Analogue signal generator (and maybe more related questions)

[Kim]

Given the level of accuracy we are trying to tune our various frequencies to and the closeness of notes in the lower range it might be worth knowing that your plate actually resonates at 169.6 rather than 170. From memory some of the gaps are only 4Hz. So a sweep setup along with spectral analysis using a mic seems like the best option.
Dom

Thanks for the clarification Dom.

Cheers

Kim

[matthew]

http://www.nch.com.au/tonegen/index.html

is free and what i have been using ... tone is continuously adjustable with mouse or arrow keys.

a knob is a much nicer control though.

[kiwigeo]

http://www.nch.com.au/tonegen/index.html

is free and what i have been using ... tone is continuously adjustable with mouse or arrow keys.

a knob is a much nicer control though.

Have had a quick play with this package and even though graphical interface isnt quite my cup of tea compared to Audiotest, Tonegen has a few more features......being able to produce simultaneous multiple waveforms of differing frequencies is quite nifty.

[Taffy Evans]

Wow! this thread takes me back 30 odd years or more, when I was playing with this sort of equipement. I have not read it all yet but in my rig I also had a dual trace ocilloscope, and a big strobe light system that I could freeze the motion of the plates if I wanted. When I think back to the long nights spent collecting data, I wonder how many more guitars I could have built. Intersting to read the aproaches taken by members now days.

[Dominic]

Thanks Matthew, this one looks good. I've downloaded about 5 so far and none of them were any good. I had given up on software SGs. This looks like the best I have seen.
Martin, that one for Macs you sent a link for looked pretty good as well.

One thing I have just read about speakers is that the output frequencies can be impacted by the way the speaker is mounted to the box. If the speaker is inset at all and not flush they say it will change the frequencies. If correct, this suggests that no matter how accurate your dial is on your generator you should measure the frequency at the guitar because it could be different than the signal produced. I don't know how sensitive this effect is but if they are saying you will get audible differences perhaps its something to consider.

Those with old analogue generators will use a meter anyway. But using a software version it might be worth checking what frequency your speaker setup is actually producing.

Cheers
Dom

[kiwigeo]

Wow! this thread takes me back 30 odd years or more, when I was playing with this sort of equipement. I have not read it all yet but in my rig I also had a dual trace ocilloscope, and a big strobe light system that I could freeze the motion of the plates if I wanted. When I think back to the long nights spent collecting data, I wonder how many more guitars I could have built. Intersting to read the aproaches taken by members now days.

Pay day coming up.....might get myself a laser interferometer

[matthew]

One thing I have just read about speakers is that the output frequencies can be impacted by the way the speaker is mounted to the box. If the speaker is inset at all and not flush they say it will change the frequencies.

That would imply that pitch of music changes depending on what speakers you use! Iseriously doubt that the FREQUENCIES will change, no matter how you mount them!

Frequency REPONSE maybe, but not frequencies. And certainly not discernably.

[Dominic]

Thanks Matthew, that makes sense. I don't know anything about audio tech stuff apart from where to put the wires so I have just been reading what I can find. And this site mentioned it so perhaps they meant change the response when they said change the frequency. Or perhaps they just don't know what they are talking about. Lots of info on the internet but it still takes a big investment in time to sift the good stuff from the rubbish.

Nonetheless, I will still check the frequency output with a good tuner as I would imagine the digital processing and the digital to analogue conversion and the processing by the amp all have the potential for errors to the actual frequency and the counter on the generator or in the software also could be poorly calibrated.

Better safe than sorry.

Dom

[peter.coombe]

Frequency response certainly can change depending on how the speaker is mounted in a box, but if you are feeding a pure sine wave into the speaker that is what the speaker will reproduce. There is no need to measure the output of the speaker. There will be a small amount of higher higher harmonics because no speaker is distortion free, but they can be safely ignored. How the speaker is mounted affects the high frequencies only, and they are high enough not to be an issue here (>1khz). The frequency response lumps you get are due to diffraction effects of the sound waves on the discontinuity caused by the speaker frame. If you flush mount the speaker then the diffraction effects disappear and the frequency response is smoother. You also get diffraction effects from the corners of a speaker box, but also only at high frequencies. Don't worry about it.

[matthew]

And the other thing is ... the frequencies you will be measuring are YOUR set of data. Your results most likely won't match Peter's or Martin's or Graham's or Trevors's or whoever. So it doesn't really matter what the frequency meter says, as long as it's consistent. You're using it as a tool to measure your own work, observe what happens when you make changes, and try to keep some things consistent. The numbers mean very little of themselves. So yeah, don't worry about it!

