One of these two was the Tension on a string for the desired pitch. There seemed to be two views on the equation given; one was that it was valid for the Scale Length, the other that it somehow included the Total String Length if it was correct. A measurement method was given in the opening post to prove or disprove the two viewpoints. We will leave that subject, and move on to the second subject presented there…That which is closely related to tone, attack, sustain, and pitch as a function of time after string excitation. To save the trouble of looking in the previous thread, here is the opening statement.

“Now about the tone/sustain/harmonic/bell/etc. thing...these, also can be measured. The units will be harmonic content vs. time, in one form or another. The equipment needed is available in many computer soft-wares for recording, but usually not with the appropriate controls and data processing capabilities.
With the appropriate computer program, the instrument harmonics (as provided from the pickup) can be captured and seen/analyzed. This method provides a way to see the amount of harmonics in a single string, to all strings at once, and as a function of their tension, where and how they are excited (picked/hammered), where the pickup is located, etc.. The basic program costs less than your volume pedal (pot type). The program provides a Frequency Spectrum Analyzer, and Oscilloscope in one package. It is used for acoustic design from speaker cabinets to room size/shape/material compensation.”

This prompted the following response from a Forum member, Joe Meditz.

“Ed's graphs are time frequency plots. The frequency response at a given time can be adjusted with an equalizer. However, since the shape of the freq response curve does not remain constant, the equalizer would have to change from the time of attack and throughout decay. This could be done on a digital signal processor. http://s75.photobucket.com/albums/i287/edpackard/?
The question for Ed is: Using your measurement data, do you think that one could make a box that you plug your steel into that will make it sound like JD's Emmons? Now that would be something!”
I had the audacity to answer yes, with qualifications.
Then, in another thread Eric West asked:

“Ed, I have a request when you have time.
Inre to two questions:
One. Is there a way you can represent a graph and timeline of the change in pitch on a plucked string over a say one second interval in hz and seconds? (Certainly sharp immediately because of the initial disturbance.)
Two. Vindictivity (sp).
When a string that has been at rest for a given time, say 12 hours is plucked hard, does it draw up for a certain period of time, in a very small incriment of a hz?
These are not trick questions, I just wondered if you have the equipment, and/or the time to put this into documented form.
Thanks for any answer”

It turned out that while I did not have an off the shelf solution, Joe Meditz came up with one, and sent Eric and I a copy of the results, and sent me an explanation of how he did it.
This all leads back to the session at Jim Palenscar’s steel shop last Dec. where we measured, photoed , and captured the sonic behavior of 32 PSGs. When we first talked about doing this, WAV recordings were discussed for the sonic properties. This was bypassed in the interest of time. After boiling down all the info gathered at the session, we are ready for the next session to go deeper into the differences et al.
The first session dealt with the signals and frequency spectrum for all open strings (from the pickup) excited by a thumb pick sweep at the 12th fret. The next session will deal with the output from each string under a variety of conditions, such as where and how picked, plus some bar location tests. The plan is to WAV record this data as well as to capture the FSA data as before.
We will try to get Joe Meditz to participate in the session so that we can tune the data gathering to his post processing methods. It is convenient that Joe lives in San Diego.

[This message was edited by ed packard on 07 July 2006 at 07:24 AM.]

I posted:
"The 25" scale would have greater tension along its scale length but less total tension than the 30-1/4" string (24" scale).
The key factor is tension per inch.
The 24" string would have the same tension per inch in its scale length and its overhang. The two segments must have the same lbs/inch in order to have stasis."

Using lbs/inch instead of psi (assuming a given string guage), my thought was that adding the overhang tension to the tension of the scale length would produce more total tension in the longer string (breakage not being a factor in my thinking).

There was disagreement with part of my statement; and the topic being closed before I could verify my thought, I'd appreciate you're giving me a reality check.

You don't "add" the tension.

If the overhang were a MILE LONG, the string would have the same tension ANYWHERE ALONG IT in order to get the scale length up to the desired tension in the scale area to the desired pitch. A MILE LONG.

The ONLY difference is for each inch of overhang there is the exact same PROBABLILTY of breakage as every inch of the scale length.

So, if the overhang was just as long as the string, the string would only be twice as LIKELY to break in the course of say 6 months. Slightly less, since no scale area strings are not subject to fingerpick, changer use or bar wear.

Scale length DOES relate DIRECTLY to string tension. Total String length in an "overhang" DOES NOT.

Why this has been such a gruelling process of trying to get the simplest physical states properties across is WAY beyond me.

Or not..

EJL

Remember, the company that put out whatever information to the contrary inre to lessened string tension with less "overhang" WENT BROKE and until the company got sold at auction, they couldn't even post straight information on what was going on.

I was there at the auction, and passed on buying it.

I'm glad that it's in GREAT hands NOW.

[This message was edited by Eric West on 07 July 2006 at 10:30 AM.]

