Buffers and Voltages

[liquids]

I've searched a bunch on this topic and learned a lot, but could use some direct info and opinions.

I'm looking at the possibly going to be eliminating my Boss TR-2 from my rig, which is also my my buffer, and will be moving toward a stand-alone buffer of some kind, as I definitely need a buffer.

I run everything off my VS 1-Spot, and it's great.  I've got two, even, one for the board and one for the testing rig/backup.   

Story is, I've messed with a few options for buffers--ROG and GGG EA Trems clipped on me with the depth down, op-amp buffers, etc, all clip on me, to some degree, with hot signals.  I experience this a lot - my bridge humbuckers are Duncan Custom Customs, and without fail, if I smack a chord, the buffers break up, if not a lot than a little too much. The TR-2 in bypass pedal basically doesn't (at least so little that it's okay), though I've experienced other Boss pedals that seem to.

Last night I tried Ormans simple op-amp buffer with a TL081, and it was close, but it did give a little grit.  Though clean with my (Seth Lover) neck pickup, I felt the sound of just the op-amp buffer was a tad better, and fuller sounding than the Boss buffer, if anything. 

I tend to place my buffers after my dirt pedals as I tend to think most dirt pedals sound a tad better 'seeing' my pickups instead of a high impedance input signal.  Next, since my rack unit has a 10k input impedance I always need a buffer in there. I kept the TR-2 in line, and had the op-amp buffer circuit before it, switching it in and out as an 'additional' buffer to hear it.   With my dirt pedals on, sure enough, the difference between the added buffer made a less pleasant tonal difference--it sounded like the input was being 'overloaded' by the dirt pedals which were near unity gain.

I'm crossing my fingers about an up-and coming Tremulus Lune build, to replace the TR-2, which hopefully won't clip (or at least less noticably) than the other Trems I've tried building. I'm considering getting a higher voltage for building a stand-alone op-amp or other buffer-type circuit for 'input' headroom.  I could buying and boxing up Dave Barbers b-buff which runs 24V, and be done with it.  That is risky without demoing it.  That being said, I'd enjoy the option to run other pedals at higher voltage, so maybe buying the Dunlop 18v adaptor would be nice.  Maybe building a charge pump and running one of my one spots into that for those that can handle it and or sonically 'prefer' it (I could easily check and or mod my pedals to make sure they'll run safely).

Will this fix my problems?  I just want to be able to run a buffer and other pedals that should sound clean (Tremolo, for example) without over driving the inputs with just my pickups.  I was super disappointed with the EA Trem because I love the sound, but no matter what I did or versions I tried, I couldn't get it to totally not to clip with my pickups, let alone what happened when running a dirt pedal into it.   

The Duncan JB is a common pickup and pretty hot -- does anyone else notice this issue?  What buffers do you use?

Last but not least, the boss buffers get a lot of bad press for the boutique guys, but it clips the least of the ones I've tried.  Maybe I could build a stand alone of the one in the trem? I looked at the schematic http://www.schematicheaven.com/effects/boss_tr2_tremolo.pdf, but as a rookie I can't follow the signal path when it's off as I don't understand how the JFET switching works, the JFET and BJT part numbers are obscure.  It seems a bit complex outside of the input FET and output BJT emitter stages, which I think I've also tried with no success, but will try again if that is all that's going on there.  Similarly, will these kind of stages have more 'input headroom' if run at 18v, etc?

[R.G.]

I encourage you to continue looking for a build it yourself solution if you want to do that.

However, Visual Sound just introduced the "True Tone" buffer which does pretty much what you're looking for. It has an internal buffer running from about 26V, and takes the same 9V power input. It was designed as a solution to just about exactly what you're describing.

[Ripthorn]

Have you tried Jack Ormans buffer at higher voltages?  What you might could do is put in a charge pump to up the supply voltage from 9V to 24V or so and see what that does.  Won't cost very much and worth a shot.  I always like to have a charge pump on hand for experimenting.

[liquids]

Just got a charge pump that I'm using extra diodes to keep under 16v until my order of higher voltage caps comes in.  That being said,  With nothing connected and getting 14v at the output of the charge pump, I'm down to 10v once I connected a tremulous lune!  So, I removed some of the extra diodes to get it so the effect sees more like 14v when connected...

