> opinion on using oxygen free copper wire in the pickups?
A trace of oxygen binds-up other minor impurities in copper. OFC is not generally better, and can be worse long-term. The conductivity is so close it does not matter.
Some special OFC is resistant to hydrogen embrittlement, a problem we do not have.
A trace of oxygen binds-up other minor impurities in copper. OFC is not generally better, and can be worse long-term. The conductivity is so close it does not matter.
Some special OFC is resistant to hydrogen embrittlement, a problem we do not have.
I'm not sure what I did to ruffle your feathers. But have a nice day anyway.
-Gnobuddy
Reading that 1994 interview ( TA Audio Topics - EE Times Interview with John Murphy on Tubes vs. Solid-State ), Murphy mentions that his work was used by Carvin in some of their solid-state guitar amps, starting in 1987. He went on to say that his concepts were "...still in production in Carvin’s SX series solid-state guitar amps."
So I went looking for Carvin schematics, and to my surprise (because Carvin folded a while ago), I found many: Guitar Amplifier Schematics
Even better, there is a schematic for an "SX Guitar Pre-Amp Rev B" with a 1994 date, the same year as the EE Times interview with Murphy.
I won't post that schematic since it's presumably still copyrighted, but you can look it up easily enough at the above link.
The schematic is complex, but it looks as though all the tube emulation is done in three op-amp stages, each with one or more clipping diodes.
As Murphy says in the EE Times interview, the diodes are never arranged to produce symmetrical clipping; and furthermore, I notice at least one of them is set up to create a DC voltage across an interstage coupling cap, a DC voltage which varies with guitar signal strength.
In other words, that diode is intended to do the same thing that grid current flow in an overdriven valve does - cause the DC operating point to vary with signal strength, which in turn varies the duty-cycle and harmonic content of the resulting clipped signal.
I wonder if there are any sound clips out there on the 'Net for this or similar Carvin amps. It would be interesting to see how this early attempt at emulating valves using op-amps actually sounded.
-Gnobuddy
It did not sell well. I have never seen one on stage or for sale. It failed because it fell short of emulating triode stages; it was an attempt, nothing more. The Soldano SLO-100 was hot for a while, but many loathed the high treble cut due to the oscillation clamp across one plate resistor and other contouring factors. Here's an analysis: Soldano Super Lead Overdrive Channel Circuit Analysis
Concerning differences of opinion, all this book knowledge is necessary, but what about a scientific test at some juncture?
What is an accurate, reasonably priced way to measure the noise of a circuit? I saw one guy using a micro voltmeter. Big bucks$.
Sounds like shifting the location of the quiet, first stage buffer/amplifier from the amp to the guitar!
How about using more powerful pickup magnets and less turns of wire? This would be treating the root cause and not the symptoms.
Your calculations COMPLETELY IGNORE CURRENT NOISE, which, as I am sure you know, is the dominant source of noise in a circuit with a source impedance above ~ 3KΩ.
The formula I used for calculating the signal-to-noise ratio of each opamp is:
=-20*LOG10((SL)/(((SQRT(POWER(VN,2)+POWER(CN*SI,2))*141.35))*1/1000000))))
where SL = signal level in mV
VN = 1K Hz voltage noise in nV√Hz
CN = 1K Hz current noise in pA
SI = source impedance in KΩ
With signal level = 10 mv (a quiet guitar) and a source Z of 25 KΩ, the TL072 comes in at 72 db S/N; an OPA 1652 is 82 db S/N. There's your 12 db. You will probably say that 72 db down is inaudible--I will disagree---84 db down probably is inaudible, but not 72.
Last edited:
This might help - it uses a dual-channel 'scope: https://www.edn.com/electronics-news/4383541/Calculate-and-measure-noise-values
I haven't tried this method myself, so I have no personal experience to offer. But the source is a reputable one.
It's more than that: because the on-board preamp buffers the pickup signal, you no longer need to use a noisy 500k volume pot. You can use, say, a 10 pot...and that will not contribute much noise, because the pickup signal has already been amplified before it reaches that pot, so S/N ratio is better.
