Variable Resonance Pickups - Fun with Buffers

[charrich56]

Here's some plots of the response curves with the values given in the schematic. The Redeemer buffer is simulated by an opamp and a 150 ohm resistor.

First plot is in resonant mode, sweeping the Tone pot from 0 to 100% with a plot line every 10%. System Q is set for 7-9 dB peaks to emulate an unbuffered pickup loaded with tone(wide open) and volume controls and some cable capacitance.

Second plot is in lowpass mode, sweeping the Tone pot the same way. With the Tone pot slightly dialed down, the response is darn near flat to slishtly past 8 KHz. It can be flatter than this but I selected the resistor value to give slightly more high frequency cut range around 5 KHz at the expense of slight peaking around 1 Khz.

[Passinwind]

There's a fairly long running thread on variable resonance pickups on Talkbass, here's the latest schematic from one of the main posters in that thread: https://www.talkbass.com/threads/a-d...#post-19323345

[charrich56]

Charlie,

That's a really good thread and I really enjoy looking at the designs there (including yours.) I followed the thread for a while; please correct me if this is wrong, but the thread is only talking about active post filtering, and not about making the pickup itself part of a sweepable or variable LRC circuit.

I really haven't come up with a name for the buffered thing, and I knew that what I really wanted to convey was "variable resonant frequency pickup" rather than "variable resonance pickup" because the latter sort of implies that we are able to control both center frequency and Q of the whole thing as in a state-variable filter.

So the goals are slightly different. We aren't going for as much sheer tweakability as much as we are going for very simple control of the resonant frequency of the pickup itself, which by itself gives enormous tonal variation, essentially re-voicing the instrument, while staying with the "traditional" general idea of what various magnetic pickups sound like, loaded by wide open tone and volume controls, and with some cable capacitance, without doing anything else. That's a relatively constant Q approach although we could revisit that with a fancier design.

The secondary goal is to be able to provide a low impedance, flat (to 8-10Khz+) response signal out of the instrument to provide a good baseline for subsequent analog and digital processing downstream.

[Passinwind]

Charlie,

That's a really good thread and I really enjoy looking at the designs there (including yours.) I followed the thread for a while; please correct me if this is wrong, but the thread is only talking about active post filtering, and not about making the pickup itself part of a sweepable or variable LRC circuit.

I really haven't come up with a name for the buffered thing, and I knew that what I really wanted to convey was "variable resonant frequency pickup" rather than "variable resonance pickup" because the latter sort of implies that we are able to control both center frequency and Q of the whole thing as in a state-variable filter.

So the goals are slightly different. We aren't going for as much sheer tweakability as much as we are going for very simple control of the resonant frequency of the pickup itself, which by itself gives enormous tonal variation, essentially re-voicing the instrument, while staying with the "traditional" general idea of what various magnetic pickups sound like, loaded by wide open tone and volume controls, and with some cable capacitance, without doing anything else. That's a relatively constant Q approach although we could revisit that with a fancier design.

The secondary goal is to be able to provide a low impedance, flat (to 8-10Khz+) response signal out of the instrument to provide a good baseline for subsequent analog and digital processing downstream.

Yep, these are really pretty different forks, just hoping for a bit of cross pollination. I don't even own a passive bass, FWIW.

[charrich56]

Everyone,

Here's an updated discrete buffer design which uses 2 JFETS (J201) and has some good characteristics, and should be easy to build. I believe that for a working pro musician level instrument, that the CAL Redeemer is a pro-quality, reliable solution, but for the DIY and experimenter crowd, this buffer should get the job done.

At least one person has expressed an interest in doing an SMT PCB for this design and I will update on this thread on how that progresses.


The Circuitlab version of the circuit is at:

https://www.circuitlab.com/circuit/7...ffer-j201-v1x/

[charrich56]

Charlie,

Thanks for the like. We are trying to do this so that the same raw PCB will work for both a 9V version, and an 18V version for the bass crowd. Going semi-audiophile with only low-ESR film caps in the signal path (no tantalum or ceramics). R1, R2, C1, C2 are on a plug-in DIP header so they can be customized for a particular application. It's looking like we can get close to a 1" x 1" board which would stick to the back of a pot. This will be useful just as a buffer, even if someone chooses not to implement the variable rez tone feature.

-Charlie

[Passinwind]

We are trying to do this so that the same raw PCB will work for both a 9V version, and an 18V version for the bass crowd. Going semi-audiophile with only low-ESR film caps in the signal path (no tantalum or ceramics). R1, R2, C1, C2 are on a plug-in DIP header so they can be customized for a particular application. It's looking like we can get close to a 1" x 1" board which would stick to the back of a pot. This will be useful just as a buffer, even if someone chooses not to implement the variable rez tone feature.

