A week in the shop

Last week I thought I’d finished the oiling of the prototype offset, but I was wrong. I’d assumed the wenge fretboard on this would be treated just like a rosewood fretboard: I’d use a specific fretboard oil (e.g., lemon oil) to preserve the moisture in the wood, rather than a gun-oil type oil to lacquer it like I would a maple fretboard (as indeed I do for the rest of the guitar). However, I tested the fretboard oil on a scrap bit of wenge before applying it to the neck, and much to my disappointment it it removed all the nice two tone colour that wenge has, all going a uniform dark brown. I had a chat with Matt over at Fidelity Guitars, and he confirmed that he treated wenge more like maple than rosewood, recommending a few coats of a gun-oil to give it a thin lacquer.

Thus the neck was remasked to let me oil the fretboard, which I did over a few days, and then I’ve left it a few more days to cure. The results are worth it: the fretboard now is nice to touch and has preserved the two tone effect of the wenge.


I think the only thing that I'm disappointed about on this neck is the clay inlays, which I did as an experiment to see how they'd work compared to using plastic inlays. Overall they fit, but they have been stained somewhat by the wenge dust as I've sanded the it back, and it's not quite the finished I'd hoped for. So somewhat a success but I think in future I'll try something else.

Whilst the neck was curing I went to mill the pick guard for this guitar on the CNC router. Unfortunately, the run did not go as planned, though for once it was not the machine that was at fault, and rather my inexperience of working with PVC and doing chamfers with the CNC router. 


The above picture shows my attempt to chamfer the holes where the mount screws would go on the pick guard (this was the first stage of cutting it). All the holes should be the same size, bar the few on the bottom left where I was testing cutting the PVC, but as you can see they vary in diameter wildly. This is because the PVC is flexible, and even though I’ve screwed it down to the bed the surface of the PVC isn’t perfectly level, which causes this variance. 

To make a chamfer you use a tapered bit, in this case I had a bit that had a 90˚ angle at the tip, like so:

IMG_0272 2.jpg

The depth at which you insert the bit into the material will dictate how wide the overall hole will be: as you push the bit in more the taper means the bit touching the material is wider. Unfortunately because the PVC here isn’t absolutely flat you end up with contact points at subtly different heights which means different sized holes, which is clearly not what we want. Even if I’d done a better job of pinning down the PVC I’m not sure I could totally eliminate this without ensuring the bed on the router is perfectly level and the PVC uniformly thick, so I think I’d need to chamfer the plates after cutting next time using some other method.

Given I’d ruined this material and I want to get this guitar finished, I decided to try a very different approach: I found some snazzy laser cutter safe material and made a scratch plate out of that:


This has a totally different vibe from what I was planning, but I’m quite pleased with the result for now. We’ll see how it looks when assembled: I can always change this easily later. It was quite nice to do prototyping timeline on this:


Left to right, that’s the first paper version cut on the laser cutter to check general sizes, the cardboard version I cut to mount the pickups and selector on to make sure they all fitted, and then final acrylic version also cut on the laser cutter. I often tell people to measure, measure, cut, and this is a nice example of that principle in practice :)

A little while back I did a test print of my own control knobs, and I realised this week I’d not bought any for this prototype offset, so I’d just make my own. It doesn’t offer a particular cost saving, particularly when you factor in time, but it’s nice to have yet another personal touch to a guitar. 

I started by making some mockups in Fusion 360 and printing them using the low resolution Ultimaker just to ensure I was happy with how they feel in hand and in context on the guitar. There’s two designs: one for the volume and tone controls, and another for the pickup selector dial, which I’ve modelled slightly on the one used on the Fender XII selector.


Having got the low resolution printouts made and I'm happy with how they feel, I can now do a high resolution printout with the FormOne printer, but first I need to know how bit to make the toothed hole inside to get a good fit. Although notionally pots have a 6mm toothed shaft, I’ve found in practice they vary a little, and this is important if you’re just going for a friction fit. So I made a little gauge and printed that out.

Screen Shot 2018-06-18 at 11.21.09.png

This gauge has 0.2mm increments from 6mm to 7mm diameter, and I’ll use that on each pot to work out the best size and then print the control to match that measurement. The taper on one end is there to remind me which end has the smaller hole on it.


This is the print just out of the FormOne, before I’ve cleaned it up using a IPA bath and then done the UV curing to harden the outside. 

Following on from my plans last week to try and make good on some of the failed parts, I took the body that the CNC router had cut into and did a rough pass to get rid of most the unneeded wood, ready for trimming it using the templates:


I also cut a third offset template, this one with the holes for a dynamic vibrator, so I have the option of making it a fixed bride or a vibrato unit. I also made sure to put the control cavity in this template, as I accidentally forgot to add it to the last set I made!

Lastly, my OMEC Teleport arrived, and has been installed on my pedalboard.


I’ve not played much with it, but I can conform it plays nicely with the iPad, but you do need to get the iPad Camera kit adapter (which is almost half the price of the Teleport itself!) or some other USB-B to lightning cable (it ships with a USB cable only). I tried it both with my own ongoing audio filter project and with Bias FX.

It amuses me somewhat that of my 5 pedals 3 of them have a USB cable sticking out the back to let me play with them in one way or another.

A week in and out the shop

This week has been a quiet week in the workshop. At the start of the week the prototype offset was at the stage whereby it just had a coat of oil on the body and neck once a day, and then since I’ve left it a full week to cure properly. In the coming days I can get back into action doing the assembly and cutting the pick guard etc.

Screen Shot 2018-06-11 at 13.19.04.png

As mentioned last week, I realised I can't laser cut the pick guard (as most pick guard material isn't laser cutter safe), so I'll have to mill it instead. Given I was milling it I've ordered a 60˚ angled router bit so I can chamfer the edges properly. Last time I tried to use the CNC router for this kind of thing the material cracked around the spaces near the screw holes, but I'm confident this time I'll have better success as Fusion 360 lets me control things better than the tools I used before.

In prep for heading up to DoES Liverpool next month to use their CNC router I started the process of getting all the wood ready for that trip. I need to make a new body blank to replace the one that was destroyed by the errant router at Makespace last time, and I have had for ages a bunch of nice wood sitting around ready to make the necks for the two commissioned offsets. The first stage was to thickness all that wood, and I once again took the kind offer of a friend to let me use his thicknesser.


