A better treble bleed circuit

[JohnH]

Hi moose212- Thanks for posting that. That Fender link is the clearest description Ive seen about what they are doing with their recent 'Tone Saver' circuits. Everything else I read was covered in marketing verbiage and epoxy.

So, the seperate encapsulated Tone Saver gadget is indeed a resistor and cap in parallel - which I agree is the best format for those two parts.

I hadnt seen the pro-series version before, which as shown on that Tele. Evidently it's a 1.2nF cap and 150k parallel resistor, all in series with 20k.

Guitarfreak was written originally to test TB circuits and can do quite a few variants. But I hadnt thought of that one and so it's missing one resistor component needed in GF versions to date. But it wouldnt be hard to add it, and I might do that since presumably there are now lots of guitars being sold with that design.

I did run a quick test in 5Spice with the new 3-part design. I didnt see anything very helpful happening though and the new part seemed to slightly reduce the effectiveness at moderate high volumes without doing much at lower volumes. But its hard to accuatately compare that way. Once the design is in GuitarFreak a full comparison across frequencies and volume levels will be possible.

[Yogi B]

Huh, so this arrangement is one I thought of when originally coming across this thread, (plus the marginally different arrangement of the series treble bleed with a parallel resistor).

Jul 17, 2017 16:57:21 GMT -5 JohnH said:

Evidently it's a 1.2nF cap and 150k parallel resistor, all in series with 20k.

If I remember correctly, I came to the decision that 820p, 330k & 22k were about the right values for me, with 500k pots & PAFs -- as that was the original context of this thread.

I can't confirm those exact values, as I abandoned the idea after real world testing proved that it wasn't a noteworthy alternative to the others, thus I 'reset' to my previous favourite treble bleed configuration.

Plus, I think any SPICE file I happened to have pertaining to this may now be lost to the aether, with a bunch of other old schematics that I deemed weren't worth saving after an effort to organise and update my personal libraries (since the update to LTspice XVII) left most of my files in varying amounts of disorder.

[moose212]

Hi John - thanks, and sorry for my delay.  I missed the new page (kept refreshing page 3...).  My intuition was that the 20k was there to reduce any bleed you'd get through the capacitor if you wind the volume all the way off.  I also wondered if it might serve some helpful purpose in cases where you were using treble bleed mods on multiple volume pots (e.g., each humbucker has it's own dedicated volume knob), but I haven't thought through the circuit implications to know if that makes any sense.  Selfishly, that's my current case - I've got 4 volume pots (sliders, not knobs) as sort of a mixing-board--type set up for 4 pickups (coil-tapped humbucker, single coil, and p-90).


If you do modify guitar freak I'd love to hear what results you get.  Thanks again for this thread

[Weshuggah]

Nov 12, 2010 21:22:59 GMT -5 JohnH said:

Improved treble bleed - Type D, Wiring diagram

Meanwhile. here is the wiring for my new arrangement using a dual-gang pot.

Hello, this is very interesting... I tried this setup (diagram above) but Oops! ..when Volume is at zero, I still hear the sound. I checked ground wiring in all the guitar and it seems ok. Where could it come from?

[reTrEaD]

Hello Weshuggah. Is that the plural form of Meshuggah? Are there more than one of you? lol

Anywho ... Welcome to GuitarNutz2!

Jul 29, 2017 9:33:22 GMT -5 Weshuggah said:

..when Volume is at zero, I still hear the sound. I checked ground wiring in all the guitar and it seems ok. Where could it come from?

I could imagine a couple of reasons for a small amount of sound bleeding through when the volume control is at zero.

1 - The problem might exist in the volume pot itself. Perhaps when you reach the rotational stop, the wiper isn't fully past the resistive element and making direct contact with the metal of the CCW lug.

2 - There is appreciable resistance in the wiring between the CCW lug of the pot and ground.

I'd suggest using an ohmmeter and measuring the resistance from tip to sleeve of a known-good cable plugged into your guitar. The resistance will decrease from several thousand ohms when the volume is at 10 to nearly zero when the volume is at 0. Reduce the scale of the ohmmeter to the lowest possible while still getting a valid reading. A reading of an ohm or two is quite normal. Tens or hundreds of ohms indicates a problem.

Post the resistance readings and we'll continue from there.

