GuitarFreak - guitar frequency response calculator

[JohnH]

Jan 20, 2021 8:43:52 GMT -5 pougnetpedals said:

New to the site. Am I missing something? It doesn't look like the link to the GuitarFreak macro works anymore? Is it still available?

Yes its there, I just checked it, download the excel file from the link at the top of the first post.

[David Mitchell]

I almost hate to ask, but what are the chances of an update for LibreOffice? Even the last version posted for it seems quite glitchy in v7.something that I have (updating mine to the latest stable version and will check again), and of course it doesn't have the newer pickup models.

I tried 6.6 in LibreOffice, online Excel, and Google Sheets, but sadly it doesn't seem to be compatible with any of them. I will gladly try again if I missed a trick!

[MattB]

The 97-2003 version of 6.6 mostly works in LibreOffice.

I can't save traces and the no-load tone setting doesn't work, but everything else I've tested seems ok. The reference trace button does work, so it's still possible to compare traces two at a time.

I copied the pickup models I wanted over to version 6.4, so I can use that if want to make graphs.

If you try this be aware that the dB parameter varies between the two versions. This doesn't appear to affect the shape of the waveform, just the overall level, so I think it's  only really important if you want to compare between different pickups.

[JohnH]

Thanks for trying GF. I recall that getting the Libreoffice one to work took a lot of fiddling to move it over from excel, and I don't have it installed any more.

[Wim]

JohnH, this is an amazing tool. Do you know if anyone has any measurements from Jazzmaster pickups?

[JohnH]

Hi Wim, thanks for your comments. All of the pickups modelled in GF are derived from testing carried out by antigua. His tests are very comprehensive and over the years he has made a very substantial collection of results. Here is one on Jazzmaster, which also refers to another:

guitarnuts2.proboards.com/thread/8311/fender-vintage-jazzmaster-analysis-review

In principle, I could take that and work out values to put into GF. But I haven't done that so far. But you can gain a lot of insight by looking at the graphs and comparing them to others. It looks like the B and N pickups tested were very similar, and much like a typical hot single coil. From those already modelled in GF, try the bridge pickup in the 'Strat Texas Special' set to get fairly close. Also try SD-SSL1 which looks even closer and was modelled in GF6.4 onwards.

[David Mitchell]

Feb 2, 2022 3:20:07 GMT -5 MattB said:

The 97-2003 version of 6.6 mostly works in LibreOffice.

I can't save traces and the no-load tone setting doesn't work, but everything else I've tested seems ok. The reference trace button does work, so it's still possible to compare traces two at a time.

I copied the pickup models I wanted over to version 6.4, so I can use that if want to make graphs.

If you try this be aware that the dB parameter varies between the two versions. This doesn't appear to affect the shape of the waveform, just the overall level, so I think it's  only really important if you want to compare between different pickups.

Thanks! I think I forgot to go back and try that version after enabling macros in LibreOffice, which, of course, makes a big difference. Oops....

Feb 2, 2022 4:55:12 GMT -5 JohnH said:

Thanks for trying GF. I recall that getting the Libreoffice one to work took a lot of fiddling to move it over from excel, and I don't have it installed any more.

That's fair. GuitarFreak is a staggering thing to someone such as me who can do a few things with spreadsheets, but certainly not this....

I'm getting this with the 97-2003 version of 6.6 in LibreOffice 7.1.8.1. Does it look like it's working mostly as expected?

[JohnH]

Hi David, I think that's working basically fine. Here it is in the latest Excel 365, set to match your plot:





The colours on the Libreoffice plot are more in-yer-face, plus some of the graphic objects have moved around. But the basic maths seems to be working and if you compare the tabulated numbers you can see that. The blue dashed trace should become a copy of the red trace if you press the reference button, similar the save and clear buttons. Do those buttons work? how about the sliders? In excel, they slide smoothly but i think in LO they may jump.

Im grateful for any other observations of how this version runs in LO

cheers
John

[asher]

What does Ld stand for? How does it differ from the inductance measurement labeled L1?

[JohnH]

Apr 10, 2023 12:29:49 GMT -5 asher said:

What does Ld stand for? How does it differ from the inductance measurement labeled L1?

Hi Asher, thanks for your interest in GF

The models that it uses have 6 components, and as you have seen, there are two inductances. I think of Ld as a secondary element that with Rd and Rl is part of the natural damping of the pickup due to eddy currents and magnetic losses. With the other parts connected to it, it helps to shape the height of the peak, and also the dip in the high mids that some pickups show. I cant really point to a single particular physical characteristic of the real pickup that is Ld. Instead, the whole 6-part model is tuned to try to capture the overall response of the pickup as a whole under different loads and frequencies . When you measure a real pickups inductance, then most of that will appear as L0, but Ld is chosen to try to match the more complex responses of the pickup, which cant be captured by just a single L component.