Take a given string, anchor it at one end, and apply tension to the other. The string stretches according to the tension applied. Each inch of the string stretches by the same amount. The anchored end has zero motion…the middle of the string shows half of the total stretch, and the end that the tension is applied to shows the total amount of stretch caused by the applied tension. The string will vibrate at a given frequency (fundamental = f1). We have a length of string, an amount of tension (pounds pull), an amount of stretch (giving a new string length and slight reduction in string diameter), and a resulting frequency.

If the pounds pull is increased, the frequency go up, the string gets longer (a bit), and the diameter decreases (a bit).

If the pounds pull is decreased, the frequency goes down, the string gets shorter (a bit), and the diameter increases (a bit).

Now lets apply a tension (pounds pull, let’s use 30 pounds) to a given length of string (let’s use 24”), to get a frequency of vibration (let’s say 440 Hz), and an amount of stretch (let’s say 0.5”). Now put a bar in the center of the string. The frequency of each half (12” either side of the bar) will be the same, and be twice ( = 880 Hz) that of the whole string (which was 440 Hz). The only thing that has changed is the length of the two vibrating parts of the string; the total string stretch, the pounds pull et al are unchanged.

Now let’s grab the string in the center, and cut it at that point. To get the string back to the frequency we had with the bar at the center (880 Hz), we will need an applied tension…it turns out to be the original 30 pounds of pull. We have half the string length, the same pounds pull, half the total amount of string stretch (= equal string stretch per inch of string), giving twice the vibrating frequency. So we can say pounds pull per given string (not inch of string), inches of stretch per inch of string, and the string length are what give us the pitch (vibrating frequency = Hz) of the tensioned string.

Summation = The tension is the same AT any point in the string, the amount of stretch is the same FOR any inch of string, and the frequency change is inverse to the length of the string.

Now let’s go back to the 24” string, and make it 30” long, and apply the same 30 pounds pull….the vibrating frequency goes down, and the amount of stretch increases (both total and per inch).

Place a bar at the 24” point. The long part of the string will now vibrate at 440 Hz. For the same applied pounds pull, and amount of stretch per unit length (inches per inch) of string. We can refer to the 30” as the Total String Length (TSL), and the 24” as the Scale Length (SL). Surprisingly, we might treat the 6 “overhang (30”-24”) as a scale length also. Pick it and it will vibrate at 4 times the frequency of the 24” section….but it has a 24” overhang.

The resulting conclusion is that for a given string, Scale Length determines the required Tension (pounds pull, which will be the same AT any point in the Total String Length, and Scale Length), and the amount of stretch, which will be the same FOR each inch of Scale Length…and this is what the basic string tension/length vibration equation says. Total String Length is only part of the happening when TSL = SL, otherwise the difference must be considered as an SL all by itself.

Long and detailed…hope that I said it correctly. The subject is worth some effort because it is one that is confusing. Your beliefs It won’t change your picking style.

OK Charlie, with that as background.
Quoting you:

"The 25" scale would have greater tension along its scale length but less total tension than the 30-1/4" string (24" scale).”

From the description above, we tune according to scale length, not according to Total String Length.

I would reword it to say---a 25” scale, with a given string, tuned to a given pitch, would need more pounds tension applied than a 24” scale would. They would both end up with the same amount of stretch per inch of string. The longer SL requires the increased tension to get to pitch.

“The key factor is tension per inch.”

If you mean …a key factor…per inch of scale, I can agree. The only thing that remains constant in both cases, and for TSL and SL, is the amount of stretch per inch of string, and PSI….did I say that rightly?

“The 24" string would have the same tension per inch in its scale length and its overhang.”

If you mean that if you cut the overhang off (or shortened or lengthened) it, you would still apply the same pounds pull (tension) to the end of the string to get to the same pitch…then yes.

“The two segments must have the same lbs/inch in order to have stasis.”

Stasis = 1. A state of balance, equilibrium, or stagnation.

I think that the Equality here is in the pounds per square inch of string, which produces a given amount of stretch per linear inch of string. If the string was a rod/board pivoting on a fulcrum, then the pounds times the length on both sides of the fulcrum would be equal…not sure how to visualize that in the present context.

Yes, it must be a question of coming to terms; I thought it was a simple matter until the explanations and disagreements got longer.

I think we're on the same page now(that scale length is the determining thing about the previous discussion, and overhang irrelevent); I was just unsure for a moment.

Charlie,
I think I understand what you are getting at.

You appear to be integrating tension per inch to get total tension. You can't do that.

The tension in pounds over the total length of the string is the same as the tension in pounds over a fraction of an inch of string. For example:

Suppose you stand on your bathroom scale and read 180 lbs. Now, suppose you have a second bathroom scale and stack one scale on top of the other and stand on the stack. Both scales will read 180 lbs!
The difference is that the total deflection of the sum of scales will be twice that of a single scale.

Hookes law says that:

F = kx

The force (compression or tension) is equal to k, the spring constant in lbs/in, times the distance the spring is compressed or stretched.

If the bathroom scale moved 1" while standing on it, its k = 180/1. This is a constant whether the scales are used indivually or in a stack.

But, the spring constant of the entire two scale system is now exactly half that or k = 90/1.

