Friday, 27 July 2012

Update on Routing Out a Guitar Body.

I thought that I'll drop in a quick update on the earlier post (May 2012) on routing out a guitar body. This is just a quick picture of the finished guitar. Works fine but cheap pots and pickups are going to be dumped and replaced with something better. At least I got the bridge in the right place and the neck angle correct. There is an odd wavy reflection above the neck pickup, but this is just the reflection. It does show the arm-rest angle and the curve of the horns etc. The truss rod cover is bookmatched in case you are looking for one. Overall, very happy with how it turned out. The bridge pickup is tapped with a 47nF shunt capacitor accross the slug coil. Pickup bobbins and thin copper wire have arrived so I'll be winding new pickups after the summer break.

Monday, 16 July 2012

Build A Guitar Humbucker Pickup Frequency Response Analyser For Peanuts

The Shuntbucker frequency plots obtained in my earlier post were originally obtained using an Agilent arbitrary frequency generator and buffer amplifier as the source and the output amplitude recorded on an LeCroy 1GHz digitising scope. I am lucky in that I have access to some neat equipment, but most people won't.  In this post I'll outline how anyone with around £15 can do the same measurements and add thoughts on how to do full dynamic impedance measurements on pickups. I hope to try rewinding some pickups soon. The rational here is when winding pickups how do you know where you are going if you don't know which way you came from.

The first part is how to get a signal into the pickup in a way that is reasonably representative of a string above the pole pieces and is frequency agnostic. The issue is more complicated with humbucker pickups as they have two coils which produce an opposite signal to an open field magnetic source.  It will need some sort of magnetic injection which, with a bit of luck you can obtain for free. You will need to source an old computer hard disc drive and some star-point type screwdrivers. Open the drive, release the magnet assembly in the corner and extract the pickup head assembly as seen in the bottom left hand corner below. The bearing has a small screw on the side which will either need the correct screwdriver to remove or a pair of pliers and just grab the screw head and turn.

Once removed ditch the bearing, its magnetic so you will want it out of the equation. The pickup arm assembly can be removed with a junior hacksaw, its just the voice-coil that we are after. There may be a parking magnet as part of the voice coil, if so snap it off. When finished you should have something like this, well after soldering a lead on anyway. There will be some drive circuitry on a flex PCB mounted on the arm, It will provide a good connection point to solder to the coil, just remember to cut any tracks leading to the drive chip from the coil with a craft knife.

The coil should have a DC resistance of about 10 to 15 ohms, high enough to drive with a domestic amplifier. However my test kit has a 50ohm output impedance so I had to knock up a buffer amplifier. It is based on a OPA548F high power op-amp. Its data sheet shows a relevant schematic. To drive a volt pk-pk maximum into the coil will only require an amplifier with under a 1/2 watt capable drive. I'll come back to drive later as you don't have to make this part.

The reason this coil is ideal is that I can mount it centrally and orthogonally to produce a good response by working both coils at the same time.  It will produce an opposite polarity field into each coil thus producing a combined in-phase signal which is what I want and what a string does. The coil construction means that the self resonance will be decades above the coil under test ensuring good measurement. Below was an earlier attempt with a slightly different coil. I used Blu-Tack to hold it firmly in place. It runs exactly in line with the split between the two coils. In this configuration a symmetrical in-phase output will be produced, as you would expect by plucking a string. If I lift the coil and hold it horizontally above the pickup It will produce a matched out-of-phase signal which will combine and cancel as you would expect from an external field into a humbucker. Moving the coil away from this position say to the side changes the output balance and phase of the coils. This would produce cancellations when attempting to measure the frequency response. 

The output of this was fed to a board with resistors and capacitors representing the volume and tone circuit of the guitar and the 1Meg input impedance of my amp. I used a 10Meg input impedance scope probe so not to load the circuit.  The coil idea proved to work very well and produced repeatable results even after disassembly and reassembly. However most people can't afford a scope, signal generator and probes, and my issue is that I can only have limited access to them at work. However I have a potential solution which although I have not yet tried _should_ work for anyone  wishing to measure pickup response. (See EDIT below)

Visual Analyser 2011 XE (VA) from is a free download unlimited real-time software scope, spectrum analyser and lab tool-kit. It uses a PC sound card to provide a real-world interface. The coil should be high enough impedance to be driven from a good headphone output. If your card does not have enough drive try lowering the output or wiring the left and right output together. You can also buffer with a HiFi amplifier on a very low setting. For the input you will need to wire a high impedance buffer between the Humbucker coil and the line input. Any high impedance audio op-amp buffer circuit should suffice. Generally the input impedance of a soundcard is quite low when compared with a guitar amplifier input stage. Most valve amplifiers have a 1 Meg input. The scope function will measure the output amplitude provided by the pickup when driven with the VA software signal generator.  With this configuration your total kit outlay will be the cost of a broken hard drive and building a buffer amplifier. If appreciative please consider a Paypal donation to the software engineer who wrote Visual Analyser. I have no ties to the author of VA who I believe is from Italy.  The software was used in the analysis of the triode / pentode switch mod harmonic test in an earlier post.

EDIT: I have now ran a quick test using VA and a normal Dell computer with an on board sound card. I placed a 470K resistor in series with the line in jack to prevent the pickup getting loaded by the low line in impedance. I used the left and right headphone output combined to drive the HDD coil. The sweep was set to 24 points an octave and the spectrum analyser set to peak hold. The waveform generator was set to sweep from 80Hz to 20kHz. I left it to run for a while to build up the plot below. The cursor measurement shows the peak at 5.758kHz. You can also notice that from 2.5kHz the amplitude rate of change increases towards the resonance peak.  It will look different from my previous plot as it records in dB and not straight mV. (Click on the plot to see full size)

On the Sillanumsoft website it also has links to an article (you may wish to google translate if English is your native language)  to add an interface of two op-amp buffers before the soundcard input and convert it to an LCR bridge. If you wish to attempt true impedance measurements then this would be well worth a look.  Another "also" is the frequency counter with pedometer function. I intend to look at this when I start winding pickups. I have a motor with a micro-reed relay switch and magnet that now gives one clean pulse / revolution. I reckon that I'll be able to use this as a turns counter. The 1+= calculator trick is a little unstable at higher speeds but that's for another day. Have Fun!