So, I've been looking into Eagle and making PCBs for various step-up regulators. This is the progress I've made so far. A panelized 5x5cm board with a couple of different designs for three different ICs (Linear LTC3525, Texas Instruments TPS61097, Texas Instruments TPS61221).

The panelizing / tabs and mouse bites were done using the Gerber Panelizer from This Is Not Rocket Science: http://blog.thisisnotrocketscience.nl/projects/pcb-design-tools/
The tool isn't public yet though. So if you want to test the alpha/beta you have to contact them through twitter ask for a copy. Props to Stijn Kuipers for taking the time to explain things to me. Bottom line: Gerbers are trickier than one might suspect at first glance.

The TPS61097 is perhaps not the best IC around but I have a few samples so I thought I'd make use of them. Can't find that chip cheaply on AliExpress either and other sites (was it Lowpowerlabs?) have long since abandoned the TPS61097 due to some issue or another. So, off to smaller footprints. The TPS61221 which was talked about for the official MySensors battery board (TPS61222 was it?) is SC-70-6 unfortunately but CAN be found for about $1 a piece on AliExpress. A pretty good deal. But for now I have some samples I got from TI. Finally, the LTC3525, favored by a lot of hardware tinkerers, is a tad more expensive, around $2.3 on AliExpress. They all require the same number of external components and can work with the pretty much the same values (2x10uF or 1x10uF/1x1uF, 1x10uH) afaik.

Sent the gerbers to DirtyPCBs yesterday, have been updating the order about six times by now Hopefully it will get processed before the Chinese New Year.

Since I have the PCBs that Meanpenguin designed I figured I'd make the first booster PCBs fit the AMS1117 pins so it will hang over the main PCB in an unobtrusive manner.

And yes, among other iffy things, the footprint for the inductor varies (a lot). I went with a couple of different footprints that I though would be hand solderable for a typical 3x3mm inductor (i..e with some of the pad exposed). In reality the inductor I ordered for testing has a horseshoe-like footprint similar to the PCB on the bottom left. The capacitor footprint also purposely differ (0603 and 0805), but the capacitors I have ordered are all 0603 so it is hopefully all good. This is my first real run with Eagle (and the Gerber Panelizer) so it has been a real learning experience for me. I hope that I will be able to reflow PCBs in the future and not have to worry about whether or not the tip of the soldering iron has anywhere to go.

Btw. Other components ordered:

• 10x Linear LTC3525DESC6-3.3
• 100x Capacitor KEMET C0603C105K9PACTU CAP CER 1UF 6.3V 10% X5R 0603
• 200x Capacitor TDK C1608X5R1A106K X5R 10uF 10V 10% 0603
• 50x Do3314-103mlc Inductor 3x3mm 10uH

We'll see if AliExpress delivers as promised. The components have all been dispatched at any rate. Prices were pretty good overall, particularly given that I save shipping and taxes compared to Mouser or whatever. First time I order "serious" brand components from Ali though. Fingers crossed!

I'm hoping that in the future we'll have ready booster PCBs on the official MySensors PCB site for all major ICs obviously. Personally I find the lack of proper low quiescent current regulators to be perhaps the most annoying problem right now.

So, in conclusion I assembled the boost modules. I don't have an oscilloscope or any fancy variable load so I can't do much testing. But both the TPS61221 and LTC3525 seem to be working as intended, they power the NRF24 (not sensor in this particular picture below) without any hiccups at any rate.

I tried some ceramic power resistors to put some load on the regulators and got something like ~60-80mA out of one fairly run down AA cell if I'm not mistaken. Basically according to spec.

All I have to do now is design a compact sensor node where 1xAA actually makes sense. Kind of like LowPowerLabs' Moteino. Feels like it's time to take this to the next level. The only thing bugging me is the need to reflow instead of hand solder, and the massive task of sourcing components from AliExpress so I can avoid Mouse/Digikey. Perhaps I should give it a rest for now and see how the official MySensors modules turn out before I decide.

Well, we looked into this for our local Makerspace. Especially if you're in the US, the Shapeoko or the Xcarve can be somewhat price effective, particularly when they have their sales campaigns and include the dewalt router and free shipping or whatever. But I still think the Xcarve is a bit flimsy and mostly propped up by a successful campaign in social media where they gave away machines to well known youtubers.

In the end we opted for building a rack and pinion RawCNC instead.
http://cncmaskiner.org/
http://rawcnc.com/
Unfortunately the guy behind the project keeps changing his mind constantly about how to sell machines, kits, parts and plans so I don't know how practical it is to start a build right now. It all hinges on finding the parts like the aluminum extrusions.

