Testing v1.4 beta.
I'm already getting a very precise Tension value for Lipo batteries where stopping draining it at 3.2V (at max) could be crucial.
The problem is, tension wise and under load:
3.5V = 3.125 %
3.4V = 1.25%
3.3V = 0.2%
3.2V = 0%
I'm already interpolating for any value of tension (for middle values).
As we can see, it drops and drops fast at the very ending, but for a low discharge scenario as we get with 328p, 1% for a 2000mAh still means many many days of use (maybe months).
As tension calculation occurs before percentage on the node i could and should cut off at the node side based on tension.
But for controller this means many days of use at 0% starting near below 3.38V because actually a percentage like 0.98% is truncated and sent like 0% because we are sending an uint8_t value.
I'm trying to change that in MySensor::sendBatteryLevel changing 8bit integer to something like double/float, but i get compile errors.
Is double/float supported for internal messages?
@clippermiami said:
@BSoft re: Capacitor Bypass. I assume you are talking about the 4.7uFd cap between the Vcc and Gnd on the radio? If so its been installed from the get-go, its part of the design and is within millimeters of the NRF connector.
Yes it is!
Sorry i wasn't aware of 4.7uF presence.
You could add more since it as been reported as better (220µF), check here: link text
Ok, since sma antenna is more power demanding and works ok, maybe capacitor isn't solution.
But i get better than 30 feet on zigzag antenna, maybe you have a not so perfect pcb unit (nRF). It is nice and residue clean?
@clippermiami
You can still use the step-up, connect the capacitor in parallel between VCCout (regulator) and ground.
Or better, connect the capacitor between VCC-GND on the NRF24 (the closest you get to nRF24 is better).
If you still get transmission problems and if possible, bypass the step-up and keep the capacitor on and check again.
@clippermiami said:
I have a good solid 3.3V on the radio Vcc so it isn't a voltage drop problem. I'm wondering if there is a problem with the radio sitting over the PCB ground plane. I have one of the Seeed Studio DevDuino v 2.0 units an the radio also sits over the ground plane on that device. I haven't gotten around to programming it yet so I can't compare the performance.
I think problem could be on the step-up regulator. It as been reported some regulators behave badly on nrf24 peak/burst operation.
On those moments instant current needs are above normal and the VCCout from the regulator oscillates terribly.
As a test, please use the capacitor workaround and check if problem still occurs.
You will have to disable BOD fuses if you want to work below 2.4V.
Please also take in account that using atmega328p below 2.4V will decrease operational frequency (below 8MHz).
More details at this post: link text
I still prefer to avoid any type of regulators.
@a-lurker If brownout detection is disabled we could power atmega328p with 1.9V minimum, but frequency of operation will decrease.
I know you get it right, i just want to standout that there is no problem of operation at low voltage if we know 2 thinks:
I think the big question beside all methods, is none take in account the non linear discharge function..
I mean, considering a Lipo between 4.2V and 3.2V, it's not with 20% at 3.4V (but with 1-1.5% instead).
For a correct percentage value, we have to interpolate the discharge function of a battery.
Another alternative is to measure current but we have to know the initial battery charge (mAh).
Nice work!!! I would avoid the step-up module... for a low discharge scenario CR123 give us almost all juice with their voltage up to 1.9V (nRF24L01 limit).
I think the intent of rPI as a mysensor controller is an easy open customizable/semi-universal database service delivery to cloud, and in that department there is no reason to consider it as a bottleneck. And as a mysensor network controller should not be a problem also, i mean, controlling nodes does not require an horsepower machine.
While server was down, I've used a recent tab from this post. This file represent the mhl backup
Due to server upload limits, removed one photo and zipped the mhl file. It is like it never happened. 
Arduino Library 1_4b1_ Call for beta testers_ MySensors Forum.zip
@BulldogLowell said:
If you are powering with 3.3 volts or higher, It seems to me that you use the on-board voltage regulator on the 3.3V ProMini by simply wiring the power supply in to the RAW pin.
This will give radio power at 3.3V, the problem is if your sensor(s) are 5V. Well, in that case it seems that you simply need to keep the input voltage at 5V+ but no higher than 12V.
No?
I am using the 3.3V ProMini exclusively on my wall-wort powered devices that require 5V in this fashion. I find it much easier than stepping down the voltage for the radio. So far, it is working brilliantly, however I wouldn't know the dynamics of how this would affect a battery power supply.
Yes, if you use the pro mini on-board regulator you have 3.3V to atmega and nRF24L01 but at the cost of regulator power consume. If i remind that regulator have an efficiency of 20 or 30%, it's pretty bad.
And yes, you could at the same time use your battery to power 5V devices (bypassing regulator), but since it's unregulated you have to be careful with the drop voltage of your battery (check if that sensor could work between that voltage range).
Yes, using on-board regulator permits battery voltage between 3.35V and 12V... if you bypass regulator, you could supply even lower voltage, but you will have a max frequency limitation.
Please be careful with one thing you said, if a secondary sensor needs 5V you maybe could not connect a battery of 12V to it, you have to check that sensor datasheet.
It is ok if you connect a 12V battery to pro mini RAW pin because there is a regulator on-board that allows you to do that, but could not be the case with a secondary board.
@axillent I want to use a Lipo between 3.35 and 4.2V. For the atmega i could bypass the regulator, but nRF24L01 does not accept above 3.6V, so i have a problem.
So i'm trying to use atmega output that gives me 3.3V.
My primary goal is minor power consume avoiding all types of regulators.
Thanks for the reply @axillent
What i want is to minimize power consume removing LDO from equation (and neither adding a more efficient one).
From my understand digital output or analog output (PWM) never gets above 3.3V for an ATMega328p 3.3V, please correct me if i'm wrong.
I'm waiting for the material to run some tests.
My real concern is if ATMega could maintain an output while at sleep, i doubt it.
@axillent said:
any li-po are rechargeable and their fully charged voltage is about 4.2V
you still can power promini directly but you will need an LDO to power nf24L01
Could we avoid the LDO by using a pro mini digital/analog output set to High/duty cycle=100%? Maybe with a capacitor for ripple suppression!?
Because i think the máx 40mA per output pin could be ok to power source the nRF24L01+.
And whats the inconvenient if we want to use sleep modes?