< $50 robot

[totalis]

I started out reading hundreds of tutorials, and following guides and everything else. one of the best was the $50 robot. This, excluding setup (soldering iron etc.) cost about £60 (inc delivery) in the UK. The total was actuall more like £150 including programmer...

this to me seemed expensive, and while i wanted (and still want) an Axon, i had to made do with lesser devices of my own creation.

Nearly 2 years on, and some (lots) of money later, i have made a microcontroller (professional PCB, only 2 SMT devices) at the total cost of ~£6 ea. Also getting servos (cheap) from hobbyking helps bring down the price of the finished robot. The programmer can be a £2 nokia data cable



tl:dr

What do people think of my MCU? is it of interest to anyone? basically i want some feedback (constructive or otherwise) if this would be of any use to anyone.

BTW, more pics to come (i will photograph them in the morning)

[totalis]

sorry...

[WaterPig Master]

Hi there,

That's a nice little board! Are you planning to take them commercial? I wouldn't know if there's a market for really small, low cost support boards for MCUs, but it could be worth looking into.

How about giving it holes for screw-mounting? I'm sure you could jiggle around the layout of the headers and fit a hole in each corner without sacrificing the awesomely small size of the board?

Perhaps add two SMD LEDs: Power indication, and one connected to any spare I/O pin you might have, so the user has a general 'indication' LED — I added one to a spare pin on my MCU board and it was hugely useful during debugging of my robot.

I am planning on making something like this in the future, just for the satisfaction of having made it myself My ideas involve a slightly different approach, where I would provide more abstraction, and have things like motor driver circuitry on the actual board — which will end up much larger than yours, I think!

Can we see an assembled version, and the back?

[billhowl]

Hi,
The 7805 may not support too many servo, so some of the center pins can connect to Vin and not 5V.
The serial out Tx/Rx can include Gnd pin so that one connector can be used.

[totalis]

Thanks for the replies i had considered having a pin devoted to an indication LED but i didnt want to sacrifice any of the IO, so i could just use a solder jumper and let people choose The power indication is a very good idea, that will be in V2, as will the mounting holes. As for the GND next to the RX and Tx, that is a very good idea and i will jiggle the board till it fits

While i understand the current limits of the 7805, i am reluctant to change as i am making a little board designed to fit over 8 pins and give another power input. im not sure how this will look yet but i personally use regulated voltage much more than unregulated, but this may change...

as for more pictures, i will post another few of the front and back of the PCB and an assembled version tomorrow

Thanks again

T

[Soeren]

Hi,

The pins on the lower half of the PCB could be moved "up" for a smaller PCB.
If you mount the X-tal and the voltage regulator "un-slanted" (which doesn't seem to serve any purpose at all?), you can crimp the board even more.
If you wanna sell it, either tin or gold plating is a must.

Edit:
You have an acid trap on the input pin of the voltage regulator and in a few other places.
You need to check the pad-track clearance - like one pad of R3 is suspiciously close to the track leading to R2
If you make mounting holes as suggested, make sure you remove the copper (on both sides) to avoid shunting the regulated voltage when mounting to a metal base.

[totalis]

The pins on the left can be moved up but there isnt much clearance to move the pins on the right, although that could be achieved by un-slanting the VREG.

 As for the slanted VREG and XTAL, on a prototype of this board i had trouble getting everything to fit (PCB costs the same aslong as they are less than 5cm x 5cm ) Also the XTAL will be reoriented, the eagle sketch was not very clear about the traces going to the XTAL.

The only redeeming quality of the slanted VREG is that it allows a mounting hole to be made in the corner

And yes the pad of R3 is very close, will be updated next revision.

Also, one thing i have noted with this version is that the programming header is the wrong way round, meaning the programming wire comes across the board, not off the edge...

Can someone please point out the acid traps, i dont quite understand what they are.

Thanks again guys

T

[Soeren]

Hi,

[...] Also the XTAL will be reoriented, the eagle sketch was not very clear about the traces going to the XTAL.

I thought it was your own design??

Can someone please point out the acid traps, i dont quite understand what they are.

As I mentioned, one is at the input pin of the voltage regulator and there's a few more of them.

An acid traps is where two traces (or a trace/pad or pad/pad) meet in an angle less than 90°. The sharper the (inwards) angle, the worse, as this is where etchant will typically be left and a normal rinse won't clean it properly.
Later on, ambient humidity acting on the residues will have various unwanted effects, like eventually cutting a track, or strange and intermittent capacitance influencing your circuit.

[totalis]

Hi,

[...] Also the XTAL will be reoriented, the eagle sketch was not very clear about the traces going to the XTAL.

I thought it was your own design??

It is my own design, but when i put the XTAL down on  the eagle sketch, it looked alot bigger and so i ended up routing the traces to avoid the silkscreen and not the XTAL traces...

Thanks for the info. acid traps are not something i had considered, i will try to avoid them in the next revision.








T

[WaterPig Master]

Hi there,

Seeing it assembled immediately makes me think you should use SMD caps and voltage regulator. It looks dangerously heavier on one side, and much taller too. It would make routing traces easier, especially if you replaced the resistors with SMD ones too. Is there any particular reason for the TH components?

Thanks for the info on acid traps — not something I've come across before.

Looking good!
Barnaby

[totalis]

Hi there,

Seeing it assembled immediately makes me think you should use SMD caps and voltage regulator. It looks dangerously heavier on one side, and much taller too. It would make routing traces easier, especially if you replaced the resistors with SMD ones too. Is there any particular reason for the TH components?

Thanks for the info on acid traps — not something I've come across before.

Looking good!
Barnaby

I chose not to use SMD caps because the space saving is not very much, and the complexity of the soldering is higher, as for the VREG, the price is the main reason for the choice of package, and it gives a relatively small footprint.  The main reason i chose TH components is that if i were to sell this (or version 2) as a kit (for $50 robot), i would send the board pre-soldered with the SMD things and so it would increase workload, also there is not very much saving in choosing SMD resistors over TH.

the only other reason for the through hole components i can think of is that i personally like to know i could fix something easily if i break it (a relatively common occurence :-))

and yet when i add power/activity LED's they will be SMD... but they get tinkered with less so meh...

thanks again guys


T 

[totalis]

One final question, would anyone buy one of these MCU's for £10 ($16, 11EUR) + postage?

This would make a brilliant introductory MCU (ie $50 etc)

They will be fully tested and i will ship them in a large amount of bubblewrap to protect them during transit.

I am thinking of selling 6 or 7 of these things. The price is low because i am only interested in covering the cost of making them.

I can also provide a brd file for use with webbots fantastic project designer.

PM or reply if you are interested.

[Admin]

Does it work?

I ask because I don't see any capacitors on the crystal . . .

[corrado33]

I was wondering the same thing admin...

You could always use the internal oscillator (I didn't look at the chip, does it have one?)

[totalis]

The XTAL is a resonator and so doesnt need the 2 capacitors, i chose to use the resonators as opposed to a crystal oscillator so that i could save space, but i think i could have used either.

They do work btw, i have made 8 so far and apart from a solder bridge (my fault) all of the ones i have work a treat. i have even used them with webbotlib to control a leg on my hexapod

i will be posting more videos in the coming week, including an overview of my sub-£100 18DOF hexapod

in fact, if you can see this...

<object width="400" height="224" ><param name="allowfullscreen" value="true" /><param name="movie" value="http://www.facebook.com/v/10150243605199678" /><embed src="http://www.facebook.com/v/10150243605199678" type="application/x-shockwave-flash" allowfullscreen="true" width="400" height="224"></embed></object>