In the end we didn't want a timing belt machine. And we wanted a vfd / spindle solution from the start not having to go through a dewalt / makita router first. These spindles are quiet and you get standard ER collets. But they are also really heavy and might not work on a standard Shapeoko without reinforcements. We opted for a 2.2kW air cooled model.

Here is someone who has custom built an unusually small RawCNC:
https://openbuilds.com/builds/rawcnc-1-5-desktop-edition.5771/

I just wanted to post some pictures of the boards I got two weeks ago. 12 boards all in all from DirtyPCBs. Sure took its time. Processing time was under a week I think, but shipping was characteristically slow. The silk screen turned out a bit wonky since I exported too many layers (duh!) and it is a bit fuzzy in other places but other than that the boards seem fine. Since it was my first ever order of PCBs I took the liberty to add a big silk screen on the back and round the corners, just to test them out. Clearly not an issue for this fab at least.

The boards are 1.2mm btw. Their standard thickness. But a bit thinner than the "usual" PCB. I like it though. Unless there is a reason to go with thicker boards?

Also pretty happy with Domoticz. Also installed Dashticz on one of my two installations (the other being at the local Makerspace). For me, at least historically, I needed something that could run well on a Raspberry. Domoticz is up and running in seconds obviously even on limited hardware. For me at least Domoticz is pretty flexible, with Blockly, LUA, Python and dzVents. I think there is also some sort of add-on for PHP. The interface and the handling of devices is a bit messy but still.

Haven't tried HA but I would certainly consider it.

@jocke4u I can't claim to have built and tested that many sensors but I have read most of the error reports on the forum over the last couple of months. I see some possibilities here:

1. Power source for the sensor is not ideal ... which could be fixed with a decoupling cap just like you did ... only ...
2. Decoupling capacitor in question is not ideal, perhaps try a larger capacitor, 10 or 22 uF or perhaps try another type, like ceramic. I assume you're using electrolytic like in the store?
3. Perhaps there is something wrong with the radio you're testing, Ebay quality control and all that.
4. Some have reported big differences in reception based on the orientation of the radio.
5. Some have also reported big differences in cable length (shorter cables to the radio or prototype board being better)
6. The NRF24L01+PA+LNA probably consumes more power at the same setting (?) (RF24_PA_LOW/-12dbm) though it should still be within the limits of the Arduino mini pro I guess? If you have the radio connected directly to the arduino I mean, feeding it power through its circuitry? If possible, and you have a 3.3V source, like batteries, you could bypass the Arduino and power the radio directly from the batteries. Perhaps there is something with your board creating an issue like the sensor itself. Like I said, I have no clue here.

@Mrlynx, @korttoma Yeah I guess that would work, and I might just do that if I have an electrician over for some other business. Having an outlet there might make things easier for future appliances that might want to reside there.

But it so happens I have an outlet just some 3 meters away from the cabinet. Safe from the elements. I thought I might make this a central power access point for the whole front of the house, using some sort of 12V outdoor adapter that in turn runs cables to any number of sensors. Shouldn't really matter how many volts are lost in transfer, and then I could finish off with either connecting the RAW pin OR using one of those neat buck converters that can be found cheaply on Ebay. I have a couple of them already and used them to power a sensor node and at least with a decoupling capacitor on the radio they don't seem to be too bad.

Which exact RFM69 is the footprint for? Cause it says "RFM69HCW" in some places, and "RFM69HW" in others. The footprint looks a lot like HW (https://c1.staticflickr.com/1/563/20328459340_62fa380554_b.jpg).

@jendrush said:

@bjornhallberg At this moment i have gateway on Arduino Uno, and nRF24L01+ LNA+PA is working ok. But when i tried nRF24L01+ LNA+PA with nano it didn't work.

Sending didn't work you mean? Or didn't receiving work either? I've only tested receiving for a few minutes using the debug setting, with the nRF24L01+ LNA+PA and Nano connected to my RPi. Seemed to work ok but it'd be nice to know for sure before investing more time with this particular setup. Also, I don't think I could fit an Uno into my camera housing. I'd have to get a 5V->3.3V converter (one that doesn't ruin radio reception) and take the power directly from the PoE splitter (which currently outputs 5V directly to the RPi TP1/TP2 pins).

Just for fun, here is the setup I have right now, don't laugh

Anyhow, the Uno allows for 50mA on 3.3V, while the Nano can do 40mA if I'm reading the specs right? Still way off from any 115mA.