[Weshuggah]

Thank you hehe basically a mix with Wes (Montgomery or Borland, whatever) with those swedish sugars!

hmm... so I get 13k at level 10, and 106k at level 0. It also goes to 150k around level 6-7.

I guess it doesn't smell really nice 

[reTrEaD]

Yes, it smells like a problem.

13k seems just fine. That would be the resistance of your pickup(s) in parallel with the resistive element of the pots. So basically the resistance of the pickups.

105k somewhere in the middle of the rotation doesn't seem unusual at all.

The 106k at minimum is alarming. Even if you misread that and it was actually 106 ohms, that's too much.

Measure the resistance from the sleeve of the output jack to the CCW lug of the volume pot.
(The lug with the blue wire marked ground in the drawing)

Also, measure from the CCW lug of the volume pot to the wiper of the volume pot when the volume pot is rotated to zero.

Post those numbers and we should be able to determine where your problem exists.

[JohnH]

Hello and thanks for your interest in this circuit. It should definitely go to zero or within a very few ohms at 0 on the knob. See how ground and hot output are on the inner pot half, so they just have to get connected together at 0 position. It suggests your overall grounds may have an issue. That 106k reading at 0 is very consistent with a 150k resistor being in parallel with a 500k pot (one that like most, is a bit lower really), as drawn, but without the connection to ground working.

Just so you know, this design was interesting to work out, and was very slightly better than a simple parallel treble bleed (150k and 1nF), but not very much. I never built it myself and I don't know who else has either. But it should work!

[Weshuggah]

Ok, I did the new measurements you asked and I read 0 ohm for both. And it was really 106K at minimum (from tip to sleeve on the jack), even 150K around 6-7. I use a 120K resistor + 1nf cap.
It's weird cause my overall grounding seemed nice, I don't get noises at all or anything...

How to know what part of the pot is the one above or below on the diagram? Not sure it has an importance..

[JohnH]

Your component values should work quite well. the diagram was intended to be a view looking from the back, knob on far side, as you wire it up. But its supposed to be a standard 500k dual-ganged pot, so it should make no difference which pot half is the inner or the outer one.

[reTrEaD]

Jul 29, 2017 17:57:50 GMT -5 Weshuggah said:

Ok, I did the new measurements you asked and I read 0 ohm for both. And it was really 106K at minimum (from tip to sleeve on the jack)

Something doesn't add up here. The wiper of the pot should be directly connected to the tip of the output jack. So the sum of the resistance from sleeve to the CCW lug of the volume pot *plus* the resistance of the CCW to the wiper, should be exactly the same as tip to sleeve of the output jack.

[monkey]

I've tried option C in my LP copy, which has Duncan designed buckers. While it does preserve treble, it also has an undesired side effect: the volume doesn't audibly roll back AT ALL until you get to about 1 on the dial, at which point it drops off a very steep cliff. Thus it seems impossible to actually user it as a volume any more, so clearly it doesn't work for me. In fact it makes the situation worse because the volume is rendered useless.

From what I've seen on here, I'm not convinced any of the other options really deal with this issue either.

So, does anyone have any practical experience of this issue and how to deal with it?

For example, I see the Kinman mod isn't discussed here, i.e. where cap and resistor are in series. What would the graph look like for this?

Or maybe I need to play around with cap and resistor values? Any advice (preferably based on experience of the same issue) would be greatly appreciated.

By the way, this is still possibly the most informative thread I've found on the subject. Good work!

[JohnH]

Thanks for your comments.

The Kinman version is plotted on page 1, date 13th November 2010.

Option C, (cap and resistor in parallel) works just fine. Its on all my guitars. And on every new American Fender Stratocaster too (at least the Pro and Elite series' as far as I know). The first statement convinces me more than the second!

It needs to go on a normal log taper pot, My base values are: a 500k pot should get 150k and 1nF, a 250k pot should have 120k and 1nF. A values or two either way may sound better to you. With that, the taper is slowed down a bit (nicely) but not to the extent that you have noted. Maybe your resistor is too low or your pot is a linear one?

[Yogi B]

Dammit, beaten to the punch...

---

In addition to what John said you could have potentially used a considerably larger capacitor (e.g. 100nF), but in that case your volume control would be acting more like a tone control (although cutting both treble and bass).