At the moment I’m feeling that the number of guitars I’ve generated is very poor for the amount of times I’ve used this favour. Hopefully we can start to put a dent in that. 

I really feel that as we reach the mid point of 2018 I’ve not had a successful first half of the year: I’d hoped to have made three or more guitars by now, but it looks like I’ll only have managed one guitar, the amp, and a bunch of harsh lessons about things going wrong. It’s very easy to let setbacks cause you to lose momentum, particularly when you have other things you need to do to pay the bills, and I look back on the first half of the year with a sense of having got distracted and not kept up pace.

To try and offset some of this (pun not intended), I’ve been looking at what I can do to try and turn some of the bits I have left over from failures into an actual guitar that I could ship as a sort of factory second. I have two bits that immediately come to mind as things I can try spruce up. Firstly, I have the neck where the fretboard didn’t glue on perfectly, leaving it with an obvious seam that meant I couldn’t use it. I think with this I could either go for an edge binding to cover the seam, or I could try perhaps heat it up to melt the glue and remove the fretboard entirely and try re-glue it. I bought some binding material, but the standard sizes it comes in of 1/4 inch is not quite deep enough to reach the seem without having to have a couple of mm off the fretboard top itself. Anyway, as other things move forward I shall see if I can find a way forward here - suggestions welcome!

The second bit is the body that the router ruined that was meant for one of the commissions. It now has a large gouge near the upper horn from there the CNC router decided my design wasn’t interesting enough, and it had a better idea. What I’d like to do with this is fill in the gouge and finish cutting the body out with templates. It’s not going to be useable on the commission, as it’ll be an obvious flaw - the guitar will be finished in stain/oil and it’d show through. But after filling I could paint this one to hide it, or I could follow the Kintsugi philosophy (on both this and the neck actually) and fill the area with a gold resin, and make it a feature that tells the tale of the troubled build process (a hat tip to my friends Marine and Dom for that suggestion).

My problem is that I’m a perfectionist about the guitars, and I want them to be flawless, and when I fail to hit my mark I sort of shy away from things for a bit as I lose my confidence that I can achieve what I want is dented. This is something I need to try come over, as its an inherent part of learning and getting to grips with hand crafting physical objects.

As part of trying to continue to develop my manufacturing skills by tackling some smaller projects I bought a cheap effects pedal kit and set about designing a case for it. Currently I’ve just started with a traditional shape, but eventually I’d like to do something less uniform in its shape and 3D print that.

Screen Shot 2018-06-10 at 12.41.35.png

Strength will be the main issue with anything 3D printed, as in the heat of the moment effect stomp boxes probably take quite a hammering. But this is mostly as an experiment, and if it fails I’ll have learned something and I can just use the metal enclosure that shipped with the kit.

I also spent a couple of days this week refining the audio effects software I was playing with last week, trying to shape it into something that is more practical to let me (and potentially others) experiment with. There’s a saying in software that once you’ve finished 90% of the product you get to do the other 90% - a reference to the fact that the interesting bit you care about is what  you think will take most of the time, but is in fact only half the problem, and you still have to do all the dull and tedious bits to make it usable by people. I spent some time this week on tweaking my software so I could save my effect filter experiments and share them between all my iOS devices. allowing me to work on them both on my phone and tablet seamlessly. It’s been a long time since I wrote iOS software in anger, and thus a lot of time was just trying to get back up to speed on this. 

I’ve been testing using an old iRig2, which is great for travel, but is very frustrating to use most of the time with my pedal board, as it requires long heavy audio cables running up to my iPad and back to the board. But this week the Orange OMEC Teleport finally hit the shops, which is effectively an audio card in a stomp box, which then has USB out the stomp box to your phone/tablet/laptop. So to try and stop me going crazy due to tripping over cables I ordered one of these to make it easier to test.


A week in the shop

This week I made progress on the prototype offset some more, found a potential alternative CNC router option to help me move forward on my backlog, and played around a bit in both Fusion 360 and with some DSP code.

One of the things that I’m embarrassed to say took me a while to get used to was the idea that wood isn’t perfect. I come from the world of electronics and software, and whilst you can argue that software is often far from perfect, from an engineering perspective it does at least let you generate pristine things. So it took me a little while to get used to the idea that I couldn’t just reject a bit of wood because of some tiny flaw; if I did that I’d have no wood to work with. You have to work with these flaws and find a way to hide or embrace them.

The first thing to do is when you get your wood is be selective about which bits you keep and which bits you cut out: can you, for instance, position a knot in the wood such that it’ll end up where the pickups or neck pocket will go and you’ll route it away? But even after that, you can still be taken by surprise as you peel away the layers of the wood as you guitar progresses, you’ll often find hidden features that you didn’t expect. On the fretboard for the prototype offset I'm finishing up at the moment I found this hole as I trimmed the board flush with the neck profile, where some insect has buried its way into the tree.


Thankfully, it turns out that fixing small things like this is relatively easy. All you need is some wood glue, and some sawdust from the wood in question: either from as you process the piece hold on to some, or from an off cut that you held onto for this circumstance (which is what I did here). You can just mix the sawdust until you get a fairly viscous paste (like a cement) that you then pack into and over the hole.


Leave that to dry overnight, and then you can sand it back flush with the rest of the wood.


You’ll still notice a mark there if it you go looking for it, but no more so than any other bit of figuring in the wood, and once oiled you’ll hardly see it. You can see more details on how to do this in this video from Crimson Guitars. This was my first go at this sort of repair, and I’m quite pleased with the results.

In general this week has been slowly applying finishing to the neck and body of the prototype. This is a mildly frustratingly slow stage where you apply a coat of stain or oil once or twice a day and there’s not much else you can do. I started the staining at the end of last week, and I was using a yellow stain, which after half a dozen coats has gone a wonderfully deep orange peel/sunset orange. It’s not what I expected it to do, but I’m still rather pleased with the results so far. This here is a stop motion of me applying the first coat:

I’m now four coats of oil in, with two more to go.


Once the final two coats are done, I'll leave it for a week to properly dry, and then I can cut the pick guard and wire it up.

