Modernize Old bypasses? (not convert to true-bypass)

[m_charles]

Hi, I have a few pedals that I really enjoy in my collection.
An old 10-series ibanez analog delay, an 80's boss CE-3 analog chorus, and a couple old boss phasers (all the bosses are the one's that use the "ACA" adapters).

They all sound great, but the bypasses are really bad on them. Not just a little shaved off the highs, but a noticable dampening of the clean tone when bypassed.
None of my newer boss, dod, etc. have this issue, the bypasses aren't perfect, but definitly acceptable.

Does anyone have any suggestions on how to clean up the buffers, or tweak them in some way to get them sounding good when bypassed?

BTW, I found a million posts about true-bypassing, trying to squeeze in a 3pdt, using a bypass loop etc but that's not what I'm wanting to do.
Just keep the pedal as is, not have to use a loop, but "modernize" the buffers.
Suggestions?
thanks!
Chuck
BTW, thanks again Pete for your help on my last post.

[DougH]

Are you sure that they have buffers? A lot of the "tone-suck" bypasses are just an SPDT with the pole tied to the output. The input Z of the effect is what causes the tone suck. In this particular case, a super-simple solution is to spider-wire or daughter-board (with a small piece of perf) a simple JFET buffer onto the input of the effect. Move the input connection from the input jack and bypass loop (that goes to the switch) to the buffer input. No need to change switches or anything. When the effect is bypassed, the buffer isolates the low Z of the input from the bypass signal and doesn't load it. I did this to a wah and it worked great.

If it does actually have a buffer, then we'll need to know more about the circuit. (I.e. we'll need a schematic.)

[m_charles]

Thanks Doug, but yes, all the old boss, most dod and most ibanez used buffered bypasses.
Thought it could be the old electrolytics but its to consistant across these older buffered bypasses to be that. Impedance?

[DougH]

I'm not that familiar with them and wouldn't hazard a guess without a schematic.

[Paul Marossy]

What kind of input buffer do these old Ibanez and Boss pedals use? An FET flip-flop should present a very high input impedance, so in theory should be very little loading of the signal and consequently loss of highs (AKA "tone sucking") when in bypass mode.

Is there a schematic you could show us for one of these old pedals of yours?

[rustypinto]

Link to the CE-3:

http://www.synthdiy.com/files/2007/bossce-3.jpg

This would even be tricky to implement true bypass to (though i know this isn't your intent). There is a slew of FETs used for the switching involved for the mono vs. stereo mode. The input buffer is also a FET, but i wouldn't necessarily blame the tone suck on this. When in "normal" mode, this goes through each half of IC1 which has a good bit of tone shaping involved (eg. the output half of IC1 has a pole at fc = ~2.4kHz, which will roll off treble harmonics). I see this on a lot of Boss units. Its often done on BBD based units with a lot of inherent clock noise. The channel resistance of each FET used in the switching doesn't help either, but this is hard to correct without dropping a better device in place (which takes a careful selection).

[earthtonesaudio]

I have a suggestion which may or may not work.  Based on the notion that each FET switch has some capacitance between the channel and gate, the any resistance leading up to the switch will form a low-pass filter.  Less resistance allows more highs through before attenuating.

So try replacing resistors 8, 15, 36, 17, 28, and 34 (and perhaps also 29 and 31 if output B sounds funny) with maybe 10k resistors.


Note:
If you're very careful you can try this temporarily without rewiring anything, by sticking the leads of 1/8W resistors in next to the existing resistor leads (adding the temporary resistors in parallel to decrease the total resistance at that section).  Friction will hold them in place.  If it doesn't work, it's easy to pluck them all back out again.

[Paul Marossy]

Link to the CE-3:

http://www.synthdiy.com/files/2007/bossce-3.jpg

This would even be tricky to implement true bypass to (though i know this isn't your intent). There is a slew of FETs used for the switching involved for the mono vs. stereo mode. The input buffer is also a FET, but i wouldn't necessarily blame the tone suck on this. When in "normal" mode, this goes through each half of IC1 which has a good bit of tone shaping involved (eg. the output half of IC1 has a pole at fc = ~2.4kHz, which will roll off treble harmonics). I see this on a lot of Boss units. Its often done on BBD based units with a lot of inherent clock noise. The channel resistance of each FET used in the switching doesn't help either, but this is hard to correct without dropping a better device in place (which takes a careful selection).