As John said the Kinman mod is discussed, here's a direct link to the post.


Finally I expect newey will be along soon to give to a formal welcoming, but until then: hello & welcome!

[reTrEaD]

Sept 3, 2017 1:24:27 GMT -5 monkey said:

it also has an undesired side effect: the volume doesn't audibly roll back AT ALL until you get to about 1 on the dial, at which point it drops off a very steep cliff. Thus it seems impossible to actually user it as a volume any more, so clearly it doesn't work for me. In fact it makes the situation worse because the volume is rendered useless.

Let's try to imagine what might cause such a severe issue.

1 - If you replaced your volume control and are using the wrong taper (linear or reverse-audio) that would push all the 'action' toward the counter-clockwise end of the rotation. More so with reverse-audio than with linear. But if that's the case, you would have the same effect even without the parallel treble bleed network.

2 - If the resistor you're using in the parallel treble bleed network is extremely low resistance (like maybe 150 ohms or 1500 ohms instead of 150k, that would also produce symptoms similar to what you've described.

3 - A defective capacitor (shorted) or a wiring error that has the wiper connected to the clockwise lug will cause symptoms that fit your description.

How to determine the exact problem?

If you haven't changed the pot, we can eliminate that.

In any case, the first step would be remove the resistor and capacitor.
The volume control should function in a normal manner.
If not, the only culprits would be a wiring error or a pot with the wrong taper.

If the volume control works normally without the treble bleed network, determine whether the resistor or the cap are causing the problem by putting just one in the circuit at a time.

[JohnH]

Fender Treble-bleed circuits

In the last year or two, Fender have started to offer variants of a treble-bleed circuit on some of their US guitars, particularly the Pro series Stratocasters which replaced the previous American Standard. Also, they have offered encapsulated and rather expensive add-on TB devices that they call 'Tone Saver'.

The 'Tone-Saver' is a little mysterious, but in the stock Pro series, the details are all clear, and are shown on the schematics. Here are the pdfs of SSS and HSS from this range:

www.fmicassets.com/Damroot/Original/10001/Fender%20Am%20Pro%20Stratocaster%20011301XXXX%20SM%20REV%20A%2010-17-2016.pdf

www.fmicassets.com/Damroot/Original/10001/Fender%20Am%20Pro%20Stratocaster%20HSS%20011304XXXX%20SM%20REV%20A%2010-17-2016.pdf

I suspect the 'Tone-Saver' as a separate component is also based on these.

So what we see is two designs. For single coils with a 250k pot, the design is a 1.2nF capacitor and 150k resistor in parallel, all in series with another 20k. For humbuckers with a 500k pot, they use a 270k and 1nF in parallel with no series resistor.

This clip from the HSS diagram shows both, cleverly built around a dual-ganged 250k/500k pot. The SSS and HH Pro Strats use the same treble bleed designs.



Testing

I wanted to test these designs using GuitarFreak 6.5, as a way to test-drive this new version, which had an extra component added so that it could deal with the Fender circuit. From previous testing, I had settled on a TB design for 250k pots comprising a 120k in parallel with either a 1nF or 0.82nF cap, or 150k/1nF with 500k pots.

Criteria

Whether one needs treble bleed, and what criteria constitutes an optimum design target is a matter of opinion. My view is to try to maintain the tonal balance of the full volume tone, across a range of reduced volumes, across a range of frequencies up to about 4khz. This is best visualized on plots of frequency response where all traces at low volume are 'normalised' to the same level, then noting the greatest differences of the traces above and below that of the reference full-volume trace. Adding the magnitudes of these '+' and '-' differences together gives a resulting range. On this basis, I think 3db range is a very good result, which can be thought of in terms of every frequency being within +/- 1.5 db from the reference, in relative terms.

With all TB schemes, there is a tendency to an initial fall off of treble in the first -2db of volume reduction (but less than with no TB), then a restoration of treble at further volume reduction and finally an slight excess of treble at very low volume. Picking the right design is a balance of compromises based on preference and the intended use of the guitar. The resistor(s) help to control the effect of eth capacitor, though they do effectively change the perceived pot taper.

The plots below show the full volume reference as a dashed line, with the low point of treble loss at small volume reductions (typically about -2db or 8.5 on the knob), then the trace at -15db reduction. All other traces for intermediate knob positions within this range are between these bounds.