This means that when stacked together, 180 lbs compresses the system by 2".

You can do this experiment: Hang a cup on a rubber band and measure the elongation. Then tie two rubber bands together and do it again. The elongation will double. Also, if you put your finger on the knot the upper rubber band of the two, it will vibrate at the same rate as the single rubber band.

In summary, keeping the scale length constant, increasing string overhang only decreases the k of the system from the tuning peg to the bridge. But the tension does not change.

HTH,
Joe

[This message was edited by Joseph Meditz on 07 July 2006 at 03:32 PM.]

And you guys are right. TSL does not contribute to premature string breakage.

Sorry guys.

[This message was edited by Curt Langston on 10 July 2006 at 09:46 AM.]

quote:

---

That is what happened to Buddy Emmons and the 25 inch scale Sho-bud. That extra 1 inch length (TSL)..

-CL-

---

NO!

TSL= TOTAL STRING LENGTH.

NOT SCALE LENGTH.

SCALE LENGTH DICTATES HOW MUCH TENSION THERE IS ON A PSG. NOT TOTAL STRING LENGTH.

JEEZ OH WEEZ!

If this whole thing is somebody mistaking "Scale Length" for "TOTAL String Length", (TSL) then it points to a group imperative to refuse to debate until this basic misunderstanding is cleared up.

Shame on us.

EJL

[This message was edited by Eric West on 07 July 2006 at 06:16 PM.]

I am sorry for this misunderstanding, but I quoted Charlie quoting himself and it looked like a quote from me! I have since edited it.

Charlie is mistaken and is trying to understand why. The tension does not increase from increased overhang.

Below are two ASCII drawings of two systems of equal scale length. The string hangs over a roller o. The x anchors it. The weight provides the tension. The first system has more overhang than the second. Neglecting friction it is axiomatic that the tension in the string is the same in both cases since the weight is the same.

 o----------x
 |
 |
 |
 |
 |
 |
 |
 |
 |
 _
/ \
---
	 
 o----------x
 |
 |
 _
/ \
---

[This message was edited by b0b on 07 July 2006 at 05:17 PM.]

These are just the top side photos. Photos exist for the undersides also, but a step at a time.

From these photos, one may see the progression of construction and mechanisms over the years.

What you are seeing is the "tone" of the open string. If any of the peaks change in height, the tone will sound differently.

This post is to see how the graph shows up, and to begin to make a common language for the subject.

We will refer to the fundamental as h1 = 1 X the fundamental frequency. h2 will be 2 X the fundamental frequency, or 2 X h1. h3 = 3 X h1. and so on.

Using h1 as a root note = I, h2 is an octave higher. h4 is an octave higher yet, and h8, h16, h32 are the same note name at higher octaves....double the frequency, go up an octave in pitch.

h3 is 3 X h1, therefore not an octave, or a note of the same name. It is a V note. If h1 were C, then h3 would be G. h3 X 2 = h6 which is also a V note (G in the example)...h12 = 4 X h3 and is also a V note, and so on.

h5 = 5 X h1, and turns out to be a III note, therefore h10, and h20 are also III notes. If h1 is a C, and h3 is a G, then h3 is an E. We now have the basic three notes of our C triad chord, and all present in the single open string at the same time.

If you look at the chart, you will see the the amount of output from each string varies, and then they all fall off in amplitude as the frequency gets higher. It is at the higher frequencies that the loading of the volume pedal, effects, or amplifier begins to have an effect upon your tone.

Remember that this signal was taken after the pickup had done its job of modifying the raw string vibrations. The pickup did this in two ways. First via it's location re string length. Second via it's electrical network...it is an electrical filter and affects different frequencies in different ways. More on the effect of pickups on the signal later.

[This message was edited by ed packard on 09 July 2006 at 07:07 AM.]

The one octave diatonic scale (it's odd numbered intervals) will give you the basic major chord in it's triad, and seventh (3&4 tones) chords. This and much more is found in the harmonics of the single open string.

Now let's take the single string harmonic content display above and smooth it by three each 1/3 octave filterings (smoothing technique). The reason is to reduce the raggedness of the presentation, which will make where we are going easier to understand. You will notice that the low frequencies are not as smooth as the high, and that the amplitude appears reduced at the low freq' end. Not to fret as this will be remedied when the lower freq' strings are added.

Next we will show the results of strumming the all the open strings at fret 12.


You will see many more peaks in the unsmoothed response more strings vibrating means more frequencies available.

The smoothed trace amplitude now appears to be less frequency dependent.

These have been open string strums at fret 12. The high frequency falloff shows all that the open vibrating strings have to offer as seen from the pickup as the result of the "strum".

In the next post, we will show the results of the OS strum. The result will be all that the strings/mechanism can supply, as seen thru some of the various pickup settings available on the BEAST. The results are smoothed.

[This message was edited by ed packard on 09 July 2006 at 08:52 AM.]

Here is the open string strum at fret 12, on the BEAST, and viewed at the pickup out put. The different curves are for the some of the different pickup combinations on the BEAST. The point is....same hardware, same strings, same excitation, different spectral response because of the pickup used for each trace.