With that great explanation of the tone shaping at IC-1, I can understand what the problem probably is. I don't know that there is a real good solution because of the circuit design and the devices involved.

I know, it's not the aim of this post, but true bypassing would get rid of all those problems you have in bypass mode. But the other problem with true bypassing things is that sometimes the effected level vs. bypass level can be enough to cause a different set of problems. That, of course, depends on the circuit.

I have a suggestion which may or may not work.  Based on the notion that each FET switch has some capacitance between the channel and gate, the any resistance leading up to the switch will form a low-pass filter.  Less resistance allows more highs through before attenuating.

So try replacing resistors 8, 15, 36, 17, 28, and 34 (and perhaps also 29 and 31 if output B sounds funny) with maybe 10k resistors.

That's not a bad idea. It might just work. At least enough to maybe make it acceptable.

[rustypinto]

Agreed, all.

If you were to pull R7, and even R21, you would probably get a lot of treble back. Plus its easy enough to put those back if you don't clip them.

[m_charles]

Wow, thanks for jumping on this guys!
Yes, seems to be something about these old 80's buffers that aren't too nice to a guitar's tone. Especially a strat most-often on the bridge pup.

I may give the parallel 10k thing a try, but from experience, I almost always end-up wishing I would have left things alone when it comes to cracking open old bosses, ibans, etc..
Would "rebuffering" as doug had mentioned be a possible solution? I could maybe fit a simple TLO71 buffer in before or after.

I'm assuming AFTER would be the place. Right?
Or would this not really accomplish what I'm after. I'm still a little hazy on buffers. I know they can be great, and necessary, just still getting my head around the science of 'em.
chuck
BTW, a very under-rated analog chorus that can be had often for under $50. They don't get really swirly like its famous predecessor, but for adding a nice clean "throb" to certain things and not masking the sound of your guitar(except when bypassed!), they're a great deal. Just don't waste your time comparing it to a ce-2. Totally different animal.

Which brings me to another question (in which I'll contradict my prior statement about cracking open these pedals).
Think there's an easy way to widen the "rate" span of the ce-3? Just get it pumping a bit quicker?

[DougH]

Agreed, all.

If you were to pull R7, and even R21, you would probably get a lot of treble back. Plus its easy enough to put those back if you don't clip them.

I would pull c4 and c5 and if the clock is too noisy put one or both of them back. May not be able to correct this without an external bypass box or true bypass mod.

Why they send the bypass signal through that IC is beyond me...

[aziltz]

If you splice in buffers, put them before.  The idea is to increase the impedance presented to the guitar at the input.  Placing it after the problem sections wont change the fact that the guitar is connected to the problem sections first.

[Thomeeque]

So try replacing resistors 8, 15, 36, 17, 28, and 34 (and perhaps also 29 and 31 if output B sounds funny) with maybe 10k resistors.

Lowering R8 and R34 will unbalance pre/de-emphasis mechanism (and will lead to crippled in/out AC characteristic) - R7&R33 would have to be lowered and C6&C19 enlarged in the same ratio as R8&R34 to keep it tuned correctly.

Why they send the bypass signal through that IC is beyond me...

R_ON resistance of switching FETs is maybe too high to switch signal directly to output..

Which brings me to another question (in which I'll contradict my prior statement about cracking open these pedals).
Think there's an easy way to widen the "rate" span of the ce-3? Just get it pumping a bit quicker?

Try to lower R10 (let's say to 4k7), this will widen the range (by allowing to dial lower speeds) and than use smaller (47n) capacitor for C28, this will shift whole range up (or maybe halving of R52 will have same effect, but I'm not sure now - probably yes), so you will have same lowest speed but doubled maximal speed.

T.

[m_charles]

Thanks Thom, you rule! Well I guess I'm gonna put on the surgery gloves. Always wanted to try this on the ce-3. Wish me luck, I'll tell ya how it goes!

[Guitar_Ninja]

Probably easier to just use a true bypass strip.

[ode2no1]

why not build a buffer and just stick in first in your chain? or even something like a SHO so that you can add a little sparkle to things and bump up the volume a notch.

[earthtonesaudio]

why not build a buffer and just stick in first in your chain? or even something like a SHO so that you can add a little sparkle to things and bump up the volume a notch.

Because the treble loss is coming from inside the pedal, not the cables leading up to it.  Boosting the treble might come at the expense of more noise and unwanted distortion.