Speaking of pick guards, I realised that my plan to laser cut the guard wasn’t going to work. Although it’s been great for prototyping, all the plastic suppliers I’ve found for traditional scratch pick guard material supply them in PVC or other such plastics that I can’t laser cut in the workshop I use due to the toxins they would emit. So instead I'll brave using the CNC router for this instead.

Speaking of CNC routers, in a bit of a breakthrough, I’ve actually found another CNC router I can try using. Regular readers are probably sick of this by now, but to recap: I’ve always made guitars using the CNC router at Makespace to rough out the bodies and ensure things like string holes are perfectly positioned. Unfortunately this year the router has started to go senile, and has twice destroyed bits I was working on, costing me both time and money. More distressingly it’s put me quite a bit behind schedule on a pair of commissions I’ve had. Whilst the CNC router at Makespace will get fixed eventually, that doesn’t help me in the near term: it being a community workshop means this process will take a little bit of time.

I was just about to start making the new guitars with a template technique rather than CNCing, when I was connected with the people up at DoES Liverpool, another community workshop that serves the Merseyside area. They’ve just got a new CNC router installed, and given I’ll be up in Liverpool for MakeFest at the end of June I plan to stay on for a few days to visit DoES and see if their machine will help me get these builds back on track. For me it's a way to get back in motion, and for DoES it'll be useful for them to have someone experienced with CNC routing show them what is possible with their new bit of kit. Very excited about this!

I revisited the guitar design I was working on a couple of weeks back a bit this week, trying out some new techniques to design the arm carve on the symmetrical guitar I was designing a few weeks ago.

Screen Shot 2018-05-31 at 13.42.50.png

Last time I played with this I tried to model the arm carve using a mesh in Fusion, which I found very hard to work with for this kind of application. A friend of mine, the very talented Tom Armitage, who has also dabbled in Fusion 360 quite a bit, suggested I tried using a technique called lofting, where you can get Fusion to generate shapes based from a set of intermediary sketches, and it can do those shapes following a defined curve. You can see a prototype version I did below: here I sweet between two small arced corners and a larger one in the middle, following the edge of the disk.

Screen Shot 2018-05-31 at 10.54.17.png

Now if you look back at the picture of the guitar body in Fusion above, you can see how I did the same thing on the guitar, just using points for the ends rather than corners. Really pleased with this technique. I’m still not sure about this guitar design overall, but it’d be nice to get it completed so I can then move on to something else, and the experience with this one will help with future designs.

One thing I’m fascinated by is trying to understand where the sounds I like come from and how I can, through a combination of guitar design choices, amp design choices and effects pedal choices get the sound I have in my head to come out of the speaker. To help with this I’ve been leveraging some of my other skills, and I coded up some simple audio filtering software for my iPad to let me tweak my guitar signal.


Whilst I watch things like That Pedal Show where they try to explain how things impact your sound, there’s nothing like experiencing yourself. Whilst I could make some silly and whacky sounds with my little bit of software, by far the most interesting was just playing with low and high pass filters and EQ effects. If you’ve never ran your guitar through an EQ unit, I highly recommend you do so, as it’s amazing what impact on the character of the sound from your existing setup that can have. If you want a guide to help you get started then I can recommend reading these notes by a sound engineer on how different EQ ranges impact different instruments, but again I'd say just go play with it, as I’m fairly sure you’ll be amazed what you can do with it.



A week in the shop

This last week has been a productive week, with lots to talk about, so let's get to it.

As mentioned last week, given the issues I’ve had with the CNC router I have access to, to be able to make progress again on the commissions I have I’ve decided to switch to using templates to get the rough shaping done. Last week I designed the templates in Fusion, so I started this week by cutting them out:

IMG_0048 2.jpg

Obviously this time the CNC router played ball, but I still don't trust it with anything critical (or at least cheap to replace like this MDF). Now that I have the templates the next stages are to make a new body blank for guitar #5, and to practice making the string holes and bridge mounts via a template. Clearly it’s very feasible to do this, as this is how lots of luthiers without access to CNC make guitars, but I need to just practice using the pillar drill with a template to line things up precisely.

I made progress on the demo offset I’ve been making. I’ve set myself the goal of getting this done before Liverpool Makefest (where I'll be explaining to people how guitars are made), so I can have something non-t style to show. I spent a lot of time with sand paper getting the body just right before starting to apply the stain. You can’t rush this stage if you’re going to stain/oil finish on the wood, as any scratches you miss will show up clearly in the stain, even if you couldn’t spot them before. So it’s a lot of starting low, and working up through the grits, from 120ish up to 2000. It’s really amazing how the texture of the wood totally transforms as you do this.


All sanded down I went with a yellow stain for the demo offset. This guitar will have a traditional mustang style vibrato bridge and single coil pickups, and I think a nice bright and fun colour will go with what should be a bright and fun guitar. Currently it’s four coats in, of what is normally six coats of stain, which will be followed by another six coats of oil. This is the slow stage of a guitar build, as you can only do at best two coats a day really, and partly why I like to have multiple builds on the go.


I also mostly finished the neck, which was roughly done, but just needed the polish to get it ready to oil when I start oiling the body. There was a lot of sanding here too, and I also needed to put the logo onto the headstock. In the past I made the laser etched logo in a slightly adhoc fashion: I exported the Electric Flapjack logo in DXF from Adobe Illustrator, and after many goes at getting the right export settings would be able to import it into the laser cutter software we have at Makespace. Then I’d guess a size based on measuring the space on the headstock, and after doing some trial runs on off cuts from the neck, line up things by hand on the laser cutter bed, and then hope I got it right :)

This time however I was able to improve on things, thanks to moving the whole task into Fusion 360 and having the logo as part of my CAD design. First I exported the logo from Illustrator as SVG and importing that into a sketch in Fusion 360. There I could size and position it in Fusion before making it part of my neck design:

Screen Shot 2018-05-27 at 11.21.23.png

Now there’s no guessing involved: I can see exactly how the the logo is going to look, and adjust the position, angle, and size all I want until I’m happy in Fusion. That done, I then generated laser cutter toolpaths (as explained in this post) for both the logo and the headstock itself:

Screen Shot 2018-05-27 at 11.32.58.png

Once I imported these into the laser cutter software, I taped some paper to the bed of the laser cutter and scored the outline of the headstock into the paper and used that to position the neck on the cutter bed before doing the etching itself:


As part of the etching you end up with a sticky residue on the surface of the headstock where the maple has burnt, and but with a little bit of sanding that all clears up nicely, and you have the finished product:


This done I’m almost ready to oil the neck, I just need to tweak the frets a little again as I sanded down the sides of the neck a little and they need refinishing.