Now we will take one of the pickup settings, and look at the spectrum that results as a function of time. We will use 0,2,4 and 8 seconds after the open strings have been strummed. This will give what we hear as sustain. You will notice that the mid/low frequency output changes less than the high frequency out put vs time.

Point 1....the mechanism/construction of the instrument has an effect upon the "tone" = harmonic content, at any time. We will see this when we get into comparing the graphics of response for different instruments.

Point 2....the pickup will change the sound "tone" of an instrument, depending upon the pickups magnetic and electrical characteristics, and its location re the string (height, length).

I have used the BEAST as the source of the information given so far because it quite different in construction, pickup capability, and mechanism as compared to the common PSG. The questions that will arise where we are going will have to do with the effect of construction, mechanism, pickups, et al on "Tone" at time, and tone as a function of time. Here we are illustrating the extreme of differences from the standard.

DD...I think that I have found a compromise re graphic size, but I cannot restore the ones that I have already posted. I will try the other sizing in the next post.

[This message was edited by ed packard on 09 July 2006 at 09:37 AM.]

The "strum" was once across the open strings excited at fret 12. The "scrub" is all strings, all frets, all excitable locations, and capture the highest (peak) output for each frequency and each position/location event. In this case, they are shown as seen thru an assortment of the available pickup combinations. Notice that we did not use "smoothing" on the "scrub" outputs...it is not needed as the frequencies in the signals are very close together...indeed the same frequencies are found on each string at different frets, as are the harmonics of each string and each fret when considered as fundamentals.

This last shows just about everything re tone and output that is available using the pickups and hardware on this extreme instrument.

Here are the "hardware" photos of the BEAST, the source of the signals for the preceding graphs. First, the integrated changer/tuner structure. On the BEAST, it is located on the players left.

Second, the nut and pickups...on the BEAST, they are on the players right. The scale length is 29.730", and the Total Scale Length (TSL) is essentially 31.25 or so inches.

These are two pickups, each with taps, and reversible summable windings for a very wide range of tone variations.

[This message was edited by ed packard on 10 July 2006 at 09:59 AM.]

What has been shown are the frequency and amplitude results of:

1. Harmonics in a single open string, excited at the 12th fret...the source of scale and chord structure.

2. Harmonics from the single string smoothed to make it easier to see/read when combined with other strings outputs.

3. Spectral output of all open strings strummed when excited at the 12th fret, and this output smoothed. Look at the unsmoothed output and you will see "bumps" where the individual strings are. The tuning shows up in these bumps.

4. The effect of pickup type and location on the strum signals.

5. Harmonic variation vs time for one of these pickup settings, and the open string strum excited at the 12th fret.

6. The spectrum available, as seen from various alternative pickup settings, for the open string strum...to be compared with...

7. The spectrum available, as seen from various available pickup settings, for the "scrub".

8. The hardware configuration that produced the graphs of the frequency vs amplitude and time.

Notice the difference of the shape of the curves between points 6 & 7 above...can you explain it? Why do you think that this would be?

There are a couple of twists to the above approaches that will emerge as we proceed. I hope Joe Meditz will be so kind as to keep me straight, and add some of the work that he is doing to expand the subject.

If you think that the TSL/SL/stretch/tension/pitch/breakage was tough to talk about...wait till this stuff starts.

Next we will take an E9 10 string, and compare the available response. Might as well make it the FESSY, as that is what Joe M has at hand.

A few questions:

Was the intial attack also fed into the TrueRTA for spectral analysis for the single string, the "strum" and the "scrub?"

I imagine that the "scrub" was done by turning on the TrueRTA and setting it to peak detect and then doing an OS "strum" at the 12th fret, then bar at 1st fret and strum at 13th, etc. Is that correct?
What was the highest fret you put the bar at?

You said you recorded the scrub position event info. Does that mean you saved 12 files of data if you scrubbed through an octave?

Joe

The better way to capture the info would probably be to record the various responses as in recording the music, and THEN capture and analyze the resulting details in whatever data diddling way(s) makes sense. This would be session 2 at Jim's.

Re "scrub"...The peak capture was turned on and the string was randomly strummed while the bar was continuously moved (slowly) up and down the neck until no increase in amplitude was seen. No attention was paid to being ON any fret...in betweens all over the place.

The highest point on the neck was at the pickup. What was recorded was the single file with the max output seen by scrubbing.

Scrub was not applied to the 32 PSGs at Jim's...just to the BEAST to see the difference as compared to strum. The question posed was...if different, how different, and what might be learned by doing it to all.

[This message was edited by ed packard on 10 July 2006 at 02:11 PM.]

The magnetic pickup is composed of a physical structure that houses a magnetic circuit, and an electrical circuit. The terms and symbols for the electrical circuit are:

R = Resistance…in Ohms of some magnitude.
L = Inductance…in Henrys of some magnitude.
C = Capacitance…in Farads of some magnitude.

