Measuring resonant/resonance peak - Questions/advice

[David Schwab]

Bartolini used a mechanical string picker once. I think it was used to measure the output of pickups.

The problem with using a string as a frequency generator is it produces a lot of harmonics.

[salvarsan]

4) I found a USB Computer ociliscope (model DSO-2250) which looks like it is also a signal generator. Does anyone with experience want to look over the specs and see if this would be suitable to do the task?

If you want to characterize audio widgets,
you might look as high as 100kHz and require a 1.0 Msample DSO.
The DSO-2250 does that easily but Protek specs egregiously misrepresent
the product as a 100MHz bandwidth scope.


There is an oscilloscopic "rule of 10" where if you want useful phase
info at a frequency F you must sample no slower than F * 10.
In other words, the DSO-2250 is a 250 mega-sample device equivalent to a 25MHz scope.

By Protek's misrepresentation, my CGR101 is a 10MHz sampling scope instead of a 2MHz device.

There are no Protek DSO-2250 software screenshots that I can find.
What are they hiding?

If you can try out the software first and like it,
it could be a good bench tool.

-drh

[salvarsan]

Bartolini used a mechanical string picker once. I think it was used to measure the output of pickups.

The problem with using a string as a frequency generator is it produces a lot of harmonics.

The harmonic content is strongly influenced by the the pick and plucking technique.

Using fingers, pluck a string at half-way up the fretted length for minimal harmonics.
A felt pick is more reliably self-damping, too.


-drh

[bbsailor]

Examine The Patient

The harmonic content is strongly influenced by the the pick and plucking technique.

Using fingers, pluck a string at half-way up the fretted length for minimal harmonics.
A felt pick is more reliably self-damping, too.


-drh

Hello Daniel

I agree with your harmonics content analysis.

If you really want to test this out you can examine the patient without too much special equipment, here is something to try to see how many harmonics are in typical guitar strings. We are using the string as a reverse-pickup, to magnetically make the string vibrate at specific frequencies that are fundamentals and harmonics of the standard note string frequencies we are examining.

Attach a 10 Ohm resistor (1 watt) to amp speaker out + and the other end to Clip Lead #1.

Attach another clip lead, called Clip Lead #2, to the audio ground.

Attach Clip Lead #1 to the low guitar string (preferibly an acoustic) behind the nut and attach Clip Lead #2 behind the bridge on the same string.

Place a cermanic magnet (N or S pole facing string) under the attached string.

Attach a variable audio oscillator to the audio input of the amplifier. Keep output volume below about 1W output (Feel resistor for too much heat!) or a higher volume if you choose to listen to the higher harmonics.

Tune oscillator to low E or 41Hz, the sweep the variable oscillator to 82Hz and the string will start vibrating when the oscillator gets near 82Hz. Also, set it at twice the frequency or 164Hz, and the string will start vibrating again, and continue to do this at all of it's harmonics (both odd and even) at different amplitudes that you can hear if you choose to listen to them and adjust amp volume without burning up the resistor or getting the string too hot.

This is a way to observe and measure what harmonics the string is putting out. Just map the amplitude of the harmonics that you want to investigate (open string or selected frets). Then, compare the guitar-pickup-under-test response against the known harmonic levels on individual strings harmonic fingerprints.

Accurately make all inputs the same level and measure the acoustic guitar output if an electric-acoustic type to get a good fingerprint for that string or that note. Knowing the open string harmonics plus some harmonics of notes up to the 1K range should be enough. Basically, you want individual string samples every 3 to 7 frets depending on your energy.

Harmonic fingerprints vary as strings age. They get thinner as they are used and continuously being stretched to keep it in tune. They also collect dirt which changes some weak harmonics more than some stronger harmonics.

The String impedances vary from 1 to 2 ohms depending on wire thichness or core wire thickness, if wound. This test makes a very high Q, amplified resonant circuit with the string forming the mechnical part of a feedback loop.

I thought this might be a more intimate way to examine the patient.

Joseph Rogowski

[salvarsan]

Outstanding!

Using the string as a transducer would never have occurred to me.

Guitar strings also have a longitudinal vibration mode
whose base frequency is determined by fretted length
(and elasticity modulus) and not by tension.

It is not harmonically related to the plucked fundamental.
This sound is strongest at the pluck release time and is
often called "pick chirp".

Novax Guitars calls it the "clang tone", has spectral plots and discussion at:
http://www.novaxguitars.com/info/technical.html


-drh

[David Schwab]

Thats a cool idea.

The old Roland guitar synth had a sustain device that had a large magnet under the strings by the neck, and then they used nylon saddles on the Tun-a-matic bridge, and ran the audio signal from the pickups to the strings.

There's a few patents on using the strings as pickups by Clevinger, who makes electric upright basses.

[Peter Naglitsch]

Even thou Rogers idea is great I imagine you can emulate it much easier (altough maybe not with the same accuracy) with a E-bow. Activate a string with the E-bow (thus eliminating the possible variations in picking if you do not have the Batrolini picking machine) and measure the output from the pickup while fretting the string higher and higher thru octave after octave. That should at least give you an idea about what freaquency the resonance peak acours at. Then you can always fine tune the result by simply tuning the note with highest output to find the resonance frequency.

[Robert M. Martinelli]

The inductance of the driver coil causes the current and thus magnetic field to drop as the frequency is increased. A series resistor having at least five times the resistance of the coil's reactance at the highest test frequency should swamp the inductance effects and cause the driver coil effects to vanish.

The drive coil should not be wound as a pickup coil, which will have too many turns and thus too much inductance and too low a self-resonant frequency.

For example, consider a coil having an inductance of 0.01 H. At 10 KHz, the reactance will be 2*Pi*10000*0.01= 628 Ohms. Five times this is 3.14 K, so use 3 K or 3.3 K. Or 10 Kohms. Great precision is not required.

FWIW I use a salvaged low voltage relay coil as a driver, and the results I obtained are satisfactory ( IMHO ), as the tests I've done so far have matched almost exactly the "theoretical" ( modeled ) behavior, so I am pretty confident they reflect "real life situations" reasonably well.

Regards

Bob