[peter.coombe]

And the other thing is ... the frequencies you will be measuring are YOUR set of data. Your results most likely won't match Peter's or Martin's or Graham's or Trevors's or whoever. So it doesn't really matter what the frequency meter says, as long as it's consistent. You're using it as a tool to measure your own work, observe what happens when you make changes, and try to keep some things consistent. The numbers mean very little of themselves. So yeah, don't worry about it!

Yep, spot on. It is the relative frequencies that matter, not the absolute frequency. So many people ask - what frequency should I tune the top (or back) to. Answer - it doesn't matter. They should be asking what frequency should I tune the back relative to the top to? That does matter, and it matters how you tune the current instrument relative to your reference instrument (which might just be the last one you made). And, it is important to have a reference instrument because the human ear has a short memory so you always need that known reference point. So, if your signal generator frequency counter is 10Hz out, it doesn't matter because it will always be 10Hz out.

[jeffhigh]

Well it might matter....
One of the key points of Trevor's books is NOT to place body or top resonances exactly on scale notes so that you avoid wolf notes.
The idea is to have them midway between scale notes.

[matthew]

haha OK. Not something I come across much on a non-fretted instrument.

[Dominic]

That's what I was thinking Jeff. Given we ultimately tune the strings with a quality tuner, it probably worth setting the various frequency peaks using the correctly measured frequencies otherwise you may find you have built closer to a note than you thought. And then given we are all probably tuning in 440 then comparisons across different builders should indeed yield meaningful data sets for comparison.
That’s the fun bit. I saw a dumb comment on the OLF about the design book and the guy was basically seemed to be saying he had been doing this stuff for ages and it was nothing new. But this guy had never been willing to share any of his techniques or his results if he in fact had any. So i am hoping that in the spirit of sharing set by Trevor and Gerard, that people will share any results they may have using these techniques and perhaps give some numbers for their good guitars. We could then compare the numbers for all our good guitars and look for correlations etc. By sharing like this we all could benefit from our collective experience and be making better guitars sooner. That’s the dream anyway.
Cheers
Dom

[kiwigeo]

That’s the fun bit. I saw a dumb comment on the OLF about the design book and the guy was basically seemed to be saying he had been doing this stuff for ages and it was nothing new. But this guy had never been willing to share any of his techniques or his results if he in fact had any.

Whats the bet the guy hasn't read the books? If hes has then he wont be needing them and will be wanting to offload them

[charangohabsburg]

That’s the fun bit. I saw a dumb comment on the OLF about the design book and the guy was basically seemed to be saying he had been doing this stuff for ages and it was nothing new. But this guy had never been willing to share any of his techniques or his results if he in fact had any.

Whats the bet the guy hasn't read the books? If hes has then he wont be needing them and will be wanting to offload them

I had read that comment (he deleted it 3 hours before answering to Trevor's reply) and I if remember well he had mentioned there that he has not read the books. It seems that he thinks that The Book might be dangerous for the luthierie school he runs.

[ChuckM]

I find that speaker size (and placement) is important, but it depends on what you're trying to do. I've used 2.5 inch (~6.4cm) and 5.5 inch speakers and found that the larger one tends to make everything look like a monopole. I tried a 12 inch and all I could get was monopoles. With the smaller speaker its easier to excite the smaller anti-nodes (even at lower frequencies) and see the variety of modes that exist as you sweep upwards.

[Dominic]

Talking about speakers, perhaps Marcus has the right idea by getting a bass speaker. I have a decent speaker but it does not like the lower frequencies very much. Particularly the lowest frequency peak at around 100Hz freaks it out a bit and I have to turn the volume down.
I know you carved plate guys use a lot higher frequencies than us flat top guys and instead go for raw power so perhaps have not had this issue but if you have any tips?
Chuck do you use small low frequency speakers?

Cheers
Dom

[ChuckM]

My experience is that small and bass in speakers don't mix. I end up using the larger speaker for up to 100Hz or so. Above that I use the smaller one. It's not that the smaller one won't excite the lower frequencies, but the volume has to be turned up and it doesn't really do much below about 50Hz. With the 5.5 inch speaker I can get modes down around 20Hz to show without a pushing a lot of power.