The R is the Resistance of the wire (winding) of the pickup. It is commonly misnamed Impedance. The R varies with the length of the wire in the winding, the size (diameter) the wire, and the material of the wire. Equal lengths, or turns of wire may have different Resistances, depending upon the size (diameter) of the wire…therein is ONE of the dangers of using pickup resistance as a measure of what the pickup will do sound wise.

The L is the Inductance of the winding. It is the link to the magnetic circuit. It’s value will depend upon the electrical winding, and the Windings coupling to the magnetic circuit. The larger the Inductance of the pickup, the less the high frequency signals allowed thru the pickup.

The C is the Capacitance of the winding, and is a function of the wire size, the insulation thickness on the wire, and the type of winding (called layer, random, scramble, etc). All else equal, thin wire gives more winding Capacitance than thick wire…Layer winding gives more Capacitance than the more random types. The larger the Capacitance in the pickup, the less the high frequency signals allowed thru the pickup. If it looks like a conspiracy between the L & C to kill the high frequencies, it is.

When we combine the R, L and C we get an electrical “network” that has parameters that may be used to describe the networks “filter” effect upon the vibrating string(s) frequencies sensed by the magnetic circuit. These parameters and their symbols are:

X = Reactance = like an AC (alternating current) resistance. The lower the Reactance, the less the effect upon the frequency. The Capacitance has a Reactance
(Xc), and the Inductance has a Reactance (XL).

Given values of Xc and XL are different at different frequencies, hence together they form a frequency selective filter. XL increases as frequency increases. Xc decreases as frequency increases. Can you see a “band pass” filter (high in the middle, and low at both ends)? For those that like equations, Xc = 1/(2 PI Hz C) & XL = (2 PI Hz L). The frequency center of the band pass filter is where XL = Xc; that is where the greatest amount of the string(s) vibration frequency as sensed by the pickups magnetic circuit will be passed thru the pickup.

The band pass filter has a property/parameter called Q. The larger the Q, the more selective the filter is re frequency; the lower the Q, the less selective the filter is re frequency. How may we control the Q of the filter, hence the sound from the string(s) vibrations sensed by the magnetic circuit? By the resistance of the wire used in the winding. Smaller wire means greater winding resistance (for a given length and type of winding), hence a lower Q filter and a lower frequency selectivity (flatter frequency response); Larger diameter wire, allows a higher Q filter.

The pickup has a property/parameter called Impedance (there it is!!!!) symbolized by Z. The Z of the filter is different for each frequency….how different? Depends upon the R,L,C,Q,X, and any other vegetables in the soup, the L of which depends upon the magnetic circuit.

What in H (coercivity here) is a magnetic circuit? Most picker folk will think of a pickup as simply some cylindrical magnets surrounded with some turns of wire. Maybe they call the magnets Alnico’s. The magnet(s) may be cylindrical, may be any of several types of Alnico (Aluminum, Nickel, Cobalt) material BUT they may also be one or more long bar magnets, may be Rare Earth, of Barium Ferrite, or other magnetic materials.

Each of these materials has it’s own properties…you don’t want to have the gory details presented here, so we will just say that the type of material, shape of material, and geometry of orientation of the material in the pickup, combined with any possible “return path” (from one magnet pole to the other) creates what we are calling a “magnetic circuit”. It is quite like an electrical circuit in the there is a force, something that flows , and something that something that limits the flow (think garden hose).

The electrical is Volts = Electro motive force, flow = current, Resistance = Ohms.

The magnetic is MMF = Magneto motive force, Flux (Latin for flow), and Reluctance.

The magnetic circuit is positioned such that the vibrating (and magnetically conductive) string(s) is/are in the magnetic field. When the strings vibrate, they disturb the magnetic field causing the flux to change in the magnetic circuit. The winding is in the magnetic circuit, and senses this change in flux, filters it according to it’s R,L,C,Q etc. and allows the “in the band” frequencies thru. What vibrations get into the magnetic circuit, and what determines the output amplitudes from the pickup?

The L of the pickup is a function of the flux coupling between the magnet and the coil. All else being equal, stronger magnets give larger L and associated frequency selective properties. A strong magnetic field round and about our coil may not be what we want.

The string(s) vibrations that get into the magnetic circuit are a function of the shape of the magnetic field where it is disturbed by the string(s) vibrations. If the pickup is located close to the bridge, the strings
vibrations that are found there will enter the magnetic field. Pickup location along the string is a determinant of the harmonic content (ratios) in our final signal. So is the focus of the magnetic field at the string…if a large length of string vibrates within the magnetic field, the sound will be different than if a very short segment of the vibrating string is sensed (think return screws and dual magnets etc.).

Here is a chart showing the vibration paths of h1,h2,h3,h4,h5 of a string, as a function of neck length. The amplitude of all h# vibrations has been set equal. You may notice that h#s that are ODD have a loop maximum at the 12th fret = ½ the neck length, and the h#s that are EVEN have a null at the same place…pick at the loop, get a maximum signal out for that h…pick a null, get a minimum signal out for that h.

Envision the frequency response for the h#s as a function of where you place the pickup re the neck length.

