Guitar PCB take two

Not only are we trying to make a working miniature guitar, we're also trying to find the best (and easiest) way to make a number of them, in a way which is consistent and repeatable. So far we've managed to make a working instrument, but it was really fiddly to make, and for anyone else to try to follow the same steps, it'd be pretty hit-and-miss as to how successful the end result would be (we're not mentioning that our guitar didn't actually fit into it's enclosure!)

So we're trying out an old idea again - making the neck PCB from fine (0.2mm) traces and connecting via a ribbon cable at one end. Here's the design toner transferred onto some copper clad board:



The board was cut to size to make sure it fit along the guitar neck. In fact, we can be quite loose with measurements and tolerances- at this stage, once we have a working PCB design, we can always amend the acrylic shape(s) to fit the PCB if necessary



Then the board was etched and the edges sanded and finished off nicely



Some half-pitch IDE cable was attached (I think this is 0.025" pitch cable. Whatever it is, it's about half the pitch again of "regular" hard drive IDE cable). This was acheived by tin-plating the end of the PCB first (it was too big to fit inside the glass jar we keep our tinning solution in)



Using our new tin-plating and solder-paste technique, the ribbon cable was actually quite easy to attach, even with a big fat pointed tip on the soldering iron. Maybe the fact the IDE cable was made up of single solid cores helped.

The larger "blob" of solder came off the tip of the soldering iron when we started the second piece of multicore cable. It's actually a tiny little bit of solder, but looks enormous on this photo!

The cable ends were checked for continuity and the new-style PCB put into place to make sure the ribbon cable wasn't too wide for the opening



Compare the amount of loose wire with the new board and the old one. With the new board, we could simply loop the multi-core cable under the entire main board PCB and attach to an edge connector.



We considered an alternative - creating a single, solid piece PCB for both the body and neck in one go. This has the advantage of being easier to assemble (although etching requires a shallow bath rather than our preferred "dunk-in-a-coffee-jar-of-gunk" approach) but the downside is that the guitar neck would appear lower than the guitar body (once the acrylic top layer is in place).



We were keen to keep the guitar fingerboard higher than the guitar body so that it accurately reflected the construction of a real guitar



To date the only way we can see this working is to have the guitar neck as a separate board, attached to the main board via a ribbon cable. If we used a single piece PCB, we could make the acrylic of the guitar neck from 3mm instead of 5mm to reduce the height (reducing the position of the body part of the PCB) and introduce a layer on top of the neck PCB, making it appear raised higher than the body?

Almost working guitar

Using our new tin-plating and solder-paste technique, we're finding attaching IDE cables to our new PCBs a doddle! These cables were attached quickly and easily whereas usually we'd be using tons of flux, solder braid and going overtime with the solder sucker!



Here's the main "body PCB" for the guitar, populated with the PIC microcontroller, darlington transistors, crystal and discrete components to make it work



With the body circuit made up, the next job was to make the fingerboard.
The fretboard consists of 12 pairs of contacts on the top face. Each pair of contacts consists of a pin that is connected to an output on the PIC and a pin connected to the gate of a darlington transistor. When a finger is placed over the two pins (separated by about 1mm) and the output pin sent high, the transistor will switch on and pull the sensing input low (if no finger is present, the sensing input remains high because of an internal pull-up resistor).

We're connecting our finger contacts to the main board using IDE cable, so connected one pin from each of the pairs to a common sensor input pin on the PIC.



We're using 1mm PCB via pins for our contacts.
The wires will be connected to the underside (copper side) of the fretboard PCB and run inside the channel of the guitar neck



The strands of the IDE cable were separated and soldered to the pins on the "neck PCB"



With the body and neck PCBs connected together, the last bit of soldering involved putting the body PCB in the correct place on the plastic cover and threading some actual guitar strings through the face and into the rows of holes at each end of the board. With the strings in place, the ends were soldered onto the PCB and trimmed short



The last bit of assembly involved routing all the wires carefully and fixing the body and neck in place onto the actual instrument



In fact, for this prototype we learned a few things that meant we couldn't actually complete it. For example, the cavity in the middle of the guitar body needs to be measured and placed so that the PCB fits inside it - our approach was to just cut a bit out of the middle layers of 5mm black acrylic that makes up the body and just hope it would all fit together at the end!
Likewise, we learned that to get a nice finish on the strings, we need to make and use a jig for bending the wires, so that they're all exactly the same length so they don't buckle and bow as the guitar top is moved into place.
And we learned that fixing broken traces can be a real pig of a job when your PCB is stuck down inside an enclosure and you've loads of wires everywhere.

So there we have it - a sort of working guitar.
In fact, you can plug it into the USB port and use it to play samples by selecting the frets and strumming the exposed strings - the problem is, if you're not very careful, the whole lot springs open and spills its guts all over the desk!

irobot "verro 500 powerscrub"

For all those pool owners out there, listen up. Meet Verro, an automatic, self contained, pool cleaner. Verro is developed by the all leading robotics company, iRobot. This robot cleans inground pools within three hours, and its tangle free (no hose required). It also has no loud pumps that it uses while cleaning, so it wont disturb you or your neighbors. This bot also has high quality brushes that scrub everywhere, and it even scrubs the waterline. So if your the type of pool owner that hates cleaning the pool, then this is the bot for you, but the pricetag is a hefty $1,100. Thanks for reading and you can see a demonstration of Verro in the video above!

Creating the guitar PCBs

Using our newly-discovered technique of tin-plating the entire PCB then soldering the 0.05" pitch IDE cables using solder paste, we managed to create some ready-to-try boards quite quickly this evening.