@ServiceXp Some old camera I have. Canon 40D + Canon EF 100mm. Plus some extension tubes. Probably not necessary but they were already attached so I didn't bother to remove them. A decent tripod. Some cfl tubes to light up the pitch dark.

Still, as you can see the focus plane is pretty shallow (f/5.0) and I could have used some sort of light tent to apply more and softer lights. Not to mention plugged in a remote trigger and done mirror lockup to reduce shake. It's just such a hassle to set everything up though .... I should really have a dedicated place where I can have all of this set up permanently.

And above all else, perhaps the biggest failure here and lesson 101 for macro shots, clean the object of interest first

@Denke Any updates to this project? Did you deploy the sensors? Any issues or has it been smooth sailing for the last couple of years?

Is there a suggested replacement for the TSL250? While it can still be found, it seems to be discontinued for the most part.

I've had no luck with the LM393 solution for my new power meter. Just can't tune it to pick up the led impulses. What sort of diode solution is everyone else using? I'm aware of the possibility of IR communication (https://wiki.hal9k.dk/projects/kamstrup) but it seems a little over the top.

While looking into the water meter I realized that we are getting a "smart" meter replacement soon. I'll see if I can speed up the replacement schedule. Anyhow, the replacement is a Kamstrup Multical 21 (https://www.kamstrup.com/en-en/water-solutions/water-meters/multical-21) which has several options for reading:

• Requesting the data via optical interface?
• Enabling the pulse function via optical interface, and then reading the IR pulse at 1imp/10lit.
• Reading the wireless mbus signal? (e.g. https://github.com/weetmuts/wmbusmeters)

@flopp Did you ever manage to stabilize the readings? Am I understanding you correctly that the readings were fine but the transmission got the reading scrambled?

Did you try to find a magnetic field with the meter instead? I have a similar meter and I'm wondering how best to approach the problem. I've seen people hook up Raspberries and cameras to do OCR basically but that seems like way too complex for an easy problem.

@koresh Package went under the "Postnord" radar! Woohoo, perfect! I will start replacing the old capacitors and finally be able to deploy the sensors. Very nice of you to include some product samples as well. Really excited about these sensors!

@alexsh1 I may have made a bit of a mistake with some glitchy chinese test probes. De-glitched cables and connected to a proper lab power supply gave very different readings (with the default git-sketch) ...

Still a bit confused over the differing readings. I get that the board I stripped from regulator and capacitors would perform better though. So do I have to do the same for all the other boards as well? Of course, I could save the worst performing boards for later and run them on 2xAA or something in places where space isn't an issue.

@alexsh1 Thanks for the suggestions! Reassuring to know that you succeeded.

1. I haven't added any sensors so it's just the default ones.
2. https://github.com/EasySensors/ButtonSizeNode ... may need some cleaning up @Koresh or do we use v2 code for v1 sensor as well?
3. I thought as much ... damn. I'm using the CW version radio with the adapter plate. I also thought about the radio but unlike the NRF radio this one should work with no counterfeit units?
4. No my other DMM was out of batteries. Found my old crappy DMM and it says the same, ~2mA.
5. I did remove both caps and it's still at ~2mA.

You're probably right about something not sleeping. Or else I've done a horrible soldering job on the four radios I've attached so far. All four sensors behave pretty much the same.

Also, at least on Domoticz they changed things around now so you have to use V_LEVEL instead of V_LIGHT_LEVEL or what? https://www.domoticz.com/forum/viewtopic.php?t=18577 With V_LIGHT_LEVEL I get some incomprehensible percent value reported.

@yury No I ran them on just a 2032 battery at the time, just had it in an external battery holder. I also removed the LDO on one of the boards and bridged the two top pads, with no apparent effect.

Unless there is something wrong with my DMM.

@Koresh I still have a bunch of v1 boards that I never got to work because of power consumption (I presume). I thought I'd make one more attempt.

I tried bridging the jumper and I'm using a 2032 battery. Consumption varies between 2 and 4mA on different boards. On some boards it fluctuates a lot. The instructions above are not immediately clear. Do I have to remove the LDO AND capacitors as well?

I'm sold on Fusion360 as it is basically professional 3DCAD that can be used for free. Full on parametric with a powerful history tree. As you've no doubt seen it has built-in CAM functionality as well, equally professional and it can output G-CODE for most CNC-controllers. So very much all in one. And you'll be glad you learned real 3DCAD instead of meddling with 123D, Sketchup or anything else.

Personally I prefer the Vectric programs (Aspire) for making G-CODE for the CNC as they are geared towards woodworking with features like v carving. Fusion360 CAM feels more like it was made for mills and lathes but can do most things as well.