Discussion in 'Pickups & Electronics [BG]' started by colcifer, Jan 14, 2012.
I don't know if anyone does. You usually only see things like "bass-5, mids-6, treble-5."
It's not easy to get an accurate frequency response, because there are many variables involved. What you use to disturb the magnetic field to induce a current to flow in the coil, and what sort of loading is against the coil will influence the way the circuit behaves.
The only thing that would sound believable to me would be something like a controlled mechanical vibrator driving a mu-metal rod (no jokes, please!) or maybe a series of tuning forks that are also mu-metal (steel for example.)
This is not an easy thing to do and even if you did, trying to interpret the results vs your strings and setup is probably equally difficult.
When it's done, it's done by using a small coil that feeds a signal into the pickup from a frequency generator. Then the output of the pickup is plotted against the signal as it's swept.
You can also probably do it with white noise.
But hardly anyone does this.
Bartolini made a mechanical string picker once! But that was to measure output, and not frequency response.
It's more difficult than it looks to get reliable results.
Using an induction method does not take the response time of the magnetic field into account. A non-ferrous disc with steel "poles" that is spun in the field will give a descent indication of frequency response. The wheel must be well balanced (like a car tire) and safety precautions observed in case the wheel comes apart. The output is basically a sine wave and tells nothing about attack and decay characteristics. A mechanical picker is needed for that and the alignment of the tested pickups and age of the string in the setup is critical to obtain usable info. Ferrous materials cannot be used in the fixtures except for the bits that excite the magnetic field.
Here's the coil method.
BuildYourGuitar.com :: The Secrets of Electric Guitar Pickups
Where does the article address the frequency response of the exciting coil? If its not flat (or a correction made for it) I ain't buyin' this.
The lower its impedance, the flatter it is. Or, if it's immediately preamplified it probably gets to within ~1 dB (or even tighter) to flat.
...funny how noone ever made a Wal-like exciter coil. Or an Alumitone-like one. That'd give a pretty flat characteristic.
Wal? They use multiple coils with lots of turns of wire! I'd rather use a speaker coil.
A year later...
While I was digging around last weekend researching pickups and electronics, I found that Lemme article. Really cool way to help see what you hear in a pickup. Even if it's not perfect, you get a sense of where that resonance is. More useful than a DCR reading at least. I'm surprised I didn't find more related threads on TB. It sounds just like the thing that'd get people's attention here. (At least it would in the Amps forum. )
Here's another interesting link I found. Repurpose an old hard drive's coil to drive your pickup.
Wonder if it's better or worse than using a cheapy guitar pickup off ebay. (I'd suspect better.)
I hope to get a testing setup like this going when I get all my parts together for the next build. Should be fun to play with different circuits and scope 'em to see what's happening. Again, not AES level, just to see how the pickup's RLC interacts with the components.
Here at Nordstrand, we use a usb digital/analog converter, 6 ohm dual driver coil, and ARTA software. The software is easy to use and you really don't even need an outboard converter to get it to work. You could use headphone and microphone inputs to run it. The problem is, it's not realistic unless you have the pickup attached to the pots and capacitor you are going to use it with. All measurements are relative to each other, but not to real life situations. It's good to have objective information, but the true test is in the studio through every amp we have. The slight differences on the oscilloscope translate into big differences live.
The author of this paper must be an RF engineer like me, most other people would not put SMA connectors on an audio amp! Of course most of you won't understand the humor in that....
I don't know if they do it anymore but for a time the University of Illinois was measuring guitar pickups electrically. They were measuring the RLC values of a pickup and then you can use the component values in a circuit simulator to make frequency response plots that include as much of the rest of the instrument and amplifier as you might care to include. There is a spreadsheet with their measurement results on that web page but it is incomplete and not entirely consistent. You should never send a student or professor to do a bass player's job, right?!
I would argue that you don't need to include the rest of the bass if you are a pickup maker. What we want from you is an indication of the tonal characteristics of the pickup itself. If the industry were to join forces and devise a standard test for this that everyone could use and then publish measurements of their offerings it would help buyers make better selections. The common descriptive terms like warm, cold, woody, sterile, bright, dark, etc, etc, etc are just meaningless gibberish since what one person might call woody another would describe as dark and so on. Objective measurements would be helpful.