These h1 thru h5 vibrations are, in order R,oct,5,ooct,3, or if the string were tuned to C….C,C,G,C,E.

The larger the amount of disturbance of the magnetic field caused by the vibrating string(s), the greater the output of our pickup will be…all else being equal. Notice that it is the “change” in flux that is sensed, not the amount of flux. Magnet strength takes a back seat to magnetic circuit design when it comes to getting the desired signals to and thru the pickup. A shaped magnetic field at the string interface trumps a strong field.

To make the L low (more highs thru), the winding may be removed from around the magnets, and placed at another location in the magnetic circuit, where the ratio of field strength to changing flux is more suited to getting our desired harmonic content thru the pickup. If we reduce the L by repositioning the winding , we can increase the number of turns, and get more output.

The output voltage is commonly expressed as

Eo = k N (d phi/dt) where:

Eo is the output voltage

K is a constant having to do with pickup structure, efficiency, et al.

N is the number of turns.

d means “change in”

Phi is the magnetic flux.

T is time.

Now we will consider how to get the signal out of the pickup. We have seen that the electrical network of the pickup has an “impedance” = Z that varies with frequency. If we place a load (like a volume pedal) across the pickup, we may well reduce the high frequency transfer because the pickups output Z is much higher at the high frequencies than at the lows.

Now place/attach an amplifier (or effect, or ?) across the volume pedal. When the pedal is full on this new Z becomes part of the load that the pickup sees…you have lost more highs, and the tone varies according to how low the amplifiers Zin is, and how far you have pressed the volume pedal.

To make the analysis of the 32 PSGs constant, we will load the pickup(s) with a 500KOhm load…like a common pot volume pedal presents with nothing else attached.
Those readers that come from the age of tubes and output transformers may remember the the unloaded output transformer had a bit of an increase in the output of the highs. This was the “undamped” case. The degree of damping (loading) determined the shape of the high frequency response in the vicinity of the falloff slope.

The pickup is very much like the secondary of the output transformer in the above analogy….the same bump occurs if the output load is left off.

We will use the pickup(s) that are on the PSGs to see if we can quantify what sound it is that the various types of pickers want/like. There will be several categories.
Then we find it a good thing to do to use one positionable pickup to see if we can identify the contribution of the various constructions et al to the sound differences of the instruments.

Bottom line = we have magnetic filters, electrical filters, spatial filters (where the pickup is mounted re the string), material/construction/geometric/coupling filters (both PSG structure, and the magnetic structure of the pickup. Pickups also have “micro phonics”, so how the pickup winding is damped (wax or?), and how it is connected to the instrument is another “filter” source.

And then there is hum….never mind…enough already!

Why all this about pickups? Because, on the same instrument, the chosen pickup makes a big difference.

In our 32 PSG analysis, we just took the instrument and pickup as they were at the shop. This means that if we see a big difference between EMMONS PP and another EMMONS PP of the same generation, we don’t know how much is due to instrument, how much to pickup structure, how much to pickup mounting…but it is a start. The obvious details of the pickup structures, mounting, and resistance were taken. The changer end photos allow you to see the pickup and its location re the string end.

The photos of both the changer, and nut/tuner ends allow you to see the physical differences between the PSG string terminating/body connection approaches…THINK FREQUENCY FILTERS…THINK TONE!

If you see something that looks stupid here, it probably is...let me know.

Thanks for the XL,Xc correction JM!
Again, as these graphics are added, they are all available in one spot = the first PHOTOBUCKET link. Someone seems to be reading this stuff as there are roughly 1000 hits per week at PHOTOBUCKET.

[This message was edited by ed packard on 12 July 2006 at 08:34 AM.]

This post will be periodically updated (edited) as we continue. check back to it from time to time.

Some of the comparison methods and presentations have been presented above, in the previous posts in this thread. We start with photographs of the changer end, and the tuner end of the instruments. Then we proceed to the FSA graphs/charts.

Our first comparison, will be a FESSY, owned by Jim West (thanks Jim), and an EMMONS PP. Here are the photos.

The FESSENDEN #2 ON OUR 32 PSG LIST...CHANGER END.

The FESSENDEN #2 ON OUR 32 PSG LIST...KEYHEAD END.

The EMMONS PP #3 ON OUR 32 PSG LIST...CHANGER END

The EMMONS PP #3 ON OUR 32 PSG LIST...KEYHEAD END

The first very noticable difference is the appearance of the pickups. The FESSY has two rows of magnets, and the EMMONS PP has one. You might also notice that the diameter of the magnets is different.

There appears to be a difference in the diameter of the roller nuts...maybe even groove shape...TBD.

Jim Palenscar made detailed measurments on the PSGs used in this analysis. We will present a listing of those measurements, and more, for all of the instruments before starting with the FSA graphs/charts. The list will be in several pieces, by category of PSG physical area, such as cabinet, changer, pickups, etc.

Here is the cabinet info list:

Here is the rear apron and neck info list:

Here is the changer and nut info list:

The lists will be continued in the next post because the maximum images per post seems to be 8.

