Query: Pedals with buffer *OR* TB option

[Mark Hammer]

We've seen a number of debates regarding the pros and cons of true bypass switching vs other switching schemes that leave a buffer on the input and only provide switching of the output or some other non-true-bypass arrangement.  Folks with considerable experience like Pete Cornish have expressed some fairly strong opinions in favour of buffering, while others like Mike Fuller have made such an industry of true bypass that it has practically attained the status of brand, or as some folks have described it - a product feature.

Musing on the bus this afternoon, it occurred to me that this needn't be an either/or choice.  For instance, let's say the input jack is wired directly to the board, and so are all contacts of a 3PDT switch.  The board also contains a jumper bank.  The jumper bank allows you to feed the input signal directly to a buffer and use one set of switch contacts to select between the buffer only and the output of the effect.  Jumpered another way, the switch uses one set of contacts to route the input jack to output or input buffer, and another set of contacts to take the output from either the input jack or the effect circuit output.  By tradition, of course, the thrid set of contacts is used to activate a status LED.

There are advantages and disadvantages to each switching method.  This would let the user have a choice of whatever works best for them.  Thankfully, the required 6-pin header takes up precious little space, and if you want jumpers, you need look no farther than the mobo of that 386 machine sitting in the corner of the basement right about now.  I suppose one could also use a small bank of dipswitches too, which provides the advantage of not having any parts to lose.

Now I'm wondering if anyone has ever done something like this.

[davebungo]

I can't see the point myself.  If you have a sequence of effects as long as your arm and a 30m cable to drive, then you will probably need a buffer at the start of the chain to minimise signal degradation, so why not just plug one in (or if you can afford it have Pete Cornish do it for you)?  There are tons of these around and they are pretty cheap, plus you don't have to worry about what is first in your chain, I mean would you provide this mod in every FX box type just in case it happened to be the first in the chain?

On the other hand, if, like me, you just have one or two pedals running off PP3s and a 3m cable to the amp input, then it is probably not worth the extra hassle of using a buffer.

Just my penny's worth :wink:

[phillip]

I prefer True Bypass myself.  Some effects (like the Fuzz Face and Fuzz Face-related circuits) really don't like to have a Boss pedal in front of them in the signal chain...it drives the input harder and makes the Fuzz Face sound more like a heavy metal pedal.  I think that as long as the signal chain doesn't have more than 3 or 4 pedals, you're fine...I only have about 3 at the most in my signal chain.

Phillip

[cd]

2 replies and both re-hashed the same old stuff without bothering to answer the question.  Read the subject line guys!!

Mark, I've thought about the same thing myself, complete with the DIP switch idea.  Even though it would be trivial to implement, practically speaking, I'm doing one-offs so I always go the either/or route.  But then again, I use EMGs and/or preamps in my guitars so buffering, cable runs, etc. is never a problem.

Come to think of it, this is already a feature in a number of FX, most notably the Line 6 4-switch modellers (some of them) and the Peterson Strobostomp.

[TheBigMan]

I've built an AB box that I use as a tuner mute, and it has a FET buffer like the Tillman one in it, bypassable by a DPDT slider switch on the back.

My first active bypass pedal is a Boss CS-2, but if I use an Ibanez CP-9 or my soon to be built Orange Squeezer then I lose a bit of high-end because of the lower input impedance of the CP-9 or of the TS-9 which is next in the chain.

[phillip]

2 replies and both re-hashed the same old stuff without bothering to answer the question.  Read the subject line guys!!

I did read the subject line, and I was simply stating my preference for straight-forward true bypass.  

Dealing with the Hammond 1590B box, real estate is at a premium and true bypass only pedal has fewer on-board parts so that the circuit board doesn't consume valuable space inside the box.

Phillip

[joegagan]

If you don't mind plugging your cord only part way in, you can achieve either with minimal fuss
[/img]

[davebungo]

2 replies and both re-hashed the same old stuff without bothering to answer the question.  Read the subject line guys!!