Finally, I also spent some time over the weekend doing setups on the guitars #1 and #4, both of which I still play regularly (indeed, I don’t play anything else). Both had seen quite a bit of playing since I last gave them a service, and they were in need of a little TLC. The rosewood fretboard on #1 was starting to dry out a little, so I removed the strings, gave it a clean with some fretboard cleaning fluid I had to remove the gunk that builds up from lots of sweaty rehearsal room sessions, and applied some fretboard oil. Between that and a fresh set of Elixir Polyweb stings it plays like new:


The neck of #4 being a sealed neck hadn’t dried out but needed a clean and a bit of a general check over as this was its first service. I tightened up all the pots again, adjusted the pickup heights a little, and also gave it a fresh set of the same strings.


Whilst adjusting the the pots I accidentally rotated one so that it shorted against the copper shielding in the body, causing the bridge pickup to drop out. It took me an hour to figure out the cause, but was easy to fix (and apply some masking tape over the copper shielding to prevent it happening again in future). Tele style guitars have really compact control compartments, something I’ll avoid on my own designs!



A week in Fusion

Those following these posts will know about the ongoing saga of the CNC router issues I’ve been having. Whilst I’m hoping that the fine people at Makespace will get things up and running again in the near future, and having ordered in more wood to replace the bits I lost the other week trying to build guitar #5, I’m now contemplating going slightly old school on this one, albeit reluctantly.

Not all luthiers are lucky enough to have CNC router access like I do, and so how do they make guitars? The typical way is using a set of templates and using an hand router to run around the edges of those. For a body you typically need at least two templates: one without and one with the neck pocket on, and then potentially more depending on what your internal cavities are for wiring. So I sat down this week and designed the templates in Fusion 360 based off Guitar #5, the first of the commissioned offsets I'm building.

I initially though perhaps I could just make the templates using CAM tool paths from the existing body shape, like so:

Screen Shot 2018-05-21 at 09.33.27.png

Just running those tool paths at the thickness of my template material (either 12mm or 18mm MDF) then I'd get a template without having to create yet another thing in my design I need to make sure keeps up to date as the design evolves.

Unfortunately Fusion really doesn’t understand what I'm trying to do here, and so will get upset if you specify a template material thicker than some of the cavities (e.g., 18mm template but a 16mm neck pocket), and you can’t make a single template this way that has features from the back and front of the body readily as Fusion thinks they're too far apart. So I quickly abandoned this approach on the grounds it's never good when you're fighting Fusion in the long run, and instead I went back to my design, and created a new top level component called templates, and then for each template I just used the sketches from all my existing guitar components to extrude the template outline and then cut out the bits I need. In the end I have three templates:

Screen Shot 2018-05-21 at 10.12.30.png

Whilst I don’t trust the CNC router at Makespace for expensive wood, I am happy to use cheaper MDF to make templates like this on the router and accept the occasional lost bit of MDF. That’s much more acceptable than say an expensive bit of birds eye maple that will be used on the neck for this guitar.

So, what will be the hard bits doing it this way? Firstly, getting the string ferrules all aligned properly will be interesting. On guitar #1 I did them freehand on the pillar drill, and whilst they look fine at a glance, on close inspection you can see they’re not perfect: a CNC machine gets you that precision where it matters, even if you are only using it to rough out. A template should be good enough as a guide, but I'll do some experiments first on scarp to be sure. Also, I suspect for the neck truss rod I’m going to have to build a jig rather than use the template, which is fine if you're making lots of, but a jig per guitar will start to get tedious very quickly.

Anyway, I’ll keep an eye on the CNC router repair process at Makespace, but for now I’ll start by making the body with templates. In addition to the wood to replace that list I have bought some extra wood to let me test this method.

In addition to making the templates for guitar #5, I also spent a bit of time experimenting with a custom guitar design of my own. Whilst I have no problem building guitars that are my own take on existing designs, it’s also interesting to see how I could make my own designs that satisfy some of my own desires for what should be in a great guitar. I started out just by sitting down over coffee one morning and sketching out a series of possible guitar shapes.


In these designs and the iterations in Fusion 360 that followed, there are multiple things going through my head. I had some specific design goals in terms of physical characteristics of the guitar I wanted. An example of that is how symmetric can I make a guitar and still have it work as a gigging instrument? I know quite a few lefties who either play or would like to play, and whilst I could readily make a left handed model if someone asked, I don’t really have the funds to just have a left handed demo guitar sitting around given how infrequently it would get played, which is sad. Another example is I’d want it to have ergonomic cutaways, but then now do you do an arm carve on a symmetric guitar? Is that going to work?

Alongside these constraints there’s obviously a lot of guitars out there that I like the designs of and I deliberately let some of those design cues from what I consider great guitars guide me here. I’m not experienced enough in guitar to yet pull off something radical like the lovely Music Man St Vincent and pull it off, so I’m deliberately taking a more traditional set of inspirations and then mixing it with my own constraints.

Screen Shot 2018-05-19 at 17.13.44.png

The above is my first attempt at following through on some of this, it’s not perfect, but I’m reasonably happy with it for a first pass at a guitar unique to Electric Flapjack Guitars. This is not yet complete, but it's had several iterations with feedback from friends.

You can see the curved lower section is a radial arm curve, which is somewhere between the traditional one sided arm carve you’d find on a Fender and a arch top style guitar. It’s not perfect, but designing this has forced me into a whole new area of 3D design using free-form meshes that I’m still trying to get my head around, so it’s a bit rough, but you get the idea. 

The controls ended up being asymmetric to follow the ergonomics of the reach of a right handed player, and would need to be swapped for a left handed build. But on a demo guitar at least it’d be playable with a quick restring for a leftie, and that’s at least a good improvement on the models I’ve made to date.