I tend to favor a mechanical apparatus of some kind to excite the pickups for testing because it better simulates reality. Rotating, off center mounted disks could simulate the vertical motion of strings and a disc mounted at an angle on its shaft could simulate side to side motion. Pickups would pick up the noise from an electric motor quite well so you would need to use a flexible shaft to a motor located as far away as possible to do the testing. Or use a pneumatic motor! It would be nice to have a standard setup so that the output level as well as the frequency response could be captured and reported. Vibrating strings would obviously be the perfect way to excite the pickups but has anyone ever devised a method to produce a consistent excitation level from a string in a testing apparatus?
There is one thing that I have never understood about guitar pickups. When you model them electrically you use a voltage source with a constant amplitude in the model. Yet magnetically coupled systems typically have a frequency response that increases with frequency since the dB/dt (and I am not going to explain what that is since only those who already understand it could comment on this) increases with frequency. So why don't electric guitars and basses need an equalization circuit to get a flat response? The old magnetic phono pickups had an equalization network to match the RIAA curve, the magnetic pickups and tone wheels in automotive ABS have a similar EQ circuit, yet guitars don't need them. I don't understand why they don't.
I can't help there, but the RIAA de-emphasis was applied to compensate for the recording pre-emphasis rather than the response of the cartridge itself.
Actually it was both, or rather both problems are addressed by the same solution. Certainly articles available on the subject on the web such as this one and this one focus on pre-emphasis to reduce the effect of the high frequency surface noise of the disk. So you probably want to emphasize the high frequency content anyway even if you use a transducer that is inherently flat. But that would walk you into the same problem at high frequencies that the RIAA standard eliminates at low frequencies. The physics of magnetic pickups dovetails perfectly with the RIAA curve, no surprise there clever engineers were involved with its selection.
And now I am going to have to discuss the issue I said I would not, dB/dt. The output voltage of a magnetic pickup is given by N*dB/dt where N is the number of turns in the pickup and dB/dt is mathematical shorthand for the rate at which the magnetic flux linked through the coil changes with time. If you know calculus you recognize that as the time derivative of the B field in the coil. In the magnetic wheel speed sensors used in automotive ABS there is a toothed wheel attached to the wheel axle so that it rotates past what is essentially a single string guitar pickup. When a tooth is aligned with the pole of the pickup the B field is at a maximum and when the gap between two teeth passes the pole the B field is at a minimum. The B field through the pickup coil oscillates in a nearly square wave fashion as the wheel rotates between Bmax and Bmin. The B field amplitude does not change with wheel speed but the frequency at which it changes increases directly with the wheel speed so the dB/dt is very low at 1 mph and very high at 100 mph. As a result the output of the wheel speed sensor will be under a Volt at 1 mph and upwards of 100 Volts at 100 mph! Pity the poor wheel speed sensor IC which runs at 12 Volts!! I've been there and done that professionally and as it turns out a good active de-emphasis filter protects the IC just fine and everyone is happy.
A magnetic phono cartridge system is a bit less constrained in that there is not a hard Bmin and Bmax. The difference between Bmin and Bmax increases as the permitted physical displacement of the needle increases. At some point the curve of B field versus distance will become non-linear and cause distortion but if you wanted to you could just make the cartridge larger to allow a larger physical movement. Practically speaking however you have to pick a reasonably sized pickup and that will limit the difference between Bmin and Bmax that your cartridge design will produce. But given the B field limitations you have other practical considerations to sort out. If you wanted to make a recording with a naturally flat electrical output with frequency then you still have a situation where for any given constant B field variation with frequency the dB/dt is going to increase with frequency and that will not give a flat electrical output. So you would have to limit the mechanical excursion of the high frequencies as compared to the low frequencies to get a flat output. If you keep the high frequency excursion relatively small you run into serious high frequency noise issues. If you let the high frequency excursion be large enough to combat noise then the low frequency excursions become huge and this causes two other problems: a) it limits the number of grooves per inch on the disk which reduces the playback time per side and b) now the phono cartridge has to be larger to accommodate the required low frequency range of motion without producing distortion. So the solution that was adopted was to keep the mechanical excursion constant with frequency. The excursion is set high enough so that the high frequency noise of the disk is well controlled and the low frequency excursion is kept the same. This limits the low frequency dB/dt and produces lower output voltage at low frequencies and that does give you issues with low frequency playback noise (wow, flutter, and rumble) but overall it was the best engineering compromise for the vinyl disk technology.