Note that the plots are based on envelopes of response to include string harmonics and picking position.


Single coil design, comparison

The pickup chosen for this analysis was a Fat50 in the bridge position, modelled in GF6.5 based on tests. the Volume pot is 250k and the tone pot was no-load so out of circuit.

All that being set, I compared the Fender TB design for single coils, to simple parallel TB with 120k and 1nf and 0.82nF.

First a plot with no treble bleed, showing how treble rolls off as you turn down volume:




120k/1nF

This is optimised for volume change down to about -12 db, and rises a little in high treble at -15 db:



120k/0.82nF

This does slightly better in the low volume range, at a small cost of deviation at higher volume:





Fender 150k//1nF + 20k

Here is the 3-part Fender design for use with singles and 250k pots:




The range of tonal deviation in the above three plots are 3.3db (120k/1nF), 3.0db(120k/0.82nF) and 4.0db(Fender) respectively 

So the Fender design doesn't seem to be quite as close to the reference as either of the above two-part designs, within this volume range.

So I wonder why there is the extra resistor to make three parts, when it doesn't seem to give a better result than with the just basic two component designs? It could be that it was designed using a different set of criteria, or perceived opinion about what sounds best. We dont know if it was designed with support of calculation, or just by ear, which after all, is the most important test. Anyway, it's interesting, but if any Fender user finds that it gets a bit too bright at low volume, know we know why.

There is one other factor though. All of these designs change the taper of the pot, making volume change less sudden as you turn down. The Fender 3-part design makes less of a change than the others. If you consider a knob that goes from 0-10 max, with 5 being mid-turn where the pot is at 10% resistance, then the knob setting at -15db is at about 3.5 for both simple parallel designs and about 4 for the Fender TB circuit. These are in comparison to a setting of about 5.5 with no TB, and -15db reduction.


Humbucker treble bleed - comparison

For humbuckers, Fender use 270k in parallel with 1nF. with a 500k pot. This definitely gets brighter at reduced volume, compared to what I found to be optimum being 150k/1nF. Here are the plots, based around analysis of Pearly Gates in the bridge position, using 500k volume and tone pots.

Here is a plot with no TB:




500k pot with 150k//1nF:




500k pot with 270//1nF (Fender):




Again, the low volume seem to be getting excess treble with the Fender scheme, though there is less taper change with -15db coming at 3.5 knob setting, whereas the 150k/1nF needs to go to 2.5 to reduce to -15db. By comparison, with no TB, -15db is at a volume knob setting of about 5.5.


Conclusion

Having plotted all that, I don't really see how these Fender treble-bleed designs are better than what I had been using, but they are probably better than no TB. I hope the above will help anyone who may be curious, or who may want to optimise this simple and inexpensive circuit to suit their requirements.

[jvin248]

.

Do you think the Fender system is designed to dampen out tolerance variation of the treble bleed circuit or the typically wide volume pot range effect on system tone? That might explain the extra components, or not, they could have simply tried quite a few mods in guitars and convinced themselves there must be something more than there is.

.

[JohnH]

I think they probably worked it out by listening tests, and were happy with the results. I reckon they could have been at least 1db happier with more science!

[ownyourtone]

Reading through all of this, there is one factor that keeps popping into my mind, which may have something to do with the way humans perceive various TB circuits: the Fletcher-Munson Curve - referring to how we perceive less bass and treble frequencies with a constant, flat frequency response at lower volumes. So, it may be that a “better” treble bleed, would actually reduce midrange, say around 1-2kHz, at a faster rate than frequencies above and below said frequency. Now, do I have any idea how one would do that? NOPE! Hahaha

[newey]

Now, do I have any idea how one would do that? NOPE! Hahaha

I have an idea how one would do that. By using an active circuit. But a simple passive design one might apply to a tone control, not so much.

[sumgai]

OOT's comment about the F-M curve is spot on. Sadly, it's not a concept easily introduced to guitarists, and/or other bias-confirmation addicts vis-a-vis tube/SS, vinyl/CD, yadda yadda yadda.

That said, newey is also correct in that we usually use active circuits to implement tone controls beyond the simple treble roll-off found in our guitars. This is because, as a direct cause-and-effect found in Fletcher and Munson's research, we perceive a lowering of overall volume when we reduce the level of a fair portion of mid frequencies. We compensate by adding an amp into the circuit, trying to restore the previous perceived overall level. (And most often, succeeding in that endeavor.)