I read the subject line.
I read the post.
I posted my response which basically stated that I could not see the point of fitting this to a pedal.  Try reading my post again.  Just because I don't agree with Mark does not mean to say I'm rehashing the same old stuff. and I'm certainly not being disrespectful:evil:

[Processaurus]

I think Mark's idea would work well for effects where an alternative bypass  is available and possibly desirable, like a tremolo or vibrato where the footswitch gently stops the LFO (like boss VB-2), or a delay that finishes the delay trail after being bypassed (like line 6), or a boost that has a tone you like where the bypass could just set it to unity gain.  It would be nice sometimes to have options like this as well as a bullet proof, no nonsense true bypass.

[amz-fx]

Folks with considerable experience like Pete Cornish have expressed some fairly strong opinions in favour of buffering, while others like Mike Fuller have made such an industry of true bypass

I actually did some testing and reported the results for everyone to see:

http://www.muzique.com/lab/truebypass.htm

My thoughts about it are:  it really makes no difference as long as you do not have a vintage pedal wired in non-true-bypass fashion on your board.  If you use all true-bypass it will be okay, or any mix of true-bypass and buffered, or even all buffered.  The only caveat is that a large number of buffered pedals in series may accumulate some noise --  some Boss pedals have as many as three buffers left in the signal path even when bypassed!  Most of the time it's not a problem though.

regards, Jack

[Fret Wire]

Now I'm wondering if anyone has ever done something like this.

Roger Mayer's Vision Series has this description on his site.

"Vision Octavia Features

MULTI-MODE OUTPUTS with HARD WIRE BYPASS  or TWO BUFFERED OUTPUTS
The Vision Octavia has three outputs, a true hard wire output plus two buffered outputs that can drive long cable lengths with no high frequency loss. These two buffered outputs are identical and are disconnected when the hard wire output is being used. This means that you have a choice for all possible performance or equipment situations whether you feel a true hard wire output or buffered outputs will perform best. In general the low output impedance of the buffer gives a lower noise floor and a more punchy sound but now it's so easy to hear the difference and make a real decision that is based on fact not hype.

Technical Description:
MODE 1: There is no jack plug inserted into the Hard Wire Output ( HW OUT1 ).
The input of the buffer is connected through switching contacts on this jack socket (HW OUT1) to the foot switching output ( "full bypass" or "effect" ) and the two buffer outputs ( BF OUT2 ) and ( BF OUT3 ) will thus follow the foot switching action. In this "bypass mode" however the input of the buffer is only connected to the Input jack socket of the unit and not to the Octavia circuit as well so it can be thought of as "full bypass buffered".

MODE 2: When a jack is inserted into the Hard Wire Output (HW OUT1).
The input of the buffer is disconnected and no output signal will appear at the two buffered outputs. The Hard Wire Output ( HW OUT1 ) will follow the foot switching output ( "full bypass" or "effect" ). In this "bypass mode" the Hard Wire Output ( HW OUT1 ) is directly connected to the Input jack socket and to nothing else in the unit and is in "full bypass" - "true bypass" or "hard wire bypass" mode as the input is directly connected to the output through the foot switch and to nothing else."

http://www.roger-mayer.co.uk/visionseries.htm

[Ben N]

It seems to me that Jack's article implies that the only real benefit of buffering within a mechanical-switched TBP effects chain is at the output of high Zout effects pedals when the effects are engaged.  I.e., the buffer should be bypassed when the effect is.  For example, the purpose of putting an output buffer in a wah is to make sure the wah effect is not lost due to interaction with a following pedal.  If the wah is TBP, there is no reason to keep that buffer in the signal chain when the pedal is bypassed.  Assuming that there is a hi-Z-in lo-Z-out buffer at the top of the chain, and the outputs of all the effects are lo-Z, there is no need for input buffers along the way at all.  The only reason commercial pedals have input and output buffers that are non-bypassable is that their FET switching requires it.

Mark's idea would seem to have more benefit for commercial production--like the RM, meeting the varying needs of different users--than for DIY, where the builder knows how he/she intends to use it. There may be exceptions, of course, like a booster that may be used in different circumstances, and for which a jumper/DIP/slide/jack switch to select buffered/direct bypass might be useful.

JM2C, of course,
Ben

[NaBo]

I was gonna chime in from the same pov Ben

For DIYers, it's easier just to whip up a tiny little buffer stomper or something and have total control over what gets buffered when and where.  But for a "normy", it'd be pretty nice to have control over that.