The 6-part model is still just an approximation, but by tuning it to follow the measured loaded and unloaded responses as closely as possible, it is a fairly good representation and good enough for using to design guitar circuits. There are better more rigorous approaches to pickup modelling discussed on this forum, but this one has the advantage of being easy to implement and to include in spreadsheets and Spice models, and its fairly easy to derive directly from the bode-plot measurements so that a collection of different pickups can be made.

[asher]

Where should I be looking to find more details about how this model is understood, particularly measurement vs derivation? I didn't see it reading through this thread— did I miss it? 

I understand everything you are saying conceptually but I would like to see more concretely how you arrived there.

Thanks!

[JohnH]

Ok sure. Here's some more...

I tried a few different arrangements of parts before settling on the topology of the 6 components, starting with 3, 4 and 5 parts. With 6 parts, in a sensible arrangement, you have 6 variables with which you can target 6 parameters. If its a good arrangement, getting 6 points right will also keep close to other points too.

So what 6 points to target?

The first 4 are the height and the frequency of each of the treble response peaks for loaded and unloaded bode plots, based on Antigua's tests.

The 5th parameter is to match the dc resistance as measured for the pickup.

The 6th parameter is less 'defined' and its to control the rise towards the peak to shape its width, and make it dip if needed. This last point is a softer parameter to make the shapes follow better. Most pickups can be modelled in this way to match the plots within a dB or better across the frequency range up to and above the peak.

Doing the processing to work out these numbers is circuit analysis based on complex numbers, with iterative cycles of optimisation, one at a time. Typically, it starts with an assumptions, then optimise to hit one parameter, and then another and the others. Each step pushes the preceding targets off a bit. But after many cycles of the whole group, they all converge to a good solution to match the plots. Then the 6 values are recorded.

[asher]

If I'm understanding correctly, what equations are you plugging this data into in order to work through the circuit analysis as described?

Is there any chance you have a sort of "show your work" example? I know that might be a lot to ask.

[JohnH]

Yes that'd need me to write some kind of a text book! I'm not set up to do that right now though.

[gckelloch]

Hiya John. I'm finding V6 really helpful with a Filtertron guitar project I'm working on. Just curious why the cable C can't be set to a custom value?

[JohnH]

Apr 12, 2023 17:22:46 GMT -5 gckelloch said:

Hiya John. I'm finding V6 really helpful with a Filtertron guitar project I'm working on. Just curious why the cable C can't be set to a custom value?

You can! It's capacitor C4 at yellow cell L14. Just type in a value in nF.

[weka]

Hi John:

I'm new to guitarnuts forum but very impressed with your library of technical info I've had a chance to peruse. Outstanding.

Thanks very much for those contributions!

Wes

[someguy]

This is such an awesome piece of software, but IMO it's a pity that one has to have Microsoft Excel installed to have all features working. It would be great to have this ported to a web app or something. At times I get the urge to do that myself, but it's really hard to decipher the semantics because of the Excel nature of it - it's all cell numbers with no meaningful names.

[JohnH]

hi someguy

Thanks for your interest in GF. It is due for an update, and I haven't done anything major to it in a few years. It does rely on quite a few excel features, mainly for buttons and sliders, but the maths of it is quite simple.

Porting it to other systems isn't perfect. But the nearest version that runs on a free software is from GF6.4 where there is a link to a Libreoffice version.

Google also has a spreadsheet, but I haven't tried if that might work.

Meanwhile, most people who work in offices, or study at colleges etc get access to excel in some form.

[someguy]

Hi JohnH

Thanks for making it! Yes, the LibreOffice version does the job, I've been using it for several years now. I think the Excel version won't work on Mac anyway because of the ActiveX requirement. Google Sheets could be a good cross-platform solution (if it works).

I'd like to take a shot at making a JS version, if you don't mind. Is there anywhere I can find just the condensed math you're using?

[JohnH]

I'm sorry but I don't have it written out. And there are several parts that combine to make the whole thing. I tend to just go straight in, and verify using a Spice program for some specific test cases.

The main electrical parts are basically a series of sequential voltage dividers, from left to right. If you can grab the very first GF version 1, it's probably the easiest to unpick. It's all based on representing impedances as complex numbers. Resistance is horizontal, inductive reactance is up and capacitive reactance is down. There's a text book needing to be written to fully describe!

