final build

When you've got it working, and decide that is all done (if that ever happens), How do you assemble/package it? Do you solder the Arduino to another pcb or strip board/whatever, with the other components, or do you connect it by loose wiring and stuff it in a box? If you use a pcb, do you use sockets so the Arduino can be later recovered/replaced, do you mill out a rectangle so Arduino pcb is flush with the base pcb? I guess if it was a production case, you'd use the Atmel chips on a purpose built pcb. I'm thinking of Nano type boards which come with header pins (not attached) and maybe nRF 24L's which have pins pre-soldered.

All depends what IT is.

In one project I am following two paths. There will be several boards (6 or 7) built with Atmega 328s on stripboard and two Mega clones used for control purposes - because of all the I/O pins and extra serial ports.

Connections to the Megas will be soldered to right-angled header pins that plug into the Mega sockets. I believe they will be sufficiently reliable and they make it easy to modify things.

I don't have any experience designing PCBs and I reckon I would waste far more time and money learning and doing that compared to the time taken to build the circuits on stripboard,


I usually get a pcb made and then when I’m making add an outline for the arduino I am using and then solder appropriate pin headers on the pcb so that I can just sandwich the two together using the pinheaders...

Thanks for your thoughts. I will hopefully be able to show what I have done in the next few days, once I've corrected my soldering errors. I never had these problems ten years ago...

I never had these problems ten years ago...

Or you were young and wild and just didn't notice :slight_smile:


... do you connect it by loose wiring and stuff it in a box?

I've done it that way!

I almost remember young, but never wild, perhaps a tad annoyed.
Chris - whats the purpose of the card toilet roll centre? Or was it originally a full roll in your design?

Ha! It keeps the Arduino and battery pack from sliding around in the box. When I close the lid, everything is squeezed into place.

Must have been some complex calculations to get the right pressure - just enough to keep in place, but not enough to blow open the lid. Maybe could be a sort of steam punk for electronics,

Here’s what I’ve done. The nrf24’s plug in. Having run down some of my electronicals, I used an 8 pin dil socket, cut in half. Makes a nice tight connection, since the header pins are ‘fatter’ than those on a dil chip… The nano has just the needed pins for this application, not yet soldered. It will fit in a standard plastic box, with either one or two battery packs on the back. The row of headers is for 5v , led, Dallas or other temperature unit, and 3V supply if needed for the nRF units. Room for something else, connecting wires direct to Nano.
Although I’ve pcb etching equipment, for a few boards, it’s a bit of a faff, so I milled these with an 0.8mm cutter. Had fun since the dxf cad files wouldn’t play ball, and the holes needed pocketing, not drilling. Still, got there in the end. When I was doing the PIc stuff, must be ten years ago now, I standardised on 10k and 1k resistors, even surface mount. One of the 5k resistors is 2 10k in parallel. One board has 5 soldering errors (shorted tracks mainly)
and one pin unsoldered, the other board has one shorted track.

Image from Reply #9 so we don't have to download it. See this Simple Image Guide


Thanks for posting image. I found it a right pita even to get it linked. Got so used to drag and drop.
Anyway, pride b4 fall, nothing works now. But it will, because it did before. It's down to my soldering, or most likely lack of it. It didn't help that I moved a few i/o's around and fogot to alter the software, and tried programming the Nano as a Uno...

I'm going to have to start again from the basic example of Robin2's simpletx/rx. It worked well on the Uno's. I've carried out a few iterations on the tx side, for handling the temperature sensor, and since changed it to the Nano, pcb as above, etc. Now it seems to have failed. Not sure if I've fried the nRF 24's or just bad connections or whatever. But, I'm also not sure about handling the transmission of the temperature values, since I thought it needs to be fixed length, since it seems both receiver and transmitter need the same length data. I used ' dtostrf(tempc, 6, 2, outstr); //get temperature to character 6 chars long 'in the transmitter,, then ' rslt = radio.write (&outstr, sizeof(outstr));' Could I not use ' rslt = radio.write (&outstr, 6)'; and is 'static char outstr[6];' big enough, or does it need to allow for a null termination, i.e. outstr[7]

I've come across various posts criticizing the use of ' dtostrf', not in this situation (yet) but how would you do it?

I thought it needs to be fixed length, since it seems both receiver and transmitter need the same length data.

The transmit buffer (dataToSend) and the receive buffer (dataReceived) must be the same size. The data sent can be any size as long as the data will fit in the buffers.

I'm going to have to start again from the basic example of Robin2's simpletx/rx. It worked well on the Uno's.

The nano uses the Atmega 328 microprocessor just like an Uno, except that on the nano it is a surface mount version and has a couple of extra analog pins IIRC.


The uno has sockets, for jumper wires, the Nano relies on my soldering, That's why I said going back to the UNo, removes one level of possible problems (but replaced by errors of cheap jumper wires). Thanks for the replies, of course. I need to check that the nRF24's are that I've been using are still OK,

Not sure if this is better off in the wall of Shame, but this is my first arduino project. A controller for LPD6608 Leds

Well, if it works, then it's OK. Back in the day, I did smaller stuff than that, encapsulated it in epoxy - nobody knew, except me, and now you, I guess.

I've found that a couple of the nRF24 are a bit iffy. These are the ones with the antena, Not sure if it's something I've done, or not. But I am testing this with only a few inches of transmission distance, and I' believe they don't work to good if they are too close together. Anyway, other than that, seems OK. Sometimes the temperature is in error (-127.00) I think I'll pull that out before sending it. Now to look at the example for two transmitters, and extend it to more. I'll probably hard wire in the tx address - i.e. read in some links on spare i/o ports. It will be easier than writing specific code for each Tx - hopefully.

I've found that a couple of the nRF24 are a bit iffy.

I bought some Ebay rf24s that I had trouble with. They had a blob instead of the chip on them. So now I make sure that they have the chip and haven't had much trouble since.

These got a tin cover, could be anything under them. The cheap pack of ten seem ok, but these have pcb antena. When I first got them, they had the range I needed, and worked when I tried them for five minutes. After that I started doing pcb, etc, and they don't work now, possibly messed up by me, maybe not.