syscomp resonant peak reading help.

[StarryNight]

he guys, I'm trying to understand my syscomp scope graph and the anomaly on one of the pickups. Have a look at the first graph showing some kind of phase flip - or is that just a scaling issue with the graph. The air coil I made might have something to do with it. Any initial thoughts would be helpful. Just starting out with this thing. cheers.

[John Kolbeck]

It's a scaling issue. The plot line should just disappear out the bottom and reappear at the top, but a bug in the software causes it to draw an ugly line from bottom to top that doesn't belong there. If you were to reverse the leads of the pickup, the phase would be 180 reversed, and it would look like the bottom plot. Whoever made those plots just didn't care enough to correct it.

[salvarsan]

Not scaling. It's a software defect.

Unless you have a capacitor in series with one of the coils, the phase plot should cross 0 at the frequency at which the amplitude plot peaks.

See the document Measuring Impedance and Frequency Response of Guitar Pickups PDF at SysCompDesign.
Table 5 on page 5 shows this and those plots have been my experience.


Time to make a bug report.

[Mike Sulzer]

Not scaling. It's a software defect.

Unless you have a capacitor in series with one of the coils, the phase plot should cross 0 at the frequency at which the amplitude plot peaks.

See the document Measuring Impedance and Frequency Response of Guitar Pickups PDF at SysCompDesign.
Table 5 on page 5 shows this and those plots have been my experience.


Time to make a bug report.

Possibly that bug report is a bit premature. How do you know what is being measured? Suppose it is gain, not impedance. Then the phase at resonance should be -90 degrees, and it should be near zero at low frequencies. That is the bottom plot. In the top plot it looks as though the polarity of the reference signal got reversed, changing the result by 180 degrees. Thus the resonance appears at +90 degrees and 0 degrees becomes 180. -180 and +180 are the same thing, and so you have a jump when you cross what should be zero.

[salvarsan]

Possibly that bug report is a bit premature. How do you know what is being measured? Suppose it is gain, not impedance. Then the phase at resonance should be -90 degrees, and it should be near zero at low frequencies. That is the bottom plot. In the top plot it looks as though the polarity of the reference signal got reversed, changing the result by 180 degrees. Thus the resonance appears at +90 degrees and 0 degrees becomes 180. -180 and +180 are the same thing, and so you have a jump when you cross what should be zero.

For a pickup, zero crossing is at resonance = max amplitude using the syscomp device and the test rig described in the PDF above.
I have reproducibly done this using the CGR-101.

Look at the phase plot (which is vertically symmetrical) and estimate where the zero crossing is. Flipped top for bottom, the zero crossing is still way off; lead polarity is not the problem.

Code defect? Series capacitor in the circuit? Marginal test rig setup? Bad moon phase? Other?
Whichever it is, the plot is wrong.

While the code is written in Tcl/Tk, the FFT bits were sufficiently clear, but it's possible that an error got in while generalizing it for the CGR-101, the mini, and the new CGR-201.

[Mike Sulzer]

For a pickup, zero crossing is at resonance = max amplitude using the syscomp device and the test rig described in the PDF above.
I have reproducibly done this using the CGR-101.

Look at the phase plot (which is vertically symmetrical) and estimate where the zero crossing is. Flipped top for bottom, the zero crossing is still way off; lead polarity is not the problem.

Code defect? Series capacitor in the circuit? Marginal test rig setup? Bad moon phase? Other?
Whichever it is, the plot is wrong.

While the code is written in Tcl/Tk, the FFT bits were sufficiently clear, but it's possible that an error got in while generalizing it for the CGR-101, the mini, and the new CGR-201.

When you induce a signal in the pickup coil, it appears in series with the coil and the output is effectively taken across the coil capacitance. This means that the output voltage lags the current (and thus the input voltage at resonance) by 90 degrees. Look at the three plots in the appropriate later section of the document you referred to. They show the 90 degree lag at resonance.

Thus your statement "For a pickup, zero crossing is at resonance..." is not true for this case, although it is for an impedance where you effectively inject a current into the parallel combination of the L and C. I agree that there is something funny with the phase at low frequencies, but it is consistent with the lower plot if we assume the phase reversal I described. This might be an issue with the driver coil (as could also be the case with the figures in the document at low frequencies).

Of course, we will have to wait to learn what was actually being measured. But in any case, I think it is very unlikely that a bug so basic has just been discovered in this well used software.