I haven't found much use for it yet, but with the smaller speaker it is often possible to get dipole and tripole modes to show between 50 and 140Hz. The larger speaker tends to overpower them and you don't really see what they are. Normally you wouldn't care about these modes, but I built a fan braced classical guitar with a round bridge and the dominant top mode turned out to be a cross tripole, T(1,3) which was lower than it would have been with a standard classical bridge. The coupled top frequency was 156Hz. Before the bridge was attached it was 220Hz.

[charangohabsburg]

Interesting bridge design Chuck! Apparently (and I think not totally surprisingly) it implies some dramatic changes in the mode distribution and general dropping of the top frequencies. How high are the peaks of those low-frequency modes you mention, compared to the main peaks?

[...] So if your sine wave generator outputs more than +-2V then you are in danger territory. It only takes a few seconds and the speaker is toast. The easiest and simplest protection is a fast blow fuse or polyswitch. I have a polyswitch in the amp that triggors below 120W. The polyswitch is connected in series with the speaker and warms up as the power goes up and after a certain current level (and heat) is reached it's resistance shoots up, thus greatly reducing the current and hence power to the speaker. Turn down the volume and wait, the thermister cools down and everything goes back to normal. You need to match the triggor point of the polyswitch with the individual speaker power rating.

I've been diving now a bit in the shallow waters of eBay and similar sites and found this polyswitch-calculator.

My next inevitable questions as an electronics noob are:

• The wattage of the bass speaker I ordered is rated as 75 - 150 Watt max. (there's nothing written about RMS). I guess I have to calculate the polyswitch for the lower vale, 75 Watt. Is my assumption OK or do I miss something?
• When searching on eBay for example for a 1.85 A polyswitch there's always mentioned also a maximum voltage, typically something like 30V, 60V, or 75V. That's also just OK I think. Or are these completely wrong items?

I am reluctant to buy from that German provider who has that polyswitch-calculator on his website because for sending a 1.50 Euro polyswitch they would charge me 17 (yes, seventeen!) Euros!

Peter, I loved the picture above the text "Glue the second brace and wait 24hrs for the glue to dry." on your plate-tuning page!

[Trevor Gore]

I mentioned earlier that I was getting together some details on a sig gen designed for Chladni testing. I finally got the stuff together and it's in the attached .pdf. The unit is built by Don Bradley, who I met when I was in California a couple of years ago.

The unit is a combined sig gen and amp, up to 45watts. It's a digital sig gen, so you dial in the frequency that you want and that's what it gives you, so you don't need a separate frequency meter. Don, in the attached write-up, describes the frequency selection as a 16 position switch, which is in fact a rotary switch that works pretty much like an analogue knob.

So with this one small unit (and a speaker) you can create Chladni patterns and you know what frequency they're at. No need to switch your computer on! The price is US$500. Not the cheapest, but maybe the lowest cost way out if you're "electronically challenged" or can't find the bits to cobble together a set-up from obsolete gear or kit parts.

If you have any queries, please direct them to Don, not me!

[ChuckM]

Thanks for that information Trevor. That is the person I was thinking of.

The following plot shows the top tap analysis for the round bridge guitar before and after the bridge was glued on. Two peaks below the main (coupled) air resonance, 37Hz and 59Hz, show up better after the bridge is on. I should mention that sometimes a mode appears in a dip on a plot. It depends on where the top is struck as to whether a particular mode is activated. I should also mention that sometimes these modes appear differently before and after the back is glued on as well.

[Dominic]

Chuck, this chart is quite curious given my understanding of the techniques so far. I am assuming the back is on at this stage and the only change in the chart is the bridge? The T(1,1)1 mode has moved a tiny bit and T(1,1)3 has not moved at all. Assuming the second peak is the top main resonance and if the back is pitched higher the top it should be, adding the bridge has not had a mass effect and lowered T(1,1)2 but it has not added stiffness and raised T(1,1)2 either. When i saw the picture of the bridge I thought it would not be wide enough to increase cross grain stiffness so I thought it would lower T(1,1)2. But something else entirely has appeared.
Interesting, perhaps you could describe the guitar a bit more so we can work out what is happening.
Cheers
Dom

[DarwinStrings]

That round bridge would go well with some Celtic decoration.

Jim

Someone please turn up the Tele!

[Trevor Gore]

I've had a real problem with my browser these last couple of days; sites being blocked (the ABC??), redirection to other sites etc. all after installing the NCH tone generator to have a look at it. I installed the full package. I just removed the whole package and deleted all registry entries. Problem fixed.

Be warned!

Trevor Gore.