Is that much voltage drop typical? Maybe I have something wrong?

[liquids]

I encourage you to continue looking for a build it yourself solution if you want to do that. 

Surprisingly, after starting and stopping for a while, trying different things on different days in a non-controled, hardly-scientific model, it seems I've concluded that bog-standard BJT buffers are the way to go for me...?  Unfortunately the lack of scientific consistency (same day, same setting, same chips, same voltages, etc) distorts some of my findings, but I'm getting there in the search for the right buffer for me...

I had ignored BJT buffers primarily due to my over-saturation with the mythology of tube lore--"silicon is bad."  Likewise all the talk on here said BJTs are good 'output buffers.'  Since I was focusing on a buffer that would not clip at the input, I ignored them.  As I said before, JFET buffers seemed to be clipping despite some higher voltages, that I vaugely think the last time I tried it at 24v it was cleanish, though my recent attempt at 18v was not.  Op-amps were iffy...I wasn't positive I wasn't clipping the input even when designed well, and they did seem to need higher voltages than 9v to function optimally...

Recently I tried again,t I got to listening harder, and at one point I had unplugged everything in line that would induce my need for buffering.  I realized that the test of switching to my hot bridge humbucker, and strumming very hard even broke up the input of my Super reverb set cleanly--albeit ever so gently and harmonically.  I realized in retrospect, that, while some examples like the JFET buffers were producing their own clipping under this test, no matter how clean the buffer was, it wasn't going to be able to eliminate that, so that helped.

I've tried BJT buffer on two different occasions, and at the one spot's stock 9.43v, I can't get it to clip. sounds very neutral, in fact, and like there is no buffering at all when A/Bing true bypass, which is usually not what my ears percieve with op amp buffers, let alone Jfet buffers.  When I plug in my rack unit which desperately needs something bufferish going into it switching it in and out (TB vs the buffer) it's a night and day difference between the stock sound and massive loading of the pickups going straight in.  It's minimal in size, and best of all, possibly no charge pump needed, though I may add one to the buffers I box up after more exponentiation, just because, why not. 

Back to testing with no rack unit, I put a volume pot after the buffer, going straight into the amp, just to see.  Same test-slamming a chord, but I turned down the volume coming out of the buffer.  Sure enough, that cleans up the amps break-up nicely, since it's not seeing such a hot signal at ITS input.  With that clipping out of the way, here I should be able to hear any clipping at the input of the BJT buffer...but nope, I can't hear anything of the sort, at this point.

Likewise, I've been running my breadboarded stuff into the said buffer, and it seems similarly transparent...it sounds like they are the most resilient yet to not clipping despite hot signals.  I could swear I read that there is an advantage to using Darlington's here, and/or that the higher the HFE of the BJT used the better the buffer for lower output impedance, if I'm not mistaken...is that correct?

Still to experiment if running this buffer into my pedals, as well as running how signals from my overdrive pedals into this buffer has much or any audible effect at the various voltage levels.  I've used buffers before, but the consistency offered by what I think I'm hearing right now, with the ease of implementation means I just may be adding one at both ends of my chain, as well as at the end of most of the stuff I build in the future...I'm becoming a believer. 

[earthtonesaudio]

Without knowing exactly the specifics of your setup and the builds you've tried, the best advice I can give is this:
To maximize headroom, use higher voltage, make sure the device you use can swing equally in both directions (above and below the bias point), and use something that can swing rail-rail.  There are lots of CMOS op-amps available that swing to the rails, and many of them use very little power-a bonus!

[liquids]

...make sure the device you use can swing equally in both directions (above and below the bias point), and use something that can swing rail-rail.  There are lots of CMOS op-amps available that swing to the rails, and many of them use very little power-a bonus!

This is the first I've heard of this, actually.  I always imagined that all op amps could swing the full voltage I fed them...ah, to be an EE and not just a DIY hack  .

Anyhow, so I've got something new to research, but as I do, can you recommend a few examples of the above? 

Likewise, on a datasheet like this http://focus.ti.com/lit/ds/symlink/tl072.pdf for a TL072, how do I find how close to the rails it does or doesn't swing? I've seen a few that call it that outright Is it Vom?  How far from the rails do the typical op amps used in guitar effects(4558, TL072, 741 4559 etc) swing, generally speaking?

