Official Luthiers Forum!

Where are you guys getting signal generators for modal testing/tuning guitars? The passing of Don Bradley means the Sig Gen unit is no more - unless he passed on the reigns before his death...

What are you guys using?

John I use Visual Analyser, which is freeware, on a Windows PC to do 'tap testing' to get a spectrum of top and back response. From these spectra I can get the main resonant frequencies - monopole, cross dipole etc.

The software also provides a sine wave generator which I feed into an old domestic amplifier and into a cheap car speaker to provide excitation of the panels to give the Chladni patterns. The sig generator is discrete, ie doesn't do a sweep, you have to set the frequency explicitly.

This provides a very cheap and effective method. From the forums there are lots of other software spectrum analysers for Apple and Android devices which people rate.

This technique I have learnt from Trevor Gore's work by the way.

Dave

There are lots of software sig gens out there, for phones, tablets, PCs and Macs. Some are more usable than others. For "home" work, almost all are at least adequate. What tends to let them down are the minimum frequency increments you can get and the ease of adjusting the frequency. The best software ones will allow frequency adjustment with a mouse wheel with 0.1Hz resolution. I did find one Russian app for PC which did this very well, but it seemed to come with a load of malware so I junked it. I don't even remember the name now.


Good ol' Visual Analyzer actually has a pretty reasonable performance, but is for PC only unless you load some sort of emulation-type package. I'm still using Version 9, which works fine on XP and Win 7, but may be problematic on Win 10. But there are later versions which probably work fine on the newer Windows platforms.


Well, VA Version 9 does a sweep and frequency can be adjusted with a mouse wheel (but sometimes with a rather coarse resolution). Here's how to set it up in v9, but the later versions will be similar.

For the sweep go to the "Wave" tab on the main window and on the sub-window under "Main" go to "Wave function" and select "Sweep" from the drop-down menu. Then go to the "Sweep" tab and select your sweep limits.

To manually sweep the frequency with a mouse wheel, go to "Wave", "Main" select "Sine" under "Wave Function" and a slider appears for you to adjust the frequency. Click once on the slider and then use the mouse wheel to adjust the frequency. Today I'm getting 1Hz increments. On other days I've got 3Hz increments and I haven't figured what controls this. If you click in the frequency number box to type in a frequency manually, the wheel will stop working, but will start working again if you click on the slider.

You can output a sweep and measure a response at the same time, if you want.

If you use a software sig gen, you still need an amplifier and speaker. As I travel a lot teaching this stuff, I wanted a box that "did it all" and a separate speaker, so I built this (sufficient build hints, rather than detailed instructions, included).



Read the whole thread (important details lower down), including the link to the courses I presented at Robbie O'Brien's.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

Any chance somebody is making one of those for sale (hint hint..)

Not me!

It would cost me more in compliance certification, etc. etc. to satisfy our nanny states than I would ever make in profit out of it.

Putting mine together certainly gave me an insight into the time and effort Don (and predecessors) must have put into the Bradley unit.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

A friend of mine is working on an all-in-one lutherie app for Android. He's got a nice signal generator setup all worked up. He has linked it to a small Bluetooth speaker, and linking to an amp should be no problem. He's approaching the beta test phase now.

Well just goes to show what I don't know!
Sorry for the wrong information on VA.

Dave

I went poking through Amazon and there are a ton of fairly cheap signal generators out there that plug into an outlet..

Say I start with a unit like this... Looks like I would need an amp to drive a speaker... Perhaps a cheap all in one 4 watt electric guitar amp/speaker combo.... But besides that - what is the downside to something like this.

Is there a different product I need to be looking at?

http://www.amazon.com/gp/aw/d/B01A0TSHR ... ator&psc=1

Or this - which appears identical except for brand.

http://www.amazon.com/gp/aw/d/B071HJ31W ... ref=plSrch

You'll need something like 12W to do 'free' plate mode testing, and more if you get into arch tops. Use a speaker that's rated for at least twice the output of your amp for this: sine waves are hard on speakers, for one thing. Also, the rating on most speakers I've seen is the 'peak' rating; what the thing can take as a short burst without blowing out. For continuous duty like this you need twice that: a '12W' speaker can't be run much over 6W of steady power for any length of time. Get the smallest diameter speaker you can that will handle the power. Ideally you'd like to couple your signal generator with a point on the surface of the plate; a big speaker can overlap node lines and cancel a lot of the power out.

Actual signal level sine wave generators are fairly easy to come by. We got a passable sine wave from an analog circuit using an 8038 chip back in the '80s. There was a GAL Data Sheet on that, authored by the designer, Matt Fichtnbaum. My folks and I put those out for a few years, and Don Bradley took that business over when we could no longer keep doing it. He then came out with the processor based unit he was selling. His was MUCH better.