[salvarsan]

"For a pickup, zero crossing is at resonance..." is not true for this case

The scope of the statement is constrained to a practical exercise of the SysComp Design device and is not a theoretical issue.

Get yourself a CGR-201 or similar and run the test.
The PDF alone shows you what to expect on page 5.

[Mike Sulzer]

The scope of the statement is constrained to a practical exercise of the SysComp Design device and is not a theoretical issue.

I am sorry, but I do not understand what you are trying to say. But since you referred to page 5, I understand that you agree with me that this is probably a gain measurement, not an impedance measurement as you first said. Thus you can see how easy it is to get the phase backwards, and the effect on the plot is a logical consequence.

[Mike Sulzer]

I would like to remind anyone who mistakenly thinks that Salvarsan is rudely implying that I can deal only with theoretical issues, and that I am thus unqualified to discuss practical measurements, of this: I have posted in this forum my own design of a pickup measuring system using a recording interface. That includes the small amount of additional hardware necessary and the Python code that controls the system. So I have quite a bit of practical experience in this sort of thing, enough to understand the kinds of problems encountered and how they can be misunderstood.

[CheopisIV]

Scaling; we've discussed this a while ago and I had the same questions

Link: Stupid question about resonant frequency test.

[salvarsan]

... Salvarsan is rudely implying that I can deal only with theoretical issues.

That is incorrect.

I did not imply that but said it outright, having stated the narrow scope of my assertion beforehand.

What was implied was that you habitually behave in a wilfully contentious manner that exhausts good will as well as patience.

Yes, you built a test rig using a sound card and wrote the software. Kudos, sir, however that is an appeal to irrelevance and not the issue.

It is certain that you DIDN'T build the CGR-101, et.al. and didn't write its application software.

If you owned one and used it to test pickups, examined the source code, you would speak differently.
I do, I did, I have, so I do.

[Mike Sulzer]

That is incorrect.

I did not imply that but said it outright, having stated the narrow scope of my assertion beforehand.

What was implied was that you habitually behave in a wilfully contentious manner that exhausts good will as well as patience.

Yes, you built a test rig using a sound card and wrote the software. Kudos, sir, however that is an appeal to irrelevance and not the issue.

It is certain that you DIDN'T build the CGR-101, et.al. and didn't write its application software.

If you owned one and used it to test pickups, examined the source code, you would speak differently.
I do, I did, I have, so I do.

It is much better to discuss the technical stuff than to be insulting.

Figure 3 on page 4 of the Syscomp document shows the set up for gain measurement. Measurement channel A is connected across the drive coil, while B is connected across the pickup. Thus A serves as the reference for the phase and amplitude measurements. Starry night's lower plot shows the expected result. His upper plot has the negative of the expected phases. This could happen just from turning the drive coil over. I think it is plausible that this is what happened.

You have said that there is a bug, but you have said nothing about it that would show how it gives the reversed phases. Can you? Can you provide any plausible information that would show that you are correct?

[zparts]

he guys, I'm trying to understand my syscomp scope graph and the anomaly on one of the pickups. Have a look at the first graph showing some kind of phase flip - or is that just a scaling issue with the graph. The air coil I made might have something to do with it. Any initial thoughts would be helpful. Just starting out with this thing. cheers.

Hey,

Maybe try using a 470k resistor in series with the probes instead of using a driver coil. I find driver coils to be sensitive to placement and phase making it difficult to get consistent readings each time. Just a suggestion.

[rjb]

Hey,

Maybe try using a 470k resistor in series with the probes instead of using a driver coil. I find driver coils to be sensitive to placement and phase making it difficult to get consistent readings each time. Just a suggestion.

I don't understand the suggestion.
What is your test setup configuration?
Are you driving the pickup coil with a signal generator, through a series resistor?
What is the purpose of 470 resistor in series with probes?

[zparts]

I don't understand the suggestion.
What is your test setup configuration?
Are you driving the pickup coil with a signal generator, through a series resistor?
What is the purpose of 470 resistor in series with probes?

I actually use the Syscomp CGM-101 for my portable desktop testing. It works great. Just connect the Input A with the Output + to one side of the resistor and connect Input B with the pickup + to the other side. Set the probes on X10 and let her rip. I didn't like the driver coil because it was too finicky and inconsistent for me.