We can do this without a compensating amplifier, but then we'd perceive a "weak tone" or some other such ill that would make the circuit undesirable.  Thus.... Well, I'm sure you can see the writing on that particular section of wall.

HTH

sumgai

[JohnH]

Interesting idea from OYOT

I suspect this might be achieved with a treble bleed consisting of an inductor, capacitor and resistor all in parallel, wired to the volume pot in the usual way. Such an arrangement has its maximum impedances at the LC resonant frequency, hence giving minimum bleed at that frequency. Would need a model to test. Whether it sounds good would be very much a matter of individual opinion.

There are other ways to go from loud, down to lower volume with less mids but relatively more treble and bass, such as a switch from Bridge humbucker to Neck single.

[ownyourtone]

Feb 21, 2018 11:24:06 GMT -5 sumgai said:

OOT's comment about the F-M curve is spot on. Sadly, it's not a concept easily introduced to guitarists, and/or other bias-confirmation addicts vis-a-vis tube/SS, vinyl/CD, yadda yadda yadda.

That said, newey is also correct in that we usually use active circuits to implement tone controls beyond the simple treble roll-off found in our guitars. This is because, as a direct cause-and-effect found in Fletcher and Munson's research, we perceive a lowering of overall volume when we reduce the level of a fair portion of mid frequencies. We compensate by adding an amp into the circuit, trying to restore the previous perceived overall level. (And most often, succeeding in that endeavor.)

We can do this without a compensating amplifier, but then we'd perceive a "weak tone" or some other such ill that would make the circuit undesirable.  Thus.... Well, I'm sure you can see the writing on that particular section of wall.

HTH

sumgai

[
You made a few points I’m not sure I see applicable. Namely, our intention here is different from the one you named-restoring the perceived volume to a circuit. No, we are trying to let the perceived volume decrease, but the perceived tone remain the same. We need a tone control that changes the sound as the volume decreases in order to trick our ears into thinking it is merely the volume decreasing. So, if you take out the need for make-up gain in a circuit (or tone control), I’m still fascinated by the prospect of this idea. I could however, see the relevance of what you’ve said, should the rate of said tone control was improper, or the midrange decreased too fast - is that what you’re saying? And, now I feel in way over my head! Ha ha Ha

[reTrEaD]

Feb 21, 2018 14:13:33 GMT -5 JohnH said:

Interesting idea from OYOT

I suspect this might be achieved with a treble bleed consisting of an inductor, capacitor and resistor all in parallel, wired to the volume pot in the usual way. Such an arrangement has its maximum impedances at the LC resonant frequency, hence giving minimum bleed at that frequency. Would need a model to test. Whether it sounds good would be very much a matter of individual opinion.

There are other ways to go from loud, down to lower volume with less mids but relatively more treble and bass, such as a switch from Bridge humbucker to Neck single.

John, a totally non-scientific and math-free evaluation leads me to believe enough factors are pointed in the right direction that this could bear some fruit.


I suspect we won't see enough effect when we lower the Q enough so that the dip is properly wide. But the cable capacitance will mitigate the effect on the high end, which is exactly what FM tells us we need. Greater retention at the low end of the spectrum, moderate retention at the high end.

As the old saying goes: Half a loaf is better than none. Even if we end up with just a small slice, it could be worth doing.

[JohnH]

Thinking too much:

This may lead to some infeasible numbers. The kind of inductor that could do useful things to mid frequencies in a treble bleed circuit might need of the order of say 100k impedance at a frequency of the order of say 1kHz. According to science, that would imply about 16H, which is quite a big mutha of an inductor.

[ownyourtone]

Feb 21, 2018 22:44:01 GMT -5 JohnH said:

Thinking too much:

This may lead to some infeasible numbers. The kind of inductor that could do useful things to mid frequencies in a treble bleed circuit might need of the order of say 100k impedance at a frequency of the order of say 1kHz. According to science, that would imply about 16H, which is quite a big mutha of an inductor.

Ah, well, alas, I appreciate you all entertaining the thought, anyhow. I’m pretty pleased with the old faithful TB discussed at length here. Just my audio-engineer-nerd-brain wondering.