It's certainly open for experimentation. It has been tried, several times, and there are some known issues:
1) If the magnets get much stronger, they pull on the guitar strings, affect the overtone structure by shifting the frequencies of overtones, and therefore do weird things to their sound. This already happens with Fender single-coils if you adjust them too close to the strings.
2) If you keep the same wire thickness and wind fewer turns, I think S/N ratio gets worse; for instance, half the turns will halve the signal and halve the DC resistance. Thermal noise voltage goes as the square root of the resistance, so it only drops to about 71% of its former value. So now you have half the signal, but more than half the noise...S/N ratio is worse than before.
3) If you keep lots of turns, and use fatter wire, you get lower DCR and thermal noise (good), but the pickup coils become huge and bulky. Les Paul himself (the musician, not the guitar) went this route, winding his own enormously bulky, low-impedance coils on the guitars he himself played. (I think they also had far fewer turns than a standard Gibson humbucker.)
4) If you Google up some clips of Les Paul, you can hear the tone he got from those low-impedance pickups. It is not a popular sound - even in Les Paul's time, nobody else wanted that thin, cold, metallic tone. Les Paul loved it, though, and went to great trouble to get it.
The change in tone occurs because the several-Henry inductance of a conventional high-Z guitar pickup interacts with the guitar cable capacitance to form a resonant peak somewhere in the 3 kHz - 4 kHz area, and this has a considerable effect on the tone of the raw guitar signal. A low-Z pickup coil has much less inductance, so this resonant peak shifts much higher in frequency, and the guitar now sounds thinner and brighter.
I discovered this effect for myself by accident in my twenties, when I wound my own pickups for the one and only electric guitar I built. There was no Internet then, and I couldn't find a single book on guitar pickup design at any of the three libraries in my little town. So I guessed - I wound several thousand turns of wire, and ended up with 4 kilo ohms DC resistance.
I didn't know how that compared to any commercial guitar pickups, but when I got the whole guitar together, my pickups had low output, and sounded very clean and bright - more like a plugged-in acoustic-electric guitar than the e-guitar sounds we're used to hearing.
There are commercially available low-Z guitar pickups; they don't sound like normal e-guitar, either. The Shadow Nanomag pickup is one example; it's added to some Epiphone Les Paul Ultra models, and is used to generate "kinda sorta acoustic" tones from the guitar. Once again, probably not what we want from a solid-body guitar. (But the Ultra's have conventional pickups too, so they can also make the e-guitar sounds we want to hear.)
As I mentioned before, Fishman Fluence pickups are yet another low-Z pickup. This time, the goal was to get away entirely from winding pickup coils. The Fluence pickups use a stack of "coils" that are actually a few concentric copper tracks on printed circuit boards; the entire coil probably has something like a hundred turns (my guess), rather than the several thousand turns in a normal pickup.
I assume both output level and output impedance are very low, but you never see either from the outside world: these are active pickups, with a preamp built into the pickup itself, and very heavy EQ applied to try and make these low-Z pickups sound similar to conventional high-Z pickups, and put out similar signal levels too, so that they work with all the thousands and thousands of guitar amps that are already out there.
I have never seen these pickups in person, but from the few clips I've heard, I'm not convinced they sound like the real thing. In the clips I heard, one of the Fluence pickups that's supposed to sound like a vintage Gibson PAF humbucker actually sounded, to my ears, cold, thin, and bright: everything you do NOT expect from an actual PAF.
(But exactly what you'd expect from a very low-Z pickup. Hmm.)As PRR said, I don't think pickup hiss is an issue for the majority of e-guitarists, including myself. I suspect those for whom it *is* an issue (such as yourself) are the ones playing music genres that call for very high gain in the signal chain.
In these same genres the guitar signal is so heavily processed before you hear it that maybe the clean tone of the guitar itself doesn't matter - maybe you could make a low-Z pickup, add a whopping amount of treble cut, and make it come out sounding the way it should on the far end of a Boss MT-2 or Peavey 5150?
-Gnobuddy
No need to dump thread, as you have valuable insight and opinions. There does come a time where one of the parties in a dispute should refrain from argument when no resolution is in sight.