And well, most DIYers don't form the bulk of the commercial fx world's clientelle  :wink:

It'd be a nice feature for pedal-makers who wanna make sure their product is always smiled upon by TBPers and bufferers alike.

[MartyMart]

Now I'm wondering if anyone has ever done something like this.

Roger Mayer's Vision Series has this description on his site.

http://www.roger-mayer.co.uk/visionseries.htm

Yup, the VooDooVibe Jr has this feature, one hard wired and two buffered
outputs.
But ....... its the size of a house brick !!
If size on the pedal board isnt a problem then its a great idea, for most
here I agree that a small DIY "buffer" that can be inserted for the whole board makes sense.
Again it depends on how many pedals/how much wire.......... :roll:

Marty.

[Paul Perry (Frostwave)]

You could have TWO stompswitches, one switches a (buffered) effect on & off, the other just bypasses (or not) the whole effect in a true bypass way.
Then, you have three useful possibilities:
1. effect on, with buffer
2. effect truly bypassed
3. effect off, but buffered....... maybe a useful tone sucker?

of course, any good effect should not only be the size of a brick, but weaigh as much too

[Mark Hammer]

Interesting comments from all.  From where I stand, they would seem to reinforce the notion that a pedal that provides a choice (and isn't the "size of a housebrick") might be an attractive option, since clearly each option seems pretty damn important to *somebody*.

The possibility of "buffer buildup" in a pedal-board with half a dozen or more pedals does present the possibility of cumulative noise as is suggested, without any real benefit to signal bandwidth, tone-sucking prevention, etc.  At the same time, the builder may not be able to assure for all time and all contexts that pedal X will always be preceded or followed by a buffered or unbuffered pedal, so being able to keep buffer-buildup to a minimum, or insert where necessary seems like a reasonable idea.  One of my own reasons for liking TB is the capacity to kill the whole box when a battery craps out without having to rewire or forfeit signal flow.  With a jumper arangement, you could guarantee that possibility if it was that important to you.  And of course, with the convenience of 3PDT switches, one is not obliged to resort to FET-based switching, OR to forfeit indicator LEDs.

Certainly there are going to be cases where space within a 1590B does not permit, but there are plenty of instances where there would be more than enough room on a sensibly designed PCB to accommodate a buffer and jumper choice-system.  After all, you generally aren't doing anything with the space just to the left and right of the stompswitch, right?  A header/jumper bank could easily fit on one side and a single transistor buffer could easily fit on the other.  PC mount your jacks at the far end of the board (along the rear skirt), and you still have plenty of space for a battery and a wide array of circuits, without having to resort to miniaturization.  And obviously a 1590BB would permit even more degrees of freedom.

The recent Dave Hunter guitar pedal book contains a CD with some 92 sound samples, 4 of which are clean signals via several different switching arrangements, including TB and buffered.  The buffered "wins" hands down in the samples, but it is also not very clear to what extent the testing/samping conditions favoured buffers or not.

Still, the thread is not about the merits of one over the other (TB vs buffered), but about the merits of having a choice vs no choice.

[Satch12879]

The only reason commercial pedals have input and output buffers that are non-bypassable is that their FET switching requires it.

I don't think the switching has ANYTHING to do with their inclusion at all.  Craid Anderton's electronic switch doesn't require additional input+output buffers.  Personally, I just think the Boss designers know what they are doing...

But anyway, I've been musing two somewhat related switching schemes:

1) Input jack -->buffer amplifier -->switch -->effect input -->effect output -->switch -->output jack

The input signal always hits the buffer amp.  Using a 3PDT switch, you do your complete bypassing of the effect AFTER the input buffer stage.  When you have the effect taken out, you still have the buffer driving the clean line to the output jack.

2) The original "tone sucking" came about because designers bypassed the outputs of their effects using a DPDT so they could get a status LED.  Reasoning stated that because the effect circuitry was still connected to the input of the effect even when bypassed you'd still get losses because of all the components creating sort of a loading diversion.  (But I always thought electrons took the path of least resistance so how could you have some sort of parallel resistance or capacitance scenario when the output of the effect circuit wasn't connected to anything when bypassed...  OK, we'll go with it for now.)  Why not move the bypass and switch the inputs of the effect? You'll get to use your relatively cheap DPDTs and status LED.  Or will you have a "diversion" on the output similar to the situation encountered in the usual scheme?