[moose212]

Dec 23, 2017 20:33:01 GMT -5 JohnH said:

...

Conclusion

Having plotted all that, I don't really see how these Fender treble-bleed designs are better than what I had been using, but they are probably better than no TB. I hope the above will help anyone who may be curious, or who may want to optimise this simple and inexpensive circuit to suit their requirements.

John - Finally getting back to my project, and wanted to say thanks for your analysis of the Fender-style option.  I believe you had the line "I find this very interesting, and I feel sorry for you if you do as well" - well I do!  Thanks again!

[benvigil]

I signed up just to comment on this thread because I'm finding this info REALLY helpful.

I wanted to make a few observations (valid or not) about the Fender vs non-Fender versions of the TB circuit and the measurements above...

Fenders Intent

Given that the Fender measurements for humbuckers and singles both measure and behave almost identically -- despite differing pots and circuit designs/values -- suggests to me that the results were very intentional. They chose those designs and values very deliberately. The question then becomes... what were their design goals?

Cable Length

I imagine that Fender made those design choices with the behavior of the majority of Fender players in mind. I would also bet most players who play live use a 20 or 30 foot cable. If, and that's a big if, that's the case, the longer cables (30' vs 10') would increase the capacitance and loading on the pickups at lower volumes more than at full volume. This alone would give the Fender design better performance -- i.e. closer to the stated goal: "maintain the tonal balance of the full volume tone" -- for the average player with a 30' cable.

The Tone Pot

Dec 23, 2017 20:33:01 GMT -5 JohnH said:

the Volume pot is 250k and the tone pot was no-load so out of circuit.

Using a no-load will obviously be brighter at all volumes. But just like the Cable Length example above, the effect of a tone pot in the circuit will vary based on the volume, but will always have SOME effect on the circuit at all volumes: very little effect at 10, and more effect at lower volumes. Again, this would flatten out the curves you graphed slightly, in favor of the Fender designs.

Fletcher-Munson Curve

I agree that this matters, but I'm not sure how. Previous observations notwithstanding, since FM principles tell us that perceived bass rolls off with decreasing volume levels, then increased dB levels at 2-2.5k (Fender designs) would be detrimental as it increases the FM effect. So this goes against Fenders designs.

Human Intelligibility

That said, since one of the FM revelations was that human hearing is most sensitive to frequencies within the band between 3 kHz and 4 kHz, it's interesting that the full volume graphs stay flat right up to about 3.5k. The fact that the Fender designs maintain more frequency content in the 3-4k range should be a plus in their column, within reason of course.

I'd love to hear your thoughts.

P.S. Also, it'd be great if you would graph the DiMarzio values w/500k pots... 300k/560pF. That should show less taper effect with a higher cutoff frequency so the bump would extend a little higher than the other tested values here.

[JohnH]

Hi benvigil and welcome to GN2. Thanks for your interest in tnis thread.

I didnt know that DiMarzio put out their own recommended TB recipe, so thanks for that. Ill be happy to make a plot for it. The values are almost exactly a factor of 1/2 in their effect compared to my favourite 150k and 1nF (2xR and 1/2 x C), so I expect it will improve consistency to a reasonable extent, and do its best work at low volume.

I think even on a Fender design, a no-load tone pot with a 250k volume is a good test case. It helps to max the differences that occur with TB circuits. It also properly represents both modern and classic Strats in the bridge position, which used to have no tone control and these days get a no-load control on US models.

Whether FM curves are important is very dependent on a users application and opinion. All fine if a player likes a brighter sound as volume reduces. I kept them out of the consideration. I'd probably want to turn down from a loud overdriven sound to a loud clean sound anyway so FM effects are less relevant

I tried working out TB values for longer cables, but its much more difficult to keep consistent tone. Actually, with a TB circuit and a long cird, the best tone at reduced volume is more similar to that of a 10' cord than a 20' or longer cord.

[benvigil]

Thanks John!

The DiMarzio recommendation can be found here and other places I'm sure. A good roundup by DrKev is where I originally found the values. DrKev references this thread on that page.

I'm getting ready to replace the pups in my G&L "strat" (removing Zexcoils and putting the original pups back in) and am planning on testing some of these combinations. I'll probably breadboard them with a 5-way switch to make life easier testing them.