At the suggestion of a friends that the single sided headstock design I’ve been using on new builds didn’t work on such a symmetric body, I’ve made a quick mock up of a symmetric angled headstock. This also took me away from techniques I was familiar with (which is good), and I did some research on how to make those. There’s several techniques to making an angled headstock, but here I’ve built it as a scarf joint, which makes it somewhat easier to machine and wastes less material.

Screen Shot 2018-05-18 at 14.33.51.png

Let me know what you think of the design. I still need to finish off the internals for this, but still it was a good learning experience. I really did try to make it symmetrical entirely to start with, but at that point you realise that between the control locations and things like the output jack location, you end up making a guitar that is going to be compromised for everyone, and given a guitar typically has one owner, that doesn’t make sense; but it was a fun challenge pushing it as far as I could. Some people have managed this, for example the wonderful guitars made by Millimetric Guitars, but for me I still think the controls aren’t ideal for gigging when you’re using volume to push an amp in and out of overdrive (but that also reflects the kind of music I like to play, and probably isn’t an issue in other genres). 

To complete the week with my head stuck in Fusion 360, I gave another run of my introduction to making things with Fusion 360 at Makespace, explaining how to get started with Fusion and how to then target the kit at Makespace (the laser cutters, the CNC router, and the 3D printers). The best thing was that at the end of the lesson of of the attendees immediately went into the workshop and fabricated the thing we’d designed in the class, which is possibly the best feedback as to whether the lesson is useful I could have :)

A week in the shop

With things still influx around the CNC router at Makespace and contracting work outside of guitar building keeping me busy again, I didn’t get much time in the workshop, except to follow up on what I wrote last week about exporting things from Fusion 360 for laser cutting. 

If you read last week’s post, I’d managed to find a way to export laser cutter tool paths as DXF from Fusion into the Laser Cut software that drives the Makespace laser cutters. Whilst it seemed to work in theory, at that point I’d not put it into practice. So the main progress I made this week was putting this into practice, cutting the pick guard for the prototype offset

As I did last week, I exported the toolpaths for my pickguard design from Fusion 360 using the Universal Laser DXF process I'd found and copied them onto the workshop PCs that control the laser cutter, and did a trial run just cutting on A4 paper. Here’s a nice video of that in action:

The end result worked quite well, fitting the guitar body nicely:


This was a bit of a relief. I was concerned about how well the bridge and the control plate would mesh up in situ; given both are heavily curved I'd only been able to model them approximately in Fusion, and I was concerned they might not fit well. But that was undue concern, as everything sits together nicely.

However, it wasn’t quite perfect, as the pickups I’d modelled in Fusion were based on random strat pickup measurements I’d found on the Internet, rather than the actual pickups I got from House of Tone Pickups . So I whipped out the callipers, updated my design in Fusion, and then I did a second test cut, this time in cardboard, so I could mount the pickups if they fitted. And indeed, my second pass was spot on:


I’ve now ordered some actual pick guard material for this guitar, so I can do a proper one to actually use on the guitar. 

If I was trying to be spot on like other major international guitar vendors, I’d actually mill out my pick guards, so they have a nice tapered edge. My plan is just to round them over slightly with a light sanding; we’ll see how well that goes. Whilst I do have access to a mill, it’s a lot slower to use than the laser cutter, and I’d like to try do some laser etching on future pick guards potentially, so I’m keen to keep my workflow like this for now.

Whilst all this doesn't seem that exciting, given that pick guards are readily available for most existing guitar designs, this for me is quite liberating. Until now I've been stuck with pick guard designs I can buy, but for the commissions I have at the moment I and to do variations on this offset design that I can't readily buy, and in the future I'd like to be doing my own designs from scratch. So for me this is quite a step forward towards that goal.

Beyond the pick guard, I’ve just been doing yet more sanding on the body for this guitar, making sure I’ve got out an many fine scratches as I can. It’s a slow and tedious process, but I’ve learned the hard way that anything you miss here will show up very quickly the moment you start to stain the body. Hopefully it’s now ready for that, but the neck needs a good going over again with all the grits before it’s ready for oiling.

The only other thing is to remind people that I’ll have a stand at Liverpool Makefest 2018 at the end of June, talking about how to build guitars. Tickets are now on sale now for the event: they’re free, and whilst you don’t need to get them in advance, it does help the organisers predict how many people will be coming and organise accordingly. I hope to see some of you there!



A week in the shop

If I've been quiet of late on social media, it's partly because I've been busy doing non-guitar work this last week to help pay those bills, but it's also because it was not the best of weeks on the guitar front either.

As promised last week, now that know how my offset guitar hangs together with the prototype bits I built, it was time to crack on with making the first of the commissioned bodies. This has been well overdue, but the issues I hit a month or so ago with the CNC router at Makespace going wrong and destroying things had set me back. Since then a lot of nice people at Makespace had put in effort to try and diagnose the issues with the CNC router, and indeed had given it a bit of an overhaul. In the end, whilst there is a general opinion that the CNC router's controller unit should be replaced at some point, no one was ever able to reproduce the issue I had, and it was felt good to use again, and indeed had been in regular use by others all this time without incident.

So, I went back to make the body for guitar #5, which will be a solid body offset like the prototype I made, but with a hardtail string through bridge, and made from american ash. I made the body blank for this back in February, and now it was time to cut it. The first stage was to spend a couple of hours in Fusion tweaking the CAM model. Because each guitar is slightly different and each body blank is slightly different, it’s never just a case of hitting print from the 3D model: I have to review all the tool paths for each stage, plan out the order in which the cuts will be made, ensure that I’ve got the right bit setup, and that it all first in with the material I’m using.

Screen Shot 2018-05-08 at 07.57.50.png

Interestingly the body blank I'd made, which I levelled back in February, had warped ever so slightly in all this time sitting around, which I’d not expected, so I also had to face it off again. This is also done using the CNC router using a very wide bit just to make sure the surface is flat. The variation in height was only about 0.5mm between the edges and the center, but that’s enough to cause a slight wobble on the bed when trying to cut the blank, and if not addressed that'd be enough to make things slightly off. That fixed I then set about cutting the body design out.


The router was just about finished the top body contour pass, when disaster struck once more. There was a horrible screeching noise, which I knew meant things had gone wrong again and the machine was trying to cut bits it shouldn't be, and I was quick enough to stop the machine before it had a catastrophic failure like last time, but not quick enough to stop it ruining the piece I was working on.