Just a couple of thoughts.

[Mark Hammer]

My interpretation of "switch history" is a little different than yours, Satch.  Not dramatically, but different.

It was entirely common during the 60's and well into the end of the 70's for FX to use a SPDT stompswitch.  This would switch the signal seen by the output jack from whatever was coming off the circuit to whatever was coming in the input jack.  That was ALL the switch did.  No status LED, just your ears.  E-H did it.  MXR did it.  They pretty much ALL did it.  

In most instances, the input impedance of the circuit was not terribly high.  That may not have been a problem when the effect was engaged (i.e., the output jack was tied to the output of the circuit board), but when the effect was "bypassed", the circuit was still tied to the input jack and loaded down the "clean" signal.  Stick two or three such pedals in series, and that was a LOT of loading.  Tone wasn't just sucked, it was sucked on an industrial level.  Of course, it doesn't take much to realize that few players at that time had pedal boards as extensive as they do now.  Hendrix generally had 3 pedals at his peak and that was treated as a veritable "jungle" of effects and patch cables.  I had 4 or 5 in 1978 and my bandmates thought I was an "effects wizard".  Most folks had less.

Generally, the loading had enough of an impact on bypass level/tone, that many pedals were designed in anticipation of that loading so as to achieve effect/bypass level balance, or come close to it.  Indeed, many here can probably attest to problems they had with loss of such balance when they thought they could "improve" a vintage pedal by adding a DPDT or even 3PDT switch to replace the original SPDT.

When solid-state FET-based switching was introduced, it became necessary to precede the FET leading to the "clean" path with a buffer.  In some instances, such as chorus, phaser, flanger, and delay pedals, all that was necessary to make the effect go away was a single FET that lifted the wet/effect signal at the mixing stage, leaving only the dry signal.  In these instances, the clean signal was going to need to be tone shaped a little to reduce the noise from the effect, so the input buffer remained in place.  If you look closely at the input and output stages of something like the Boss CE-2, you will see a complementary precomp/decomp circuit built into the input and output stages which gives the high end a little nudge upwards on the input, and a slight level nudge downwards by the same amount on the output.  Since that nudge downwards also attenuates any noise acquired in the BBD chip, the whole thing works like a poor man's Dolby noise reduction system.  The important thing to note, though, is that when the input signal is split into a dry and wet half (with wet sent to the BBD), the split usually happens immediately after the pre-emphasis circuit.  So, if you want to get your original tone restored, you have to always run the clean signal back through the complementary de-emphasis circuit in the output buffer stage.  Two buffers on, ALL the time, rain or shine.  The FET switching arrangement desn't necessarily *require* those precomp/decomp circuits, but the precomp/decomp arrangement assumes that the FET will be where it is.

No one, as far as I know, utilized FET switching *before* and *after* the entire effect circuit.  It has always been embedded within the circuit with input and output buffers being what lay on the other side.

For whatever reasons, CMOS switching is somewhat different, and it is possibly to stick a 4066 or 4016 or 4053 "around" an effect circuit.  Designers prefer the flip-flop, FET, momentary-switch thing though.  Perhaps it is cheaper and less complex or current-demanding than the CMOS switch.  I know the current estimates for the Anderton ones are easily 6-8 times the anticipated current demand for the BOSS or DOD type.

[bwanasonic]

For me the most convenient way of dealing with the either/or is a seperate TBP loop box. My main preference for a TBP option is for cases of undesirable pedal interaction with vintage style effects. I also like a relatively (for me) uncluttered signal path for clean playing, for maximum sparkle. I can't say I have found the buffers in Boss / Ibanez effects to be incredibly useful on their own either. But all that aside, in answer to the original question, the more options the better! If I was in the market for a commercial pedal, I would consider the added flexibilty of a switchable bypass scheme a plus.

Kerry M

[Ben N]

Thanks. Mark--that is indeed what I had in mind, re: fet switch and buffers.  And, BTW, Happy P-over.

Ben