[This message was edited by ed packard on 12 July 2006 at 05:27 PM.]

And to go with the long winded (but far from complete) blurb on pickups, here is the pickup info for the 32 PSGs. This list is not "fleshed out" with much of the info needed, but you can see some of that info in the "changer photos" for any given instrument when we cover it.

[IMG]

There are a large number of frequency spectra vs. time presentations in the data bank for each instrument. The one chosen to give the most feel most quickly is the open string strum at fret 12, captured at 0,2,4,8 seconds after excitation. The results are double smoothed at 1/3 octave to simplify the viewing.

The top trace is the 0 second capture. It is done with a different algorithm than the other three traces. The object of trace 1 is to capture the peak of the frequency output at the time of excitation. To push the "Bell" analogy that Paul F brought up to describe a PSGs sound, trace #1 would be the clapper hitting the bell, and the almost instantaneous sounds emitted.

Trace 2 shows the frequency content remaining after 2 seconds.

Trace 3 shows the frequency content remaining after 4 seconds.

Trace 4, the lowest one, shows the frequency content remaining after 8 seconds.

You can see the bumps in the traces. Most of these are "tuning dependent". The nulls occur where there are no strings/notes in the open strings of the chosen tuning. For all instruments except the BEAST, the E9 neck is used. The BEAST is C9.

The frequency spectrum for the 10 string E9 starts later than for the U12s. If you look at the 60 HZ point on the E9 instruments, you can see the bump caused by the 60 Hz hum.

The first of this series is for the BEAST.

The second of this series is for Jim Wests FESSENDEN...1 year old.

The third of this series is a 1968 EMMONS PP bolt on.

Compare the three instruments for frequency content, frequency loss vs. time, and depth of the frequency dependent notches.

These first three are chosen to show the extremes of the PSG world. A 14 string C9 with a 30" neck keyless vs. a 12 string Uni with a 24" neck, vs. a EMMONS 1968 PUSH/PULL bolt on.

Look up the details of the instruments in the above lists and photos.

The pickups are quite different. The List gives the BEAST pickup location as the neck, and the maker as Danny Shields. The pickup being used for the graph is the bridge pickup, and it is a Di Marzio.

For the purpose of showing ALL that the pickup can pass, aload of several megohms is used as opposed to the more common 500K Ohms. The result is a bit more high frequency response.

[This message was edited by ed packard on 15 July 2006 at 11:01 AM.]

Apply your "pet" theories to the differences and see if it can explain them. Look particularly at the top traces, their distance above the second traces, and the 60 Hz (hum) amount. How about the frequency depoendent bumps differences between instruments?

Could the differences be due to body to hardware connecting screw torque values?

Perhaps on e might have a misadjusted/bad pickup, but then the rest are still different.

The hardware photos are not put up here...they, and the graphs are available at http://s75.photobucket.com/albums/i287/edpackard/?start=40 for those that might like to prowl around.

Looks like I botched the numbering, but you can figure it out until I clean it up. #30 & 31 S/B # 31 & 32.

[This message was edited by ed packard on 18 July 2006 at 10:44 AM.]

OK, so when I'm winding a new E9th 3rd string (an .011 for example) onto a 24" scale, to get to G# (gotta crank on it and hope it doesn't go flying...)I reach a given tension.
To get to the same G# note on a 25" scale (regardless of whats "behind the roller nut") the tension to get to G# will be a little higher, because of the longer span.
Because the "free span" is greater, won't the changer on a 25 inch scale, also have to impart just a little more tension on a raise to get to the target raised note than the changer on a 24 inch scale?
It would seem that a longer scale length makes the string work harder for the same results.

[This message was edited by Ray Minich on 18 July 2006 at 10:45 AM.]

Now that you have seen that all SHO BUDs are not equal, lets look at the EMMONS under the same test conditions. Thanks to JD for bringing his instruments in for the tests.

These will be the last of this type of image posted in this thread. From now on we will announce which instrument's graphics have been added to the PHOTOBUCKET link, and those that are interested can go there to compare all 32 instruments (as they are put up).

If I look at the fundamental, it looks like JD's Emmons has about 15 dB loss from time zero to 8 seconds, where the Shobuds are 20-25 dB loss for time zero to 8 seconds.

However the harmonics decay on the Shobuds in a more linear fashion with respect to time (the Emmons loses more upper harmonics compared to the fundamental in the first 2 seconds than many of the Shobuds).

At 8 seconds, both guitars have lost similar amounts of harmonic content, but the Emmons has lost about 5 dB less fundamental signal voltage, so you could say it has more “sustain”, but if having strong upper harmonic content at time 2 seconds is important to you then maybe you would prefer the Shobuds.

No 31 Sho-Bud's first graph shows a very bright freq resp. Its winding resistance is low being at 10.8k. OK. Low winding resistance ==> brighter sound. But then you look at No.32 which has a lower winding resistance of 10.1k and see that it is not as bright. Both pickups are Sho-Buds. One axe is a ProII the other a Professional. Are ProII's considered to be, in general, brighter?

