Input Buffers, Active Pickups, Bass, Etc...

[railhead]

Quick question: what are your collective thoughts regarding the use of input buffers in effects that will ONLY be used for basses with active pickups?

Personally, I don't see the point -- and I remove them from the circuit -- since the active pickups should be doing their own "input buffering." That said, I've read some threads at *other* forums where builders are adamant about using input buffers within circuits that will be used for active basses.

Opinions? Thoughts? I'm talking about OA-based circuits ala TS, 808s, etc...

[R.G.]

If the active pickups are any good, buffering after them is not necessary. If the buffers are good, they won't hurt except where the active pickups put out so much signal that they overwhelm the buffers, which I've seen happen.

It's almost like you have to know the details of the pickups, buffers, effects, and everything else involved, and that makes a general rule difficult, doesn't it? 

[railhead]

Heh, yes. I suppose a basic input buffer like the good ol' Tubescreamer wouldn't hurt, but I've never noticed any need for them.

[Cliff Schecht]

It makes some sense to remove the input buffers from your effects, until you need to use said effect with a passive pickup system. I never remove the buffers because they don't usually hurt anything.

To add onto what R.G. said, signal levels coming out of active pickups can be well over a volt (EMG 81's are 2V P-P!!), so make sure that if you do leave your input buffer stage in that you have enough headroom to swing completely without clipping. I've never ran into these problems with EMG's or any active pickup systems that I've made but it is something to be aware of.

[railhead]

I see your point. Perhaps I need to move to using JFET buffers with high headroom (rather than TS-style tranny buffers), and using something like what, 2M resistors in parallel, like this?



If I'm thinking correctly, wouldn't that setup an RC filter that'll get around 1.6Hz corner frequency?

What's the best way to determine what the value of R3 should be?

[earthtonesaudio]

What's the best way to determine what the value of R3 should be?

Find how much current you want through the device, and how much output current drive you need.  I pretty much always use between 2k2 and 10k.  Less resistance increases power consumption, more reduces output current drive (but if you don't need much...).

[railhead]

If I'm using 9v for Vcc and I want to push say, 3.75 ma, using the formula:

(Vcc - (Minimum Rds(on) * Ids)) / Ids = Total Resistance

I get: (9 - (0 * .00375) / .005 = 2400 ohms

That's about a 2.4k. Is that the proper math, RG? What about the corner freq?

[Cliff Schecht]

First off, get rid of R1. The configuration you're using is a self-biasing configuration and doesn't require a "voltage divider" input. Just put a 1 meg resistor to ground on the gate. For your emitter resistor, go with something like 4.7k. It should give you the best trade-off between voltage swing, current output and power consumption.

[earthtonesaudio]

I believe there are 2 reasons to use self-bias as Cliff suggests: low parts count, and ground referenced input. 
But if you want maximum headroom or reasonable repeatability, I'd go with the voltage divider bias.

[railhead]

Okay, I read through Orman's lab book, and he has a section regarding simple buffers -- and he has one listed just like I drew here. This is his:


(link to buffer page)

He likes the "divider" version in the case of having no Vr bias source, which typically suits my situation. He also has 3k3 at my R3 location, which is higher than my 2k2 math. I guess my main question is now about the accuracy of the math/formula I quoted above -- and the corner frequency.

[Cliff Schecht]

I take back what I said earlier after rethinking the circuit. If you're using an source follower type design as shown here, the split resistor input does make more sense. As someone else said, it should maximize headroom. Honestly, the value of R2 in that schematic isn't super-duper critical because in reality, the input stages of all of your effects should require almost no bias current (we're talking pA-nA here for good op amps).

As for your cutoff question, it depends on what you are feeding the output into. Your output cutoff is determined by 1/(2*pi*Cout*Rload), with Cout = 10 uF and Rload equal to the input resistance of the effect this is feeding. With a 10 uF cap on the output and assuming a load resistance of 1 meg, you have a cutoff of .016 Hz! This is, of course, also in conjunction with your cable resistance and capacitance which should be negligible if you use decent quality cable and avoid 100 foot runs of cable.

[railhead]

I was calculating -- or attempting to calculate -- the corner freq using Jack's RC Calculator. If I'm using it correctly, I put in .1uF (C1) for the cap and the 2M split resistor gives me 1 million ohms. This calculates to 1.6Hz. It looks like your math is coming after the buffer toward output.