It doesn't look like much in the picture, but it's enough to mean that this body isn't fit for purpose. That gouge it made there in the upper horn goes most the way down through the body, so it's not going to be recoverable: this body is now scrap.

I was somewhat heartbroken when this happened. Firstly, although a body blank doesn’t look much, it’s already a reasonable investment to get a guitar to the point you’re cutting the body like this: you have to buy the wood (and good wood isn’t cheap); you have to joint the parts to make a body (I even hand planed this one); you have to thickness it, etc. And all that investment is gone in an instant and you're back to square one. But secondly, and more distressingly, this means I’m never going to trust one of my guitars to this CNC router ever again: fool me once and all that jazz. This means the method I’ve used to make guitars until now has just been taken away from me.

Initially I was briefly angry at Makespace for not taking more action after the machine failed the first time, but in reality they’re in an awkward spot too. A community workshop is just that, a workshop owned by the community. The directors of Makespace set the tone, and they want to encourage fledgling enterprises like mine to scale up from hobby to something more sustaining, but at the same time most members are just amateurs who couldn’t justify commercial rates required to run machines like with very high uptime. It’s an interesting problem for maker spaces in general, and nothing unique to the one I use; and whilst that debate is interesting perhaps, it doesn’t help me get guitars made, so there’s no point dwelling on it here.

There’s a few options forward at this point. Firstly, I know that the community owners of the CNC router at Makespace were looking into alternatives after the first incident, but given it’d happened once and never again, there wasn’t a particular sense of urgency. Perhaps now things will move forward there more quickly, but in reality I suspect not quickly enough for the two commissions I want to get done right now. Not every luthier has a CNC router though, so I could switch to more traditional methods of making templates for my guitars and hand routing the bits I use the CNC router for. The templates themselves I can make on the CNC router using MDF: although I don’t trust that machine, if it just messes up occasionally I’m happy to burn through the much cheaper MDF than I am expensive guitar wood. But I’ll need to find my own plunge router as the one in Makespace is missing its depth gauge, and I’ll need a new set of bits for this style of building. Or I try find another CNC router in Cambridge I can use, but given the type of people I see using the one in Makespace, I assume currently that sort of thing doesn’t exist.

Anyway, if I’ve appeared quiet on things like Instagram and Facebook of late, this is why: it’s been a dark week for Electric Flapjack guitars having to realise the way you've done things until now has to change radically. Ultimately I’ll find a way around it: it’s temporarily knocked the wind out of my sails, but I really enjoy building guitars so I won’t let this stop me. The reward you get when you see someone play a musical instrument you made far outweighs the frustrations of having to change how you do things. 

I’m also fortunate that my commissions have been very understanding about the issues that have presented themselves - my first concern when this happened was about the deadlines I’d set for the these builds and how I was going to let people down, so I made sure to communicate the state of things with them, and they’ve been supportive, for which I’m very grateful.

On a more positive note, I think I’ve finally found a path forward for getting my designs in Fusion 360 onto the laser cutter at Makespace, which thus far had eluded me. I could (haha) target the CNC router and the 3D printers in Makespace from Fusion, and the laser cutter was the last thing I wasn’t able to join the dots on until now.

The main issue with the laser cutter we have at Makespace (and at many other similar spaces) is that the software used to drive it, Laser Cut 5.3, is terrible. It has only the most basic of drawing tools in the software itself, so you can’t use it as a design creation tool, and its DXF import (DXF being the standard 2D vector format for exchanging CAD files) is utterly rubbish. For example, in general it’ll accept DXF generated by Inkscape (another tool I’m not a fan of), but getting DXF generated by Adobe Illustrator is something of a gamble each time.

Fusion also complicates matters: you can export 2D sketches as DXF format, or you can export 3D bodies as STL (a 3D vector format), but you can’t export the face of a 3D body as DXF, so if your design is not directly from a sketch, then you’re stuck. As an example, look at this pick guard design I have for the prototype offset I’ve been making:

Screen Shot 2018-05-08 at 08.10.33.png

The outline of the pick guard and the screw holes come from a 2D sketch that I then extruded, but the neck pocket and the pickup holes are created from subsequent operations that remove material from that extrusion, which means I have no sketch in Fusion to export that looks like the final design, which means I can't generate a DXF file to feed to Laser Cut 5.3. In actual fact, it wouldn't matter anyway, as the DXF generated from Fusion based on sketch doesn't import (or at least never has for me, and the Fusion 360 forums suggest I'm not alone).

After much playing around and a few dead ends, I finally found a nice simple route for this. You can use the CAM facilities in Fusion to generate tool paths as if you were going to cut it using Fusion rather than Later Cut 5.3, and then rather than generate the machine driving g-code that you’d normally do as a last stage, you use this DXF exporter plugin, which just generates a DXF that follows the toolpath for the thing you wanted to cut, and that DXF files generated by this do seem to import nicely to Laser Cut 5.3.

Screen Shot 2018-05-07 at 10.46.29.png

Clear as mud? So, in Fusion, you can go to the CAM section and create tool paths for laser cutting just as you would for CNC routing or milling. If you want a quick overview, then you can watch the latter half of this video here. But at the final stage when you’d go to what Fusion calls “post processing” rather than selecting the type of machine you have, you select the “Universal Exporter DXF”  as your machine, and you’ll end up with a DXF file that’s just got the lines for the paths in that CAM section (this means you can also generate a DXF per tool path if you want). This DXF then (at least in my limited test cases) imports cleanly into Laser Cut 5.3.


This means I can now cut some test pick guards to check pickup mountings and general positions before cutting the actual things.

As an aside, one thing I learned during all this is that Laser Cut 5.3 has a demo version that you can use at home for testing things like this. Whilst the binary for the full program requires a hardware dongle to run, in the same folder there’s a binary with “_demo” in it’s name that’ll run enough features to let you do some playing around. Given how hostile Laser Cut 5.3 is to use, having a copy that you can practice on away from the laser cutter is great.

Two weeks in the shop

I missed last week’s update due to a quick visit to the Isle of Skye for my sister’s wedding, which meant a slightly shorter working week last week and this. I did use it as an opportunity of going past Glasgow to deliver the Red Rocker guitar to its owner at last, which was nice. Still, I did manage some shop time over the last couple of weeks, so here's what I did.