It's certainly going to take some time before this data is digested!

Joe

I tend to think about breaking the spectrum into octaves starting with C1,C2,C3,C4,C5,C6 etc.

Having established such categories, we may talk of the fall off of octaves re the 0 sec levels in terms of db loss per second...or octave n being xdb lower than octave n+1 at t=y. This might be extended to -db per octave per second, or similar.

Joe...I tend to think of hum as being X db below the 2 second signal level (would prefer 1 sec signal level, but that was not taken last time.

The EMMONS numbers look to be about -40db for that categorization.

The Sho Buds look to be about - 30 to -40 db. #30 seems to be -30db, but #30 has weird curves to begin with.

Changing the wire size from 38 to 40, or to 36 can greatly affect the Ohms. Magnet Wire Sales (MWS) has this type of info available on line.

How the wires from the pickup are handled (shielded, twisted, twisted with ground, etc), plus where/how the ground is tied to the chassis can make or break the hum thing. Just touching the strings on some instruments makes a large difference. Our tests were "hands off" the strings before capturing the signal(s).

Lots of fun ahead.

#5 MSA had a HI Z load on the pickup (5 or so meg-ohms, hence the undamped high freq bump on the top trace. Mentally roll off the hi's for the other traces a bit.
http://s75.photobucket.com/albums/i287/edpackard/?start=40

[This message was edited by ed packard on 19 July 2006 at 03:03 PM.]

Re HUM: compare the hum levels between the aluminum necks/bodies and the non metallic necks/bodies...

[This message was edited by ed packard on 20 July 2006 at 11:56 AM.]

The photos are the best source for determining if a typo or transposition was made re either the spreadsheet, or the FSA labels. #27 & #28 are both EMMONS D10s, and both JDs. If you look closely at the keyhead photos for them you will see spaghetti tubing on the center strings overhang. JD is very tone fussy in his choice of instrument, and how it is decked out, right down to the strings. He was there for the session, and I thank him for bringing the instruments.

This data comes from a first run at gathering data on a variety of PSGs. Each PSG had 8 traces taken for the open string strum excited at fret 12 with a thumb pick.

4 of these traces were without and changes activated, and 4 were with P1P2 activated. The "tuning dependent" bumps are visible, and any change influence on tone/sustain etc..a non issue at this scale.

The open string strum fret12 was chosen as it had the least people effect/influence involved. Other strums will be captured, as in up the neck, etc. These will involve the bar, and the player more. "Pick" of the individual strings is in the cards. Pick of the same note on different strings will show the differences in harmonic content for the same notes in different locations.

"Scrub" is another approach that will be used. In the intro graphs where the effect of changing pickup settings on the BEAST were shown, both strum and scrub were captured. The os strum gave a top trace that was sort of flat. The scrub gave a top trace that was "arched". One might ask why. OS strum shows what the open strings give for amplitudes of their fundamentals and harmonics as passed thru the pickup. Scrub shows the highest output for all places on the neck and strummed at all places. This means that any and all fundamentals and their harmonics may be found in many places, and may have more amplitude as seen by the pickup than those limited ones of the open string strum. The nulls and loops chart for the neck shows that for an open string the most ODD harmonic signal would be seen by the pickup if it was positioned at fret 12...The even harmonics would be minimized at that point.

Now look at the magnitude of the signals as you mentally move the pickup toward the bridge.

The "arching" of the scrub signal response is showing the band pass filter characteristic of the pickup...the Q of these pickups is very low, so they are quite broad on top.

I am sure that there are typos and transpositions in the data. The real point is that the differences between instruments and associated hardware may be seen in the charts/graphs, even within the product of a given maker. If there were no differences to be seen, then the exercise would be as one chap in a long past thread predicted ..."pointless". So far I have not seen the need of a "picking machine".

This is just the first step in what may be a long journey. Joe Meditz has done some interesting things on the subject, and I assume that he will make these known when the time is right.

ADDED: The pickups on #27 & #28 are both "PITTMAN" pickups...so we not only have two EMMONS instruments, owned by the same tone particular picker, but the E9 pickups are by the same Mfg. I E mailed PITTMAN to get some tech data on the pickups, but they/he did not acknowledge the E mail.

All the instruments were FSA tested in the same place in Jim's shop, and facing the same way, so that any hum info would be comparative.

Re Paul F's "Bell" description of PSG tone from another thread...The top trace in the graphs would contain the sound of the clapper hitting the bell casing = the attack from picking the string(s). The 2 second trace would be like the amount of (freq & amplitude) of the casings vibration at the 2 second mark, etc.. The greater the drop from the first to the second trace the more damped the bell casing would be in the analogy.

The Sho Bud that you mentioned would have a very sharp (relatively speaking), and a large damping factor. It would sound "dark" after the initial excitation, but "bright" at the moment of attack (assuming no volume pedal activation to "sneak into the sound)...seems to be one of the preferred Nashville/Bakersfield sounds...and we have only addressed the open strings so far.

[This message was edited by ed packard on 21 July 2006 at 07:21 AM.]