Basically, we are bantering about ways to reduce idle hiss in high gain amps and dirty, relatively inexpensive fuzz boxes. If one builder might consider using high end opamps or foil resistors to squeeze out the last drop of noise (like me), and another does not care about a few dB, they are both correct in their decision!!
They call it "rough". Besides the procedure would have to be repeated for each change in UUT.
How about a test jig where many changes can be made in succession and yields fairly accurate readings immediately? I watched a video of a guy shaking out noisy transistors from a batch, using an amplifier and speaker. If it was loud the component was tossed. The process was repeated with the remainder until only the quietest ones were left. Not scientific, but much better than a hope and a dream fueled by extra cash.
A very low noise amplifier is required to boost the noise signal above the noise floor of the scope, or DMM for that matter. Here's a readable paper with a practical circuit. It is cheap and "easy" to build.A quick change jig may be added to this: View attachment Noise Measurement Opamp tidu016.pdf
I would use NiCd batteries for power and to avoid error caused by the power supply. These are the quietest batteries according to this paper: View attachment Voltage Noise in Chemical Batteries NIST.pdf
In retrospect, I see that NEITHER FET opamp (TL072 or OPA4562) adds much in the way of current noise to the total noise figure, like a bipolar amp does. And I see that the voltage noise added by a TL072 opamp is about the same as a 25KΩ resistor. BUT----is the thermal noise of a pickup coil the same as a resistor? The pickup has ~ 4KΩ of DC resistance plus ~ 2 Henries inductance and ~ 200 nF of capacitance. The noise calculations I find define the thermal noise only in terms of the RESISTANCE portion of the impedance; if so, the pickup only generates about 8 nV√Hz. Combine this with the opamp's noise contribution would give the PU + TL072 = ~ 26 nV√Hz and the PU + OPA1652 = 12.5 nV√Hz. That's a decrease of 6.3 db in noise---definitely audible and WELL WORTH the $1.27.
Last edited:
I don't think 200nFd is right, even including cable and first tube. Ah, 200pFd? Plus cable and tube C, say 1,000pFd?
The *pickup* is 4k resistive bass to low-mids and ~~100k resistive at its top-peak, roughly 3kHz. That upper peak is the most annoying, though it is not wide.
As Gnobuddy keeps saying, as soon as the axe knob comes off of "10" a lot of this reactance is swamped-out by the pot. Half down you can just pencil "50k" and nobody can show you wrong. Those in-between 5 and 10 points deserve attention, but not too much. If it mattered, we'd put the first stage IN the guitar. Microphones did that before WWII.
The *pickup* is 4k resistive bass to low-mids and ~~100k resistive at its top-peak, roughly 3kHz. That upper peak is the most annoying, though it is not wide.
As Gnobuddy keeps saying, as soon as the axe knob comes off of "10" a lot of this reactance is swamped-out by the pot. Half down you can just pencil "50k" and nobody can show you wrong. Those in-between 5 and 10 points deserve attention, but not too much. If it mattered, we'd put the first stage IN the guitar. Microphones did that before WWII.
PRR is right. The impedance peak of 100k or more at around 3kHz resonant frequency is in fact a pure resistor of several 100k . That is the real part of the complex impedance and a matter of circuit losses, 10k of dc resistance being one contributor. With zero DC resistance and no losses at all the ideal parallel resonant tank would have an infinite impedance.
Last edited:
You might want to mention that to Pat Quilter, and point out the foolishness of his company's switch to OPA1654 opamps from TL074s in his guitar amps; he said that it reduced the noise floor of them by 8 db.
Last edited:
That's probably the definition (the marginal cases don't matter).
Impedance is both resistance and reactance. Only the real resistance hisses. However the effective resistance in a maze of parts can be different at different frequencies. In this case the pickup is a tuned circuit and sure can have several resistance values.
First plot. Apparent resistance vs frequency for a guitar input network at several pot settings. The level part of the Z is the R. In the lows it is just the 4k DC resistance. But it is also "level" at the top of the 3.5kHz peak. This hisses like a ~~100k resistance.
Not trusting my eyeball, I found a noise plot function in my sim. It does well on passive networks (active part models don't always have the details). Second plot shows ONOISE, referred to pickup internal generator, at several pot settings. At "10" the LF hiss is very low. It rises to a peak at 3.5kHz.