Am I not looking at how to use Jack's calculator correctly?

[R.G.]

My gripe with all packaged calculators is that if you don't already understand what's happening behind the scenes, you don't know if you applied it right or not, and do not develop a sense of whether the answer is correct. Worse yet, dependence on packaged calculators forever prevents you from learning.

The frequency of a single RC rolloff is ALWAYS F = 1/ (2*pi*R*C) with r in ohms and c in farads. So to use another "calculator", the kind with buttons, you press
"2"  "x"  "3.14" "=" "x" (value of r) "="  "x"  (C in farads) "=" then press the "1/x" button and you have the answer. What's neat about this is that when yo do it this way, you actually understand the calculations.  On calculators without scientific notation, a 1uF cap is 0.000001 (five zeros); on a calcuator with scientific, that's 1E-6.

There's a good rule of thumb in there - if you can't do it on an arithmetic, non-programmable calculator, you don't understand it.

Crutches are to help you walk when you can't do it any other way. But you don't find a pair that you like and keep them the rest of your life.

[earthtonesaudio]

I like to use the calculator on my cell phone.  It really makes you work (6 button taps) if you want to do something really math-y, like use parentheses.   

[railhead]

Hey, I never said I understood what was going on with the RC Calc -- that's why I'm asking. I saw the notes about fattening-up a TS and how the freq cutoff was calculated -- and I asked if I was understanding its use based on a different circuit. I see now that I was looking at the "math" at the input, rather than toward output like I should.

Second, thanks for showing me the full equation.

[Cliff Schecht]

My gripe with all packaged calculators is that if you don't already understand what's happening behind the scenes, you don't know if you applied it right or not, and do not develop a sense of whether the answer is correct. Worse yet, dependence on packaged calculators forever prevents you from learning.

The frequency of a single RC rolloff is ALWAYS F = 1/ (2*pi*R*C) with r in ohms and c in farads. So to use another "calculator", the kind with buttons, you press
"2"  "x"  "3.14" "=" "x" (value of r) "="  "x"  (C in farads) "=" then press the "1/x" button and you have the answer. What's neat about this is that when yo do it this way, you actually understand the calculations.  On calculators without scientific notation, a 1uF cap is 0.000001 (five zeros); on a calcuator with scientific, that's 1E-6.

There's a good rule of thumb in there - if you can't do it on an arithmetic, non-programmable calculator, you don't understand it.

Crutches are to help you walk when you can't do it any other way. But you don't find a pair that you like and keep them the rest of your life.

I made a point last semester to ditch my calculators (I carry three: TI-89, HP49G+ and a Sharp scientific) and start crunching everything by hand. Since my advanced communications circuits teacher made us take our tests without using calculators, I had to refill my math toolbox pretty fast. I ended up with an A in the course and a much firmer understanding of the math that goes behind a lot of communications techniques (although my synth background already gave me a good grasp on Fourier and time-domain signal analysis). Unfortunately, I'm still not very good at simple math (2+2=chair!). I have no problem deriving complex transfer functions using only pencil and paper, but I still go to the calculator for simple number crunching.

I found my transfer function skillz especially useful in my RF communications project lab, where I had to design all of the hardware for my project. I was the only person in the class who could design and tune all of the functional blocks and I spent a lot of time helping others to get their projects working. My Futaba-compatible 72 MHz receiver worked beautifully and I got an A in there as well .

[runmikeyrun]

I have an active Ibanez that I hot rodded to slam my tube amp- wired the two stock pickups in series and then put in a dimarzio X2N guitar pickup which is about the hottest pickup they make.  Needless to say it doesn't play well with a lot of pedals.  If i turn the volume knob down i get a good result, so i've been playing with padding the input on my pedals.

[sjaltenb]

I plan on using this same simple buffer for my main guitar input splitter (using the extended Splitter design that Jack has on his site) to feed the tuner out and the FX chain, and also the same design as a final output buffer to feed the two amps. This seems simple and effective, any suggestions/problems?

[DavidRavenMoon]

Some active pickups, like EMG, don't have as low an output as they like you to believe!  It's around 10K on some models.  The output impedance of their preamps is about 2K.  So a buffer can still help some active pickups.