The prototype offset guitar is coming along. I did the fret work on the neck, getting the frets levelled and polished, which took an entire day. With this build I’m trying to keep track of where time is spent on a build, allowing for the fact that I’m still learning as I go. But frets I’m kinda used to now, and I suspect this is par for the course. 


I also spent a good chunk of time sanding the body down. My theory is that every finer grit level of sand paper only exists to show you imperfections in the sanding you did two levels back. It’s amazing how you’ll think you did a great job at 120 grit, getting the tool marks out, and then by the time you get up through 240 grit, it still looks good, but then at 400 grit there’s suddenly a bunch of tool marks showing again! So then it’s back down and up again through the grits. It’s tiring work, but time invested here will pay dividends in the final finish.

As I mentioned last time, my aim was to get this prototype strung up (at least partially) so that I can have confidence that my offset design (which although based on an existing popular guitar, is my own CAD design) hangs together right. I finally got the last of the bits for this guitar, a Mastery bridge unit, so I was able to sit down and put some strings on it. 


It all seems to sit together really well. Now I’m happy with this, I can next week start cutting the wood on the two commissioned offsets I have, which is a great relief. This one I’ll continue to polish up and finish as the others go along, but right now my priority will switch to getting the other guitars moving.

Whilst waiting for parts to be delivered at the end of the first week, I did finally sit down and write some code for the OWL effects pedal I got recently. The OWL pedal is an open source hardware project that lets you write your own effects code and load it onto the pedal. Lack of time had meant that despite having it a couple of months I’d not tried anything yet, so I used a spare couple of hours to correct that.


I wrote a simple, but complicated, echo effect as my hello world program. Echo effects are fairly simple to write, you just need a memory buffer into which you write the sound as it comes in and then read it out later based on some delay. You can of course play with this, and vary things like how much signal you blend back in (how loud is the echo) how long the delay is (who quickly you hear an echo) and then how quickly the echo decays. I mapped a bunch of these to the dials on the OWL, and in the end I could come up with quite an array of sounds for not much code. I can even lock in a sound almost like a repeater, though without a button to map that to it’s not very practical.

The code is on github if you’re interested. I have to admit I was inspired a little on this by the Echorec video by That Pedal Show, which was the first video of theirs I ever saw, and in the end I just used my effect to badly play Pink Floyd riffs :)

A week in the shop

Making guitars is hard, which is also what makes it intersting and rewarding; I suspect this is true of most things that one chooses to apply oneself too. I know that's why I enjoy working in software and hardware so much. Like software and hardware though, it's very easy to get ahead of yourself and forget how hard some things are after some success (this is why software engineers are classically so bad at making time estimates).

With the prototype guitars, this month was when I finally admitted that doing the offsets was different enough from the t-styles I made before that I was going to end up assembling both a prototype body and neck before cutting wood on guitars #5 and #6. I was already building a prototype neck with cheaper wood, for the sake of a week or so I may as well do a trial run of a body. So it is I'm now assembling guitar #7, a prototype offset that will act as a guide for guitars #5 and #6.

Screen Shot 2018-04-16 at 09.54.54.png

This won't add too much delay to guitars #5 and #6, as I don't need to finish guitar #7 off to a high standard immediately, I just want to get it to the point where I've strung it up and checked it makes notes and the intonation is correct. The electronics and the final finish are bits I'm familiar with and I can do after guitars #5 and #6 (or at least alongside for the finishing stages, as that bit is so slow anyway).

When I set up the prototype series, the idea was I was already well versed in making t-styles, so it'd be a fairly quick and easy job to do a couple more. Due to what customers asked for, I ended up not doing t-styles, but rather these to being offsets, which has pushed me a lot more, and has been fun and interesting. But with the additional troubles caused by machine madness this last month and the gluing failure, my confidence has been knocked a little, and I want the guitars I ship to my customers to be the best guitar I can build. So to remove my doubts about the new design and process, I'll apply what I preach with measure, measure, cut: guitar #7 will let me be confident that guitars #5 and #6 will be as awesome as I expect, and more importantly, as awesome as my customers expect.

Overall the impact will be a minimal delay on starting on #5 and #6, all being well, but I'll be a lot happier in the end result, and over the lifetime of a guitar a week or so is not that significant. It'll also cost me a bunch of money to validate the design, something potential luthiers should factor into their costs better than I did; but eventually #7 will become another demo guitar, which is handy as one of my demo guitars just shipped out.

I really don't like delaying guitars #5 and #6 more than they already have been with the CNC router adventures of the last few weeks: I'm acutely aware I have people waiting on me, but I'm also not going to ship something I'm not happy with, and this extra prototype will let me get there.

So, that explained, let's go see what progress we achieved this week.

At the end of last week I'd got the prototype neck to the stage where it was glued and the fret slots cut. This week I moved it through until all that remains is the fret levelling and shaping.

First up, the inlays. Normally, if you're going for traditional dot inlays on a dark wood the neck, you'd just buy some plastic pearlescent disks and glue them in, which does look lovely (I did this on guitar #3), but if you look at the vintage guitars they used clay inlays, and I've seen quite a few custom guitar builders do the same of late. I was interested to try this technique and see how it works, and the prototype neck seemed a good opportunity to give this a whirl. There's two approaches one could take here: make the inlays apart from the neck and glue them in, or drill the inlay holes in the neck and fill them with clay and sand them down. I opted for the second approach, as it felt like less steps, and in the future it might be nice to do unusual shapes, where moulding the clay into the neck would make more sense.

Before applying it to the neck, I did test this technique on a scrap bit of wood. I drilled some 5mm holes in the wood and filled them with some air drying clay (with the best will in the world, one probably shouldn't fire a guitar neck, and besides, it won't fit in my oven ;). This is the scrap immediately after putting the clay in:


Then after leaving it to dry for 24 hours I sanded it down and I got this:


I'm quite pleased with the results, though it'll be interesting to see how well it wears over time. Still, I was happy enough that I applied this to the prototype neck I've been building.

