IC Buffer Prototype

[Satch12879]

Hey everybody,

I'm pretty new to the board so I'd like you all to know that my focus is on making really clean, useful, powerful items - less interested in cloning boutique effects as I am putting together original solutions for myself and others.

Last night I finished up prototyping an IC buffer box based on Jack Orman's treatise at AMZ and the GGG IC buffer project.  I know some of you may scoff but I personally find more value right now in such a device as opposed to a booster or a fuzz or some other effect.  And to me it seemed an even more logical place for a beginner to start.  

Anywho, after playing around with various other kinds of circuits over the past year or two with varying degrees of success, I think I've started to get this thing figured out.  My box is really simple, pad-per-hole construction, 1M input impedance, TL071 opamp, and a spiffy blue LED.  Essentially it's the same as what appears in the GGG project file and Jack's article.  I stuck a DPDT toggle onboard to true-bypass the buffer so I could make sure that it was working, which it does but pops like hell (left out the 4M7 pulldown resistor).

I'm glad to say that it's very quiet and essentially noiseless (besides the standard remove-your-hand-from-guitar ground buzz) along with being very subtle; I could barely hear the difference.

Does anyone have any ideas for modifications or improvements overall that I could think about? Particularly I'd like to maximize the RFI and noise rejection as well as optimize the frequency response.  I think the final version will have a split output and perhaps a mute stomp switch.  Oh, and can anyone point me at a good discussion on transformer isolation as in what I should look for in spec'ing them out?  The GEOFEX Splitter/ABY is interesting but doesn't really detail design choices too well.

In addition, I'm planning on a transistor version to see which one sounds better and is more versatile.

Thanks and I'll keep you all posted.

[cd]

http://www.geofex.com/circuits/what_are_all_those_parts_for.htm

[seanm]

http://www.geofex.com/circuits/what_are_all_those_parts_for.htm

Resurecting an old thread. I notice that many stomp boxes Use Rrf but no corresponding Crf. For example, Boss seems to always slam a 10k resister after the input.

Pros? Cons?

[onboard]

I think RG's write-up mentions that leaving out Rrf can make RFI worse. Do Rrf and Crf work together as a lowpass filter to reject RFI, or is it another concept?

[R.G.]

The GEOFEX Splitter/ABY is interesting but doesn't really detail design choices too well.

I'm serving the largest market.

80% of the people who read design alternatives have their eyes glaze and their hands go numb when they read the techie deliberations.

The vast majority of people want a nice, canned schemo/parts list (maintained up to date with Mouser order numbers) prepared layouts, and printable user's manual. I did that for a while.

I quit putting design alternatives into buildit stuff quite a ways back because the yahbutcanchajust- and buthowcome- and waitaminutecanyoudoanaccuratepartforpartlistcauseIcantfollowit- just never stops.

And the pure design articles seldom get read. They make up about 2/3 of my notes here on "the answer is in article such-and-such at GEO".

I get frustrated and tired of typing.

But I never get tired of real design discussions. So:
What design choices were you after? The primary thrust of the isolated splitter at GEO was to take the concept for the isolated guitar input at the Jensen transformer site and convert it to be usable with transformers you can get from a mail order place for under US$5.00 instead of the generally-unavailable Jensens at $50.00 a pop. I picked the easiest available transformer - the Hy-Q 10K:10K at Mouser - and made up for its low end rolloff and low primary inductance by buffering the bejeezus out of the primary.

There are no optimum designs. Or, there are no generally optimum designs. All designs optimize for some things and let other things fly. This one was for cheap and easy to build. If you are after low noise, low microphonics, flat frequency response and in general high fidelity, buy the Jensens, or find other audio hifi transformers similar to them. They're hands down better - at ten times the cost and maybe twenty times the wait to find parts. In particular, the Hy-Q's have to be themselves shielded inside a metal case to keep them from picking up 60 Hz hum and RFI. The Jensens have multilayered magnetic and electrostatic shields; that's part of what the $50 is for.

In addition, I'm planning on a transistor version to see which one sounds better and is more versatile.

"More versatile" is the opamp version, hands down. It's lower power, less finicky about power supply and rejection, less prone to RFI (that internal dominent pole thingie), and has a lower output impedance, which is why the opamp is in there in the first place. A transistor version, unless it's built to be essentially a discrete opamp will not do as well on those scores. "Sounds better" will depend entirely on whether you ever drive it to clipping (mostly a function of how much voltage you give it to work on) and whether you like the uncorrected distortions in a low- or no-feedback buffer, and the "better" will depend heavily on the person doing the listening, as do almost all listening tests.

If you're really interested in this stuff, go get "The Art of Electronics" and Graeme's opamp books. Most "treatises" on buffers and amplifiers can be found in there.

Resurecting an old thread. I notice that many stomp boxes Use Rrf but no corresponding Crf. For example, Boss seems to always slam a 10k resister after the input.

It's there, they just don't buy a separate part and stick it in the board. It's there as the parasitic capacitance of the board traces and the base-emitter capacitance of the first transistor. That's why they use a 10K - it's a bigger resistor to make use of a much smaller Crf to get RF rolloff. Thermal noise, by the way, is proportional to series resistance, so the noise of that 10K is added to the noise of your source. 1K's and discrete caps are more expensive (the cap that is) but they do the same or better job because you know where your rolloff frequency is, and the 1K has less thermal noise - which may or may not matter to you. It doesn't to Boss.

Do Rrf and Crf work together as a lowpass filter to reject RFI, or is it another concept?

It's just a dead-simple first order RC filter that starts somewhere above audio and heads for negative infinity DB as frequency goes up. You have to have them both - even if they're both parasitics.

Using a ferrite bead for Rrf is a tack some designers take. Works great, as the bead has zero losses at audio, and is not only inductive, but lossy at RF.

[seanm]

Resurecting an old thread. I notice that many stomp boxes Use Rrf but no corresponding Crf. For example, Boss seems to always slam a 10k resister after the input.

It's there, they just don't buy a separate part and stick it in the board. It's there as the parasitic capacitance of the board traces and the base-emitter capacitance of the first transistor. That's why they use a 10K - it's a bigger resistor to make use of a much smaller Crf to get RF rolloff. Thermal noise, by the way, is proportional to series resistance, so the noise of that 10K is added to the noise of your source. 1K's and discrete caps are more expensive (the cap that is) but they do the same or better job because you know where your rolloff frequency is, and the 1K has less thermal noise - which may or may not matter to you. It doesn't to Boss.

Wow thanks. I would never have thought of that. If 1K is better, would 100R be even better, or do you start to get diminishing returns?

"More versatile" is the opamp version, hands down. It's lower power, less finicky about power supply and rejection, less prone to RFI (that internal dominent pole thingie), and has a lower output impedance, which is why the opamp is in there in the first place.

So do I even need to worry about RFI with an opamp? And if so, what is a good frequency for Rf rolloff calcs. I notice that Radio Navigation starts at 9 Khz! Do you just pick a nice round number like 100 Khz?

I am very interested in RFI problems because the basement we practice in has serious RFI problems. The whole subdivision has problems with a transmission tower so everything picks up this one station. I used a TRS to 2 mono 1/4" plug adapter with 2 short 6" patch cables to an effect and got a really good radio