The frequency counter in the Data Sheet was a jury rig designed around Radio Shack parts that seldom worked off the bench, and was hard to trouble shoot. Once you got it working it was pretty good. The Shack is gone, and so are most of those parts. Single chip counters are available, and we switched to one toward the end.

I still use my old hardware unit, and am unlikely to change at this point. These days it makes more sense to do it with software. As with Don's unit, you can get a cleaner wave form, higher resolution, and a wider frequency range digitally.

This may well be a stupid question, but here goes anyway. When Chladni testing, should the guitar back be resting on the bench, or should it be supported off the bench somehow so it can interact with the top?

Thanks in advance

Steve

So... If I hook up the system like this
Function Generator ---> Amplifier ----> speaker

A. Is there any obvious reason that the $50 function generator I linked won't work or is a bad choice?
B. Can I just use a cheap guitar amp or do I need a real audio amplifier?

Thanks

JSDenvir asked:
"When Chladni testing, should the guitar back be resting on the bench, or should it be supported off the bench somehow so it can interact with the top?"

Here's where you run into differences of opinion. It boils down in many respects to what you're trying to find out. If you want to know 'the real frequency' of, say, the 'main top' or 'monopole' mode, then you want to isolate the top as much as possible from any other resonance that would couple with it. That means, at least, that you want to block the sound hole and, possibly, damp the back, usually by resting it on the bench. The problem here is that when you're playing the guitar the hole is open and the back may or may not be damped, depending on how the player is holding it. Since what happens in the playing position is what people are going to hear, maybe that what you want to test. But, of course, then you're seeing a combination of stuff, and it might be hard to disentangle it.

I like to test them by supporting the guitar up off the bench with foam pads around the edge, so that the back is about as free to vibrate as possible. This is not really 'free', as it would be, say, when the guitar is hung up from the head, and it's not really like what happens when somebody holds it, either. What it is, is easy to do and reproducible. Any test that can't be reproduced is not worth much, and any test that's too difficult to do won't get done, so that's not much help either. In the end, there's no such thing as a perfect test: you have to settle on something you can do and then always do it the same way. At least then you can compare your data.

One point: if you flip the guitar over to look at the back modes, be sure to support it at least one soundhole diameter above the bench. Otherwise the bench top interferes with the air flow through the hole, and changes the 'air' resonant pitch. This couples with both the top and the back, so changing it will mess up your data.

Most function generators should work well if they produce a fairly clean sine wave, and are stable enough. The old analog rig we made had some higher harmonic in the signal, and usually worked well. Sometimes you'd activate a particularly lively mode when driving the plate an octave lower, which is confusing until you understand the problem.

Any amp that sounds pretty good should work well enough for Chladni testing. Disconnect all the effects...

One thing you really need to know is the frequency of the signal you're generating. A decent sig gen should do that, but worth checking that you still have the accuracy and resolution in the low ranges we are interested in. A lot of these things are sold to radio/electronics enthusiasts who are usually working in the MHz range rather than the few tens of Hz we are interested in.

Almost any transistor amp should be fine. You need at least 15w RMS and if using a guitar combo you want its internal speaker to cut out when you plug in the Chladni speaker (or you need to temporarily re-wire it). A small dedicated "hi-fi" amp doesn't cost much if you want to go that route. You need to be able to reach the frequency and volume controls when you're doing this stuff, so don't forget the ergonomics.


It's not uncommon to get some higher harmonics in there due to interaction between the speaker and the amp. It's not usually a problem, but just something to be aware of on the odd occasions it is. A spectrum analyser app (good ol' VA, for example) will show you what you have if you need to find out.


What Alan said. Always be aware that you may be testing your bench and your room as well as your guitar!

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

This is my standard plug for a Mac-compatible signal generator app. I've looked at many signal generator apps for Macs, and for ease of use, accuracy and cost, my favorite is AudioTest from Katsura shareware. I checked it with an oscilloscope, and there's no DC offset; the signal is clean and accurate. With the mouse cursor over the frequency text box, the scroll wheel can be used to change the frequency, it can be typed in for a starter frequency, then swept with the mouse. It displays the frequency, which updates in real time. Mice with a "clicky" scroll wheel work best--Apple's wheelless mouse won't work. The resolution with the scroll wheel is 1 Hz per click. I've been in touch with the developer to get that selectable, and to make play/stop controlled by the space bar. He's considering making those improvements for his next revision. My old shop computer is connected to an old Pioneer receiver with off-the-shelf cables. The 4" speaker comes from PartExpress. It all works well, except that I have to keep the speaker from getting too warm. I still fry one every now and then.

Pat

_________________
formerly known around here as burbank
_________________

http://www.patfosterguitars.com