Inlays sorted, I then went on to hand carve the neck. I used to fear this bit when I started, but now I've done a couple, I really enjoy the process of hand shaping a neck. In theory could do this in Fusion 360 and get the same result, but it'd take me several goes before I could know that the Fusion design felt right in my hand, whereas with hand carving the neck I constantly take both quantitative measurements with callipers and qualitative measurements with my hand to check it feels right. It's amazing to me how much a change of just 1mm on the neck profile will lead to a completely different feel in the hand. 


The back of the neck carved, I then radiussed the fretboard down, really bringing out a wonderful pattern in the wenge. It was also here that I sanded down the clay I'd put in the inlay holes I drilled to finally see how the looked in a real neck, and they look great to me. 


The final stage here was to fret the neck. I went into this bit full of confidence, and it was at this stage that I was knocked down a peg or two (sort of the inverse of how I feel about carving the neck). This was my 4th fret job, so I thought I was getting to be an old hand at this, but immediately I was struggling, with the very first fret refusing to go in and making a mess of the nice clean fret slot I'd cut. On reflection I'd done several things wrong, which I'll share here to hopefully save someone else the frustrations.

Firstly, I'd not fretted a neck for a a few months, and you forget just how hard you have to hit them to get them home. If you're about to do a fret job and haven't done one for a while (or at all) I'd strongly recommend getting some scrap, ideally of the same wood as your fretboard so it has the same hardness characteristics, and just knocking some frets in. You spend so much time being gentle with the wood on a guitar to protect the finish, having to really hammer hard the frets in feels (at least to me) alien and wrong. But if you don't they won't go in neatly, and there's a risk they'll start to deform the fret slot if you've not got them lined up perfectly. Swift, very sharp, taps are the order of the day here. It took me several frets to get back into the swing of things.

Secondly, I'd over radiused the fret wire, making my life harder for myself. Fretwire ships at a very shallow radius (about 12" or so). Ideally you want your fret wire to have a slightly tighter radius than that of your fretboard: this means you can then tap in the two sides of the fret wire into the slot, before just tapping the middle home. With this fretwire I made it a tighter radius than I needed, so I had quite a bit of more work to do getting the centre home, which means the sides are having to shuffle more than ideal. Basically I made things hard for myself here, so don't make your fretwire radius too tight.

Finally, I'd never used wenge before on a neck. Wenge is a funny wood: the two colours you see on it are very pretty, but one is quite soft and one is quite hard. The hard bits didn't want to give when I hammered the frets in, and the soft bits were happy to deform any time I struggled. Urgh. With the previous necks the wood had been quite uniform in nature, and that seemed to help. Wenge is definitely usable and it's just me not being experienced that is the issue here, but then that's another reason to practice on scrap before you try on a real neck.


Anyway, after much stressing and sweating, I got all the frets home and reasonably level. First job for next week is to level and polish the frets so I can then put this neck on the prototype body, and string it up to check my measurements are as good as I think they are.

Given the issues with the CNC router, I was still a bit nervous about how well it'd cope on a longer job, despite the fact it'd cut the neck well, so I decided to cut a test run body that cemented the idea of building guitar #7. This is very similar to the body for guitar #5, but rather than use a hardtail style bridge, which I think is actually the better choice for a working instrument of this style (and this is probably why hardtail are the only option Fender make any more on this type of guitar), I opted on this one to go for an old style dynamic vibrato bridge with a floating bridge and tremolo arm. Why do this? Mostly just because of learning: I've at least made four guitars with fixed bridges, and none with floating bridges, so given the opportunity I thought I'd go with the termolo. From a making of the body point of view there's only a very minor change: rather than drill the string holes through the body and bore out the ferrule holes on the back to seat the strings, you instead carve out two cavities for the springs on the bridge, and bore two holes to hold the pegs for the floating bridge. The complex bits of the body, the overall outline and the cutaways for the comfort curves, remain untouched.

So, design tweaked slightly for the new bridge, I set about making my first offset body in wood. Making a guitar body on the CNC router is a two stage process, as you need to do both sides in separate passes. To make sure everything is aligned you use a technique called pinning: you drill four holes equidistant from the center of your design in the corners of your material on the first side, then when you finish side one you remove the material and drill the same four holes in the CNC router bed, put metal dowels in the bed, and mount your material the other side up on those pegs. Now your design is perfectly lined up for you to do the second side. I'd always advocate doing the pinning holes first, as once you've made these, you can even abort your design run and come back another date: although the origin on the router may change, so long as you have reference holes on your material you can always make new reference holes in the bed and get back to a know position. During the making of this body about half way through the first side I knocked the material out of the mounts I was using, but because I had already made the pinning holes this wasn't a problem, as Just flipped it over, made pinning holes on the other side, and then could get back to being properly aligned. 


It took 5 hours to route the body, tweaking the Fusion toolpaths as I went. This is the first full body I've made from Fusion and the first time I've routed one with properly 3D surfaces, so it required a bit of reworking as I went along. After that it took me a few more hours of sanding and tweaking with hand tools to get to the shape I wanted. I don't do everything in Fusion, there's still quite a bit of hand work I do to the body afterward to get it just right: I rounded all the edges with a hand router, I adjusted the comfort carves to be just right using hand rasps and sanding, and so forth. Oh, and lots and lots of sanding, as the routers never give you a perfectly smooth finish.


Although it worked out all good, it was much more relaxing to do this on a non-customer body, and I'm now ready to cut the bodies for the customer guitars. The only thing holding me back is just double checking that I've got the intonation right by finishing the prototype neck and stringing it up. 

The one other thing I did this week that took up a bit of time to prep for was I ran a class teaching Fusion 360 to other Makespace members. I taught 10 people how to use Fusion to make a simple device (a kitchen roll/tape dispense) and then how to fabricate it on 3D printers and the CNC router. 


I did this mostly because people had asked, but it is also good for me: the more expertise there is using Fusion 360 in Makespace and the more people use it to target the machines in Makespace the better it'll be for me when I get stuck and need help. I've already seen people who were in the class doing things in Fusion I've never tried, which is great!

The little blue amp demo video

Somewhat belated, but here's a small video of our first amp build, the little blue amp, a small 5 watt tube based amp for home/rehearsal use:

The content isn't great, as I was a bit rushed that day trying to do too many things, but hopefully gives you a good idea of how the amp works :) My thanks to Tristan for salvaging something from the mound of video I recorded; next time I'll get him to film the lot :)