Assignment 1: set up a basic RC circuit....

[someguy]

Got it, thanks. I'm a bit familiar with this stuff from my studies in audio DSP, so maybe I'll figure something out if I have the time

Could you give me a hint on what pickup model / equivalent circuit you're using in the latest versions?

[ssstonelover]

Oct 6, 2008 1:26:26 GMT -5 JohnH said:

Edit 05/12/2019

The latest version is now GuitarFreak 6.6, which is a minor enhancement since versions 6.5 and 6.4

GuitarFreak6.6 (excel 2013, 2016, 365)

Version 6.6 should also run in Excel 2007 and 2010 but I haven't tested it.

I also down-saved it to Excel 97-2003:

GuitarFreak6.6 (excel 97-2003)

John

Guitar Freak is not downloading currently, or I'm doing something wrong

[newey]

JohnH-

I don't know if this is the same issue as ssstonelover is having. but here's what I get when I try to open it:

[JohnH]

Thanks for the alert and the clue! I havnt checked the download feature in quite a while, but I'll see if I can figure it out and fix it soon when I get some quality IT time.

Meantime, if you'd like to pm me an email address I'd be happy to send it to you. Let me know what Excel version you want to use too.

[ssstonelover]

JohnH I also did see that message newey reported, but not every time I tried to download it (weird) today. Weirder yet, a few days ago clicking on the link opened Excel.....
As it turns out I have a copy of the file, so I'm OK, so this was more a test on my part than a need, but I had given the link to a friend last night (and I do hope that is OK) and he reported today he could not open or download the file, so that is when I noticed the issue.

[stevewf]

Aug 6, 2024 12:25:14 GMT -5 ssstonelover said:

Oct 6, 2008 1:26:26 GMT -5 JohnH said:

Edit 05/12/2019

The latest version is now GuitarFreak 6.6, which is a minor enhancement since versions 6.5 and 6.4

GuitarFreak6.6 (excel 2013, 2016, 365)

Version 6.6 should also run in Excel 2007 and 2010 but I haven't tested it.

I also down-saved it to Excel 97-2003:

GuitarFreak6.6 (excel 97-2003)

John

Guitar Freak is not downloading currently, or I'm doing something wrong

Weird. Both links work consistently here, using Safari browser on a mac.  

[ssstonelover]

I think this is a Chrome or Win-11 issue.

stevewf, I retried using MS Edge as the browser, and the correct link popped up with the message asking if I would like to save it to my computer.
So, that works

However no matter what I try using Chrome, no link appears.... settings issue, security issue, IT issue, who knows?
How maddening...

[JohnH]

Thanks Beth, I'll be at my desk later today and I'll see what my Win10 and Edge office pc will allow.

[stevewf]

A workaround, perhaps, follows. But for now, maybe this text works:
GuitarFreak6.6 (excel 2013, 2016, 365)     https://www.jocidapark.com.au/circuits/GuitarFreak_6_60_051219.xlsm
GuitarFreak6.6 (excel 97-2003)       https://www.jocidapark.com.au/circuits/GuitarFreak_6_60_051219_97-2003.xls

copy one of the URLs above and paste it into your browser. DON'T just click it. Mouse-drag the text itself.

[^^^ edited, replacing "http" with "https", and using [noparse] as mentioned by Yogi B, further down this thread]

---

If that works, then the trouble lies in what I'll call a "domestic link interpreter" (service provided at the redirect.viglink.com address).
If that doesn't work then the trouble lies in your browser (which depends on your operating system) or internet provider.

---

Maybe the problem extends beyond only John's spreadsheet, and occurs at other parts of the forum too. So here are some troubleshooting steps, using the GuitarFreak link as an example.

Right-click the hyperlink and choose "copy link [to clipboard]" (sorry, I don't remember how it's done on Windows)

Paste that link in some temporary app like notepad or even your browser's URL field

I got the following:

http://redirect.viglink.com/?key=71fe2139a887ad501313cd8cce3053c5&subId=1212256&u=http%3A//www.jocidapark.com.au/circuits/GuitarFreak_6_60_051219.xlsm

Remove part of the URL (which I've stricken out in the example above; it's everything between the two pairs of // forward slashes, and including one of the pairs);
Add the "s" to make your browser use the site's certificate (i.e. https)

https://www.jocidapark.com.au/circuits/GuitarFreak_6_60_051219.xlsm