Outstanding. Yesterday I finished up my last round of mods to my prototype board, confirmed that things work as expected, and finished up the necessary board revisions for v1.1. I'll try porting your design to LTspice later this week so I can do some "what-if" sims. Using a plug-in header for modding efficacy is a nice idea.

[charrich56]

Charlie,

Thanks!

If you are firing up LTSpice, if you were interested in testing THD, I'd be interested, since CircuitLab doesn't have that facility. The Cadillac JFET to use here would be an LSK170A and I will be doing some boards with that. I had to ask Bob Cordell through Linear Systems for the SPICE model for that one. Let me know if you need it. I assume that you would already have the J201/ MMBFJ201 model or could get it easily.

The target current drain I shot for is 600-700 uA. It's pretty easy to set with R8. That was the best compromise between battery drain and drive capability since we are driving about a 2.5K load if the tone control is used.


-Charlie

[Passinwind]

Charlie,


If you are firing up LTSpice, if you were interested in testing THD, I'd be interested, since CircuitLab doesn't have that facility. The Cadillac JFET to use here would be an LSK170A and I will be doing some boards with that. I had to ask Bob Cordell through Linear Systems for the SPICE model for that one. Let me know if you need it. I assume that you would already have the J201/ MMBFJ201 model or could get it easily.

The target current drain I shot for is 600-700 uA. It's pretty easy to set with R8. That was the best compromise between battery drain and drive capability since we are driving about a 2.5K load if the tone control is used.

I think I probably already have the LSK170A model, but I'll ping you if not. I'll be happy to run THD sims once I get things going. I can do S/N too, but of course modeling specific physical builds (ie shielding and layout) really accurately is still out of my wheelhouse at this point.

That's a sufficiently modest power draw to make most people happy, I think.

[nickb]

I've been listening in the background as by sheer co-incidence I bought a single coil Far East p/u for $1.86 just to see if I could turn lead into gold. Been too busy to get further than that. No surprise it uses a rather weak ceramic magnet.

Here are a few Linear Systems FET model cards that you might find helpful.

.MODEL LSK389A NJF BETA=0.0378643 VTO=-0.4025156 LAMBDA=4.783719m IS=3.55773E-14 N=1 RD=10.6565 RS=6.8790487 CGD=3.99E-11 CGS=4.06518E-11 PB=0.981382 MJ=0.794653 FC=0.5 KF=5e-18 AF=1
.MODEL LSK389B NJF BETA=0.0350699 VTO=-0.5375582 LAMBDA=4.951m IS=2.76212E-14 N=1 RD=7.82168 RS=6.9525888 CGD=4.36E-11 CGS=4.36E-11 PB=1.2 MJ=0.8892205 FC=0.5 KF=5e-18 AF=1
.MODEL LSK389C NJF BETA=0.0278541 VTO=-0.800434 LAMBDA=0.0122435 IS=2.45217E-14 N=1 RD=12 RS=5.8 CGD=4.22E-11 CGS=4.23E-11 PB=0.9265487 MJ=0.6098477 FC=0.5 KF=4e-18 AF=1
.MODEL LSK170A NJF BETA=0.0378643 VTO=-0.4025156 LAMBDA=4.783719m IS=3.55773E-14 N=1 RD=10.6565 RS=6.8790487 CGD=3.99E-11 CGS=4.06518E-11 PB=0.981382 MJ=0.794653 FC=0.5 KF=5e-18 AF=1
.MODEL LSK170B NJF BETA=0.0350699 VTO=-0.5375582 LAMBDA=4.951m IS=2.76212E-14 N=1 RD=7.82168 RS=6.9525888 CGD=4.36E-11 CGS=4.36E-11 PB=1.2 MJ=0.8892205 FC=0.5 KF=5e-18 AF=1
.MODEL LSK170C NJF BETA=0.0278541 VTO=-0.800434 LAMBDA=0.0122435 IS=2.45217E-14 N=1 RD=12 RS=5.8 CGD=4.22E-11 CGS=4.23E-11 PB=0.9265487 MJ=0.6098477 FC=0.5 KF=4e-18 AF=1

Something in simulating pickups, a bit of a detail really, is that the voltage source is ideally modelled to rise at 6dB/octave as the coil output is proportional to the velocity of the strings. For the cheapo pickup I got, this is the model I used:

LtSpice ASC file
Cheapo_corrected.zip



My implementation plan is to include a graphic equalizer as part of the preamp so I've a very different direction. The VR idea is very attractive due to the combination of flexibility and simplicity. Anyway I'll bow out now and let you get on with it. Good luck and much success to you

[charrich56]

Nick,

Thanks for the modeling data and I appreciate your good wishes. Best of luck to someone who seems to share my hobby/passion/addiction.