Last weekend the guy who works on my amps and I were shooting the breeze about circuits.  We were talking about bandwidth and such of the common, cheap-o, and often low-fi op amps used in effects.  He was mentioning that there are some high quality op-amps out there usable for mics (?) and high end studio mixer boards etc, and that they can be $40 a pop, which to me is inconceivable....maybe I can understand more of this now.

That being said, there's probably no reason to go that far just for a buffer if indeed a simple, small, cheap bjt transistor circuit will do the job 'transparently' at an easy to get 9v, but having options and learning more in the process is what this is all about for me.  Thanks for the info offered--more, please. 

[earthtonesaudio]

Yes, in the TL072 datasheet you linked to, the maximum output swing is listed as V_OM, but note that's at a certain supply voltage.  Check the graphs for a closer approximation.

If you go to Mouser or Digikey and just search for "rail-rail" you'll get a list of many op-amps that might be suitable for your needs.  Just be sure to check the datasheet before you order, many rail-rail op-amps are slow, or only operate at 5V, etc.  But there are ones that operate on higher (+/-15) voltages and are only a little more expensive than general purpose types ($2-$7 on average).

[Gus]

If you are liking BJT EFs.  A few hints for you

The "standard" two equal value resistors bias at +9VDc is not the best way to work at +9VDC.

Look in "The Art of Electronics"(TAOE) for the EF power amp used as an example of things people overlook sometimes

An "effect standard"  EF say 10K emitter R and base at 1/2 supply voltage.  Two weak points IMO first the E is .6VDC lower than 1/2 supply voltage, matters a bit at +9VDC not so much at higher supply voltages.  Two the transistor can "pull harder" to The +9 rail than the resistor "pulls down"  This matters more when what you are driving needs more current drive/lower input resistance.

IMO You want the emitter to be higher than 4.5VDC not lower it depends on the load after it for the most symmetrical drive the lower the input resistance the more you offset "up" to a point

Simple thing to do is offset the bias use a smaller value resistor base to +9 and a higher base to ground.  Remember (Hfe x emitter resistor) || top bias resistor || bottom bias resistor is the input resistance or in the case of a TS type EF the resistor from the Vref to the base || (emitter resistor x Hfe)

The higher the Hfe the higher the bias resistor values you can use

[liquids]

If you are liking BJT EFs.  A few hints for you
The "standard" two equal value resistors bias at +9VDc is not the best way to work at +9VDC.

An "effect standard"  EF say 10K emitter R and base at 1/2 supply voltage.  Two weak points IMO first the E is .6VDC lower than 1/2 supply voltage, matters a bit at +9VDC not so much at higher supply voltages.  Two the transistor can "pull harder" to The +9 rail than the resistor "pulls down"  This matters more when what you are driving needs more current drive/lower input resistance.

IMO You want the emitter to be higher than 4.5VDC not lower it depends on the load after it for the most symmetrical drive the lower the input resistance the more you offset "up" to a point

Simple thing to do is offset the bias use a smaller value resistor base to +9 and a higher base to ground. 

Great information that I had not though about, but seems simple enough. If i stick with ~9v, I can do the calculation depending on what input impedance I go for.  I was thinking through it for 2 ~5M resistors, for example, and realized I wouldn't have the proper resistor values, but then it dawned on me that I can, for example, just add the proper additional resistance needed in series with the 'lower' resistor going toward ground--especially, since space on a board for such a simple buffer won't be an issue.

Look in "The Art of Electronics"(TAOE) for the EF power amp used as an example of things people overlook sometimes

I have the book, and have been working through it...slowly     I'll have to look at that, thanks for the tip.

Remember (Hfe x emitter resistor) || top bias resistor || bottom bias resistor is the input resistance or in the case of a TS type EF the resistor from the Vref to the base || (emitter resistor x Hfe)
The higher the Hfe the higher the bias resistor values you can use

Some of what you said here has led me to other posts, etc., regarding calculations.  That being said, it's hard to understand what the above equations exactly are.  I.e., what does || represent? 