Taking a rough integration over frequency bands we have a narrow but high peak amounting to a microVolt, with most else amounting to a half-uV.
The turn-down noise is flat at 1.6uV. This is the noise *referred to source*, inside the pickup. Referred to the amplifier input the "5" hiss would be 2X less or 0.8uV. (in 5kHz bandwidth not the 20kHz band often assumed in hi-fi) So yes the 12AX7 and TL072 are not "noiseless", giving like 4dB Noise Figure.
Attachments
Who is Pat Quilter? PPR just pointed out that the TL072 adds 4dB to thermal noise. Following this there is no way to improve noise figure by more than 4dB. Which is audible - but not impressive.
Another cup of tea is the residual noise with vol pot at min, i.e. amp input shorted. In that case voltage noise of TL072 with 18nV/sqrt(Hz) makes an audible difference compared to OPA1654 with 4.5nV/sqrt(Hz). For regular playing the guitar this is a minor effect - but to be frankly - I like it dead quiet when turning down the vol knob and just for that reason I use OPA1678 in my pre-amps.
Last edited:
So we should trust advertising claims for one consumer product more than we trust the science of thermodynamics and statistical mechanics, developed by some of the world's greatest geniuses, and tested a million times in thousands of applications and experiments over several centuries?
Appeals to authority are not a part of science or engineering, because authorities (being only human) frequently turn out to have the proverbial feet of clay. The universe does what it does, regardless of advertising claims, and regardless of who says what; the purpose of science is to find out what the universe actually does, regardless of what we think it should do.
One of the most powerful things about science is that it doesn't depend on anyone's say-so. The concept of repeatability is a fundamental prerequisite of science, i.e. it's required that the same experiment can be done anywhere by anyone to independently verify (or disprove) a claim.
(Of course the experiment itself has to be done competently, and sometimes this requires so much money and so many resources that it is hard to replicate.)
Fortunately, there are no such difficulties in this particular case. You could, if you like, try the experiment yourself - compare the actual noise output from a TL072 and an OPA1652, both set up with a 27k metal-film input resistor from (+) input to ground, suitably low-impedance metal-film feedback resistors, and powered by a pair of 9V flat batteries.
Set them up for a high voltage gain of, say, 1000x (+60 dB) each. That should provide a couple of mV of output hiss. Follow up with an additional mic-preamp or other low-noise, high-gain amplifier, and monitor the output with headphones.
This will quickly let you find out what your own ears have to say about the hiss level from each circuit. (And then by all means follow up with the oscilloscope measurement outlined earlier in this thread, to get some quantitative data as well.)
The biggest challenge will probably be to get the 60 Hz mains hum (and its harmonics) quiet enough; I think a grounded full metal enclosure, and high-pass filtering the output at, say, 1 kHz might do the trick. We're after hiss, after all, not rumble, and the high-pass filtering will help drop hum levels.
As ITPhoenix says, it's not a religious battle or anything; if it makes you happy to use the OPA rather than the TL0, there's no significant damage done. (And so significant benefit, either, for the reasons I've talked about in several earlier posts.)
Thinking back, the last time I was bothered by audible hiss in a well-engineered product was when I was still listening to audio cassettes in the late 1990s. Even with Dolby C and Dolby HX switched in, hiss was audible. The signal output from the tape head was so small - maybe 1 mV, maximum - that it was quite hard to get the playback electronics itself to have a noise floor more than 60 dB below full signal, never mind the hissy nature of the tape itself, so you could also hear hiss if you pushed "Play" without a cassette actually in the transport.
I've heard obnoxious amounts of hiss since then, but only in bottom-of-the-barrel electronics products; the $100 Kustom shoe-box sized "P.A." I used for a while as a monitor, and the larger Harbinger P.A. I tried briefly. Both these products were basically powered speakers with a crude audio mixer integrated, and in both cases, that mixer was hissy.
Other than those two bad apples, though, hiss from noisy audio electronics hasn't been an audible annoyance in my life for a long time.