But we don't see this with guitar pickups and I have not seen a good discussion of why not. In fact the only discussion I have seen of the issue is between the various voices inside my own brain! At one level I understand what must be happening: for one or more reasons the B field variations at low frequencies must just naturally be higher than they are at high frequencies. How exactly does that happen? There is where I become unsure although there are some fairly obvious things going on that should have this effect. One is that string size decreases with frequency. The fat low strings will produce larger reluctance and therefore B field variations than the thin high frequency strings. So that kinda explains the string to string flatness. But what about the variation in frequency of a given string as you fret up and down the neck? It could be that the amplitude of vibration and therefore the B field amplitude naturally decreases as the vibrating length of the string decreases. I think that to an extent that is true although it is also true that the pickups may be at the peak point of the vibrational motion for high frequencies whereas unless your bass has pickups at the 12th fret they are not at the peak point for low frequencies. And this would tend to increase the dB/dt for at least some notes as you fret up the neck. A third fact is that while a wheel speed sensor deals with about a 100:1 (40dB) frequency range and a phono cartridge with a 1000:1 (60dB) range any given string on a bass produces a frequency range of 4:1 (12dB) or less in most cases. The Ibanez Portamento has a 30 "fret" neck but it is pretty unusual and even that is less than an 8:1 (18dB) frequency range. So I would suppose that part of the answer is that there IS an increase in output as you fret up the neck but it is compensated for well enough by other factors so as to not be bothersome.
Wow. Mammoth post there, but perfectly on target, khutch. If I may add one thing to that last point about the variation in frequency as you fret up and down:
This may be the most solid (no pun intended) explanation of the dB/dt behavior up the string - and would then express itself as both the effect of the string's suppleness at that fret and as the pickup's frequency sensitivity. The big question is, are those two systems separable, i.e. can you measure one while filtering the other one out.
Oh, now you've done it! I've never thought about the rate-of-change aspect of phono cartridges but what you say makes perfect sense. The brief coverage of this in my training only mentioned pre-emphasis/de-emphasis in the context of noise reduction.
Now you've got me thinking - how can the RIAA de-emphasis curve deal with cartridges of various designs/manufacturers? Or, aside from level differences, is the principle of operation the main determiner of frequency response - with individual variations such as mechanical inertia/resonance contributing a less significant "signature tone" to a particular model?
As far as guitar/bass pickups are concerned, there doesn't seem to be much (any?) info on this. Maybe it's simply that there's no need for such precision. It's a pity there aren't any 88 pole magnetic piano pickups. Dealing with a 7+ octave frequency range might have prompted some research.
Could it be that it's just harder to vibrate a given mass at a higher frequency? Of course the vibrating mass decreases when you fret a string but I wonder how these two variables interact. EDIT: Thanks to Stealth for a third variable
All of this must surely be relevant in the design of dynamic microphones. Now off to Google.
It was an early example of an industry standard. Once it was in place everyone designed all their equipment from the recording studio through to the consumer equipment according to the standard.
But then piezo-electric (crystal) cartridges used to be fairly common and would obviously behave completely differently as they aren't subject to the same dB/dt thing. I've just skimmed through the Radiotron Designer's Handbook chapter on this topic but drew a blank again.
EDIT: Sorry, I've taken the thread off on a tangent.
I never used one but I am pretty sure the piezos went into an ordinary line input rather than a phono input on your preamp.
I think they used a dedicated high impedance input rather than a line input. But that brings us back to the RIAA curve. If it was designed not just for noise reduction, but also to compensate for the characteristics of a magnetic pickup, then there would be a frequency response problem with a crystal pickup.
Separate names with a comma.