[R O Tiree]

In your shoes, I'd make 2 buffers - one for the start of your effects chain, and one for the end, both running off 18V.

The "Input" buffer should have a gain of less than 1, so the rest of your effects chain can cope. Say you go for a gain of 0.333, this might bring the voltage swing within reeach of your effects chain. At the end, the "Output" buffer should therefore have a gain of 3, so you restore the original signal strength to go into your amp. You might get away with 0.5 -> 2 or have to drop to 0.25 -> 4... that's a matter for experimentation.

Now you are happy, because you don't have to turn your guitar's vol pot down, your effects are happy, because they're getting a signal strength they can handle, and your amp is happy, because it's getting the right signal strength that you're used to.

[Gus]

|| means parallel. 

[Andi]

I would simply put a charge pump (regulated back down to about 15v) and an op-amp follower in a box.

This does not, of course, mean that this would be the right approach.

[liquids]

The suggestions here have been at least through provoking, many ahve been or will be implemented. Thanks for the help.   This has not only been mentally stimulating but very practical! That being said, more research and tinkering has lead to more questions...

1) I've looked at Jack's basic buffer page dozens of times.  Only now have I noticed that it says, with a JFET buffer, a resistor from gate to Vr (i.e. biasing a JFET like a BJT) would yield higher headroom on the JFET buffers.   I was under the impression that JFETs didn't benefit from a Vr since they are 'always on,' for lack of better terminology    Now I notice the Boss buffer and the TR-2 schematic I mentioned likewise use a resistor to Vr off the base, not ground.  Ah, the joys of being sophomoric.      In short, I'm now wondering why that makes a difference, and how much difference it makes...I guess I'll find out with my ears sometime soon, if nothing else.  I'll likely stick with BJTs, but now I've got something to go back to the drawing board about, and experiment with.  More Art of Electronics to read, for sure!

Back to talking BJT buffers. 

2) Gus: I tried to do the math and it seemed that a "ratio of ~5:7" resistance value would make sense at +9.4v (1spot power supply)--that is, in practical terms something like using the 'noiseless biasing' R.G. touts with ~10K and ~15K resistors, and then 1M from there to base would keep the BJT base ~1+ silicon diode drop more positive than half the supply voltage, as I had interpreted your comments to advise, along with yielding the standard input impedance.   

But now I've seen the input buffer used on the 3 transistor Fuzz and you (Gus) touting it elsewhere, which follows along the lines of your comments here.  In that buffer, you are using closer to more like a 1:2 ratio, which keeps the base about two silicon diodes more positive.  This may be splitting hairs and be a small difference, but is it?

3) Following that same previous post, surfing TAOE, p. 91 / section 2.15, I see a BJT EF power amp where the input is biased by a PNP BJT...which might suggest keeping the base 2-diode drops more positive than 1/2 power supply voltage.   Is this what you were hinting at for the concept or is there something else in TAOE you were referencing?

In your shoes, I'd make 2 buffers - one for the start of your effects chain, and one for the end...

These ideas were useful and appreciated.  I've determined to go with at least two buffers, actually, over just replacing the one as I originally was intending.   I've found my volume pedal, which I believed was buffered, is passive.  It always sees a buffered signal at it's input.  That being said, I've found it interfaces even better with the Lexicon Mx200 that follows it (has a 10k input impedance!) with a buffer after it as well, given the loading effect if the volume pedal isn't full tilt.  I had always suspected that was the case, but believing it was active, I didn't ever try it out.  Now I know it needs a stand alone buffer to follow it as well.

Along your lines of tweaking the 'input' buffer to have a gain less than 2, I do also plan on integrating a volume control after the "first buffer" to able to control signal strength--which is obviously a different thing than controlling things from the guitar's volume pot. Of course, a volume control after a buffer will need a buffer to follow it as well, to keep from loading down what follows it.      So now I'm talking three buffer stages, with a a volume pot in between two in the same box.  Whew! 

[Gus]

I used resistor values you could find at places like Radio Shack.  680K etc is often not found at place like Radio Shack

Page 91 of TAOE note +15VDC and -30VDC for a +-15VAC output swing

You can also use a diode in the "bottom" leg of the voltage divider to move it "up" a diode drop for the Vbe drop.