-Gnobuddy
As I understand you are not a "loud" guitar player and never play with hi-gain where hiss becomes annoyant, so you will not perceive hiss. But this is your individual experience and point of view.
Agree completely. I never intended to imply otherwise!As I understand you are not a "loud" guitar player and never play with hi-gain where hiss becomes annoyant, so you will not perceive hiss. But this is your individual experience and point of view.
-Gnobuddy
And a big thanks to PRR for that simulation!
PRR's sim used a 250k volume pot, typical of Fender guitars with single-coil pickups. It would be nice to see the same simulation re-run with a 500k volume pot, typically used with humbucker pickups. (Simple theory - flat impedance - predicts 3 dB more thermal noise from the guitar, and therefore, a noise figure that drops from 4 dB down to 1 dB for a TL072.)
Most of my own guitars have humbuckers, and it's probably safe to say that the majority of e-guitars used in high-gain genres also use humbuckers.
Agreed. And with a 500k volume pot, even less than 4 dB improvement is possible. Maybe only 1 dB, if my rough calculation above isn't altered too much by the pickup's frequency-sensitive impedance.
Yes, but only if the noise from the TL072 is loud enough to be heard. This will depend on how loud you play your guitar.
In my case, I don't hear any hiss from the TL072 at my typical playing volume, so I there are no improvements to be had from a quieter op-amp - inaudible is inaudible!
I will say that personally, I never turn down the guitar volume to zero. If I need zero volume out of the guitar amp, I turn the master volume on the guitar amp to zero, not the guitar volume control.
I think we have now reached the point in this thread where the science has been pretty thoroughly clarified, and now we're down to personal opinions alone. Which, of course, will vary widely in any healthy society.
-Gnobuddy
Not quite. We don't know what will be plugged into the amp or box (unless it's our own), and the HB/500k combo makes a significant difference. Not everyone uses the pot on 10 all the time. How about the pickup selector, one lower setting for rythm and another for lead? And how about some room for additional volume in changing conditions, like live?
Moreover, the component values on the periphery make a difference, and these opamp numbers that have been proffered have significant differences that may affect the final outcome. A particular opamp as buffer might be a different story than, say, 2 in series for a total gain of 60 to replace the first stage of an AX7, or G=3 in my case.
Actual noise testing should be done at this point, for the extremists.
Last edited:
If you can get input circuit losses down to 50 Ohms (and have no other goal), just use LT1028. It is real quiet with near-short input.
However if you ever turn-up from dead zero, at 50k source the noise current flowing in the source is off the graph, 3uV-4uV.
Yes, OPA1654 does look fine over the range of guitar impedance.
Last edited:
It does look very nice in many respects. I like the lower power supply voltage requirements, for instance. Noise performance is better than it needs to be, but there's no harm in that.
However, I think it's only available in surface-mount packaging - at any rate, I can't find a through-hole version.
For me, that's a major turn-off, especially when there are no audible benefits. People with younger eyes might not have the same problem.
FWIW, Digikey Canada has the OPA1652 in SOIC 8 packaging for $3.09 ( OPA1652AIDR Texas Instruments | Integrated Circuits (ICs) | DigiKey ). They also have a nice little SOIC 8 to DIP 8 adaptor board for $3.68 ( LCQT-SOIC8-8 Aries Electronics | Prototyping, Fabrication Products | DigiKey ).
At nearly seven bucks for the opamp plus adaptor, with little or no hiss reduction, this combo isn't for me. But maybe someone else will think it worth while.
-Gnobuddy
Most of my e-guitars have the typical Gibson control layout, with independent volume and tone knobs for each of the two pickups, and a two-way selector switch. And yes, some of the time, I set one vol/tone pair for rhythm, and the other for lead, and switch between them.
(But my favourite sounds usually involve both pickups mixed together, so sometimes I will turn down both volume knobs for rhythm, and turn both back up for leads.)
Me, I tend to do that with the master volume on the guitar amp. But, as that video of Zakk Wylde playing his "Farewell Ballad" shows, a really good guitarist integrates the volume and tone pots on the guitar into his / her playing, using them as one more mode of expression. I should learn to do more of that myself.
-Gnobuddy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.