The cables were soldered quickly and easily and the joints were strong and consistent; no messy de-bridging with solder braid was needed (which often causes the pcb tracks to heat up too much and start to burn or lift)



A final, populated board, with a PIC18F2455 in place and extra components to make a USB HID device. This was an early board and we'd already attached what will become the USB socket using the "old-fashioned method" of drilling holes, splitting the cable ends and pushing them through to solder on the underside of the board.



Maybe with this new technique, we'll be able to either use SMT chips OR trim or bend the pins on the larger chips to enable us to solder them directly onto the (tinned) traces of the board without having to drill the through holes first. If that's possible, it will mean making PCBs would take a fraction of the time it does at present: it's all the drilling that takes forever and the less drilling we can get away with, the better!

Of course, for this board, we still need two rows of 6 holes for the strings to go through, but for future boards, we'd need only a few holes for things like the crystal and 470uF capacitor (although there are SMT versions of these with relatively large pads available too.....)

Miniature guitar PCBs

Our range of miniature playable instruments - guitars in this case - is coming along nicely, thanks to the new laser cutter which makes an amazing job of cutting out the enclosures from a mixture of 3mm and 5mm acrylic.

So now it's time to add some electronics to actually make them work!
We've designed a PCB that can be hand soldered (no tricky SMT type soldering here) although using soldering paste instead of regular solder will help making fixing the connecting ribbon cables in place a little easier.



We deliberately chose a design that uses through hole components, but it's important to keep the profile (height) as low as possible - after all, the cavity inside each instrument is only 10mm high. We used a layout that would allow us to shape the board to fit every shape and size of guitar, so a single board can be used for every design (it would be a nightmare to have to create a different board for, say, the Flying-V and another for the BC Rich Warlock style guitars)



By using the same board for all guitar shapes, we've also got to be sure that the bridge and string holes are in a similar position on every guitar. Note how the row of six holes on the PCB line up with the six holes at the neck and bridge on the guitar body. This will allow us to push the strings through the body and solder them to the underside of the PCB

Soldering made easy

The key to good soldering has always been to tin everything properly before starting. If you're not too confident with your soldering, you can also use flux/rosin to help the solder to flow (we flood everything with flux before soldering!)

The problem with fine-pitch soldering is getting the components tinned, and especially tinning fine multi-core cable. Tinning PCBs can be a problem, but with plenty of flux and a tiny amount of solder on your tip, this should be possible, even with a big tip. We've always found tinning multi-core cable to be really tricky!
The tiny thin strands in each core separate, the solder creates bridges between the different cores, separating the cores makes the wires heat up and melt the plastic coating and generally everything gets a bit messy!

While we're not quite ready for SMT soldering again yet, being able to pre-tin PCBs creates some exciting possibilities (though only really if you're excited by this sort of thing). Here's an example of how tinning can make soldering really easy (which is great if you're still having problems with big chunky through-hole components and massive 2mm pads!)



The job here is to solder the 0.05" pitch IDE cable to the traces on the edge of the PCB. Normally we'd use standard 0.1" pitch holes, split the cable at the ends and thread the individual wires through the holes and solder to the underside of the PCB. By using this new method, we don't have to drill as many holes on the board!

[EDIT: see last photo in this later post for an example of the "old way" of soldering IDE cable to our PCBs]

As always, we begin be covering the PCB traces with plenty of flux. Then just splodge a lump of solder paste across the traces to be soldered.



Hold the wires in place and heat up the solder paste with the tip of the soldering iron. We started by tacking the outer wires by heating the tracks rather than the actual wire. When the cable was held in place, we then put the tip of the iron directly on top of each core of the cable. You can see the solder paste melt and turn shiny. After a second or two, the solder starts to appear on the top of the wires - transferred, perhaps, by some form of capilliary action?



Anyway, whatever the reason, the end result is some beautifully soldered wires, no bridging (any excess solder paste is either burnt off or just runs towards the pre-tinned solder tracks) and a nice, strong, solder joint on each core of the cable.

The last thing to do is to test the continuity between the exposed ends of the cable (the opposite end to the one connected to the PCB) and the traces on the PCB. By testing "outside" each end of the solder joints, we can be confident that if we get continuity, the joint is good.



Any solder paste should do. We used this stuff. It was given to us by Matt from BuildBrighton. He said it was about 12 months out of date but had been hiding away in a cupboard and he had no plans to use it. Although the paste has a slight "crust" when we opened it, the paste underneath was still soft and usable.

Cold solution tinning powder

Here's a spanky new product (ok, maybe not to you, but it's the first time we've ever seen anything like this) from Farnell - http://uk.farnell.com/mega/600-020/tin-plating-powder-90g-for-1-litre/dp/1550282 - it's tinning powder for homebrew PCBs. You might be able to get the same stuff from Rapid, but we're Farnell diehard fans here, though you may find the same stuff cheaper on eBay.

Wherever you get it from, make sure you buy the stuff that works at room temperature. Dissolve the crystals in warm water (the instructions recommend 50degC but I just made a brew, drank it, gave it a few more minutes then used the left over hot water from the kettle.)

Why bother tinning your homebrew PCBs?
Well, firstly, it makes them easier to solder to. We use loads and loads of flux for the same reason, but as you can see from this board (below) the exposed copper tarnishes quite quickly. Tinning helps reduce tarnishing on your finished boards (and gives them a nice shiny coat which looks a lot neater!)



About twenty quid gets you 90g of powder which, according to the instructions, should make a litre of tinning solution. In powder form, the compound has an unlimited shelf-life. Made up, it's expected to last about six months (but they say that about Ferric Chloride and we're still using the same batch nearly eighteen months after making it up!)



Just make up the solution and drop the board in.
We left ours in for less than ten minutes, face side up. Rather than use the whole lot in one go, we made up about 200ml and just added a bit of powder at a time, stirring until it had dissolved. The result was pretty impressive:



Nice shiny silver traces make for easier soldering!