Some of the Chinese pickups are indeed made for a price, but they have gotten their act together in magnet fabrication and can make some good PU's. Sourcing from there has its perils but you can get good quality. I just needed some PU's quickly since I didn't have the spare time to wind my own, and I needed 1 Henry. The ones I picked turned out sounding very good. I will be working with a local winder to home in on a good PAF-style humbucker wind to work with the VR control, and a couple of other types.

I really want to support small, local luthiers, builders, and hobbyists and not just outsource and mass-market the whole thing. But I also can see that the VR tone control could find a home in some high-end instruments.

On the pickup modelling part, a simple AC voltage source works fine through an inductor as a first-order model because the inductor has, as physics and luck would have it, an exactly opposite slope with frequency, and the modelled AC voltage source will supply whatever current is needed to maintain the same voltage. This matches what you would see with a real magnetic pickup which would otherwise be intolerably bright through a flat signal chain. It looks like you are doing a bit more sophisticated pickup model, which is great.

I appreciate your comments on the flexibility and simplicity of the VR tone control. It somewhat amazes me that this approach was not picked up on, long ago.

-Charlie

[charrich56]

Hey Charlie,

Joel de Guzman (Cycfi Research | Infinity in a Day, and his FB page) turned me on to a couple of op amps which are a combination of low current draw and low noise. I've got my eyeballs on the TI OPA188 and have designed and ordered some boards for a buffer using that part. It turns out that the MMBFJ201 SMD part isn't just a SMD version of the J201; it's got very different specs and is just about unusable for this application. The Linear LSK170A is good but you need two of them at $6.00 a pop. The LSK389A is also 6.00 but it is a dual, and would sound quite nice in this circuit.

[Passinwind]

Hey Charlie,

Joel de Guzman (Cycfi Research | Infinity in a Day, and his FB page) turned me on to a couple of op amps which are a combination of low current draw and low noise. I've got my eyeballs on the TI OPA188 and have designed and ordered some boards for a buffer using that part. It turns out that the MMBFJ201 SMD part isn't just a SMD version of the J201; it's got very different specs and is just about unusable for this application. The Linear LSK170A is good but you need two of them at $6.00 a pop. The LSK389A is also 6.00 but it is a dual, and would sound quite nice in this circuit.

Have you found a reasonable source for the LSK170A and/or LSK389A? Trendsetter wants $24 for shipping a single piece and DIYaudio never seems to have the A spec pieces in stock.

I haven't tried any CMOS based opamps in many years, but the slew rate of the OPA188 is a bit below my usual lower limit. Might be fine in this application though.

Hope to get back to some circuit modeling soon. Things took an unexpected swerve when the first luthier I demo'ed the new preamp for asked me to provide a few samples for his NAMM booth next month. I built three different ones last week, but I fully expect that my friend will end wanting something a little different than any of these.

[charrich56]

Charlie,

I have gotten samples for LSK170A and LSK389A direct from the manufacturer, Linear Systems. I just got pricing from those from Kevin Hicks at Linear and basically any qty. up to past 200 both the 170A and 389A are about $6.00 each. Obviously the 389A is a better deal if you need more than one JFET in the circuit, as I do.

They are good to deal with for small purchases, generous with samples, and they are aware of the audiophile DIY crowd, and also the high-end audio/recording crowd (because their JFET's are used in some mondo expensive ribbon mic preamps.)

They also have other JFETS and they are a second source for some of the Fairchild, Intersil, Toshiba, etc. parts.

My opamp SMD boards are just about to show up from OSHPark, and I'm going to build them first most likely.



Linear Integrated Systems, Inc.
4042 Clipper Court, CA 94538
Main: 510-490-9160
Fax: 510-353-0261
Sales@LinearSystems.com

Linear Systems || Home

[Passinwind]

Charlie,

I have gotten samples for LSK170A and LSK389A direct from the manufacturer, Linear Systems. I just got pricing from those from Kevin Hicks at Linear and basically any qty. up to past 200 both the 170A and 389A are about $6.00 each. Obviously the 389A is a better deal if you need more than one JFET in the circuit, as I do.

They are good to deal with for small purchases, generous with samples, and they are aware of the audiophile DIY crowd, and also the high-end audio/recording crowd (because their JFET's are used in some mondo expensive ribbon mic preamps.)

Cool, thanks. I have something else in the works that would definitely constitute a legitimate reason to request a few samples. Opamp based designs are one of my main comfort zones though, so I'll be especially interested to hear how your v1.0 board works out.