Graphtech Piezos and Tremolo don't play nice

[finn]

Howdy, dear members of this excellent forum!

----
UPDATE: Solution found!
The root cause of the problems was, indeed, that the signal of the Graph Tech piezo saddles was considerably stronger than the signal of such an acoustic piezo pickup, for which the preamp has been designed. The big voltage peaks which were caused by the changing string tension when using the whammy bar overloaded the input of the piezo preamp completely. After experimenting with numerous different component values, the solution I chose, was a simple 100nF capacitor parallel to the piezos, which reduced the signal at the input of the preamp to a manageable level.
----

This is my first post after joining here, having found and read an old, excellent piezo-related thread by JohnH ( guitarnuts2.proboards.com/thread/3718 ).

I have a weird problem and hope that someone here might be able to help me fix it.

In short: When I use the tremolo arm of my Strat "partscaster" any faster, the sound of the piezos cuts abruptly out&in. If I move the tremolo arm veeery slowly or the tremolo arm is kept steady in any position, there are no problems. Ever. I have wiggled the wires, measured the grounding etc. and really don't think that the problem would be in any faulty cconnections.

I posted my question with a video to Reddit two weeks ago, but haven't really received any replies with much value from there.
www.reddit.com/r/Luthier/comments/19dukhr/graphtech_ghost_piezos_dont_like_the_tremolo/

If you, however, please check my short video there, you'll surely understand my problem.

Meanwhile I do have a theory what might be the cause from the electrical point of view.

Because the relationship between pressure and the voltage created by piezoelectric materials is direct and proportional and the pressure change caused to the piezo saddle by influencing the string tension with the the tremolo is a LOT bigger than the pressure changes caused by the vibration of the strings, a sudden increase or decrease of pressure caused by using the tremolo bar might create a spike in the electric potential, which would either discharge somewhere™, where it is not supposed to or - more probably - simply oversteer the piezo preamp to total clipping.

If the tremolo bar movement is slow, the change in the electric potential would perhaps be handled more gently, possibly simply discharging it through the inner resistance of the piezo elements.

If this would, indeed, be the cause of my problem, the next question would obviously be, what would be the remedy? Perhaps some kind of a high pass filter with a cutoff e.g. below ca. 50Hz? Would some simple resistor or capacitor do the trick? Connected how? Which value(s)...?

(Could I really be the only(!) one, who has ever encountered this problem...? )

[newey]

finn-

Hello and Welcome to G-Nutz2!

Your explaination of a possible mechanism for this problem makes sense, but then I thought the same as you:

(Could I really be the only(!) one, who has ever encountered this problem...? )

Are you using a commercially-available piezo system or was this a DIY effort? And, where is the piezo trnasducer mounted?

Also (and I'm certainly no expert on piezos, so I could be wrong) it is my understanding that a piezo pickup is not sensing the strings directly, the change in pressure to generate the signal comes from the vibrations of the guitar body itself, not from the strings. I say this because I have seen/heard piezos work just fine when mounted on all sorts of places on a guitar, not just at the bridge or under the strings. A piezo disc can be pressed between the neck heel and the body at the neck attachment and it will work just fine.

[sumgai]

finn,


First off, to The NutzHouse!

We see right away from that video that you're are truly Nutz, with a guitar like that you'll fit right in here.

I had a Graphtech Ghost system for years on an old Strat, it didn't exhibit the problem you're seeing. The only thing I can think of is that you're changing the downward pressure of the string across the piezo too quickly to be handled properly. My first thought, seeing as you've been dealing with this for awhile with no solution, would be to contact the nice folks at Graphtech.

But there's also a forum that you don't list as part of your search for answers, you might want to give them a try:

Vguitarforums.com

That's the mecca for all things related to synth guitars/basses, and by extension, to pickups for those systems. Might be worth a shot, who knows....

HTH




sumgai

[finn]

Thanks for your reply. I use the Graphtech Ghost piezo saddles, which pick up the sound of each string separately. The wires from the 6 piezos are combined together with a "summing board", which connects them in parrallel.

[finn]

Feb 5, 2024 12:46:00 GMT -5 sumgai said:

We see right away from that video that you're are truly Nutz, with a guitar like that you'll fit right in here.

Thanks 

Before anyone asks, that axe has a Guitarbuild body, a Mexican Fender maple neck, Fender noiseless pickups, a Freeway switch and the second 5-way-switch installed as a "varitone", changing the capacitance of the tone pot.

The lower tone pot is bass cut and in the lower "horn" of the body is a three-way slide switch for selecting magnetic or piezo pickups or both when the mono output is used. The piezo and magnetic signals are summed to the mono output with a small mixer circuit, which AnalogWorkshop kindly designed for me.

There is also a tele-like stereo output jack, where the magnetic and piezo signals are both in their own channels, the magnetic ones being completely puristically passively wired.
 
The volume pot push-pull switches the mid-pickup out of phase, the upper tone pot push-pull bypasses the tone circuitry completely and the lower tone pot push-pull switches off the piezos, when the stereo jack is used. The piezos are fed through a "Fishman-esque" (chinese) preamp (even with a tuner! ), which you can see mounted in the pickguard (BTW: was a really tight fit, depth-wise..)

[mikecg]

Hello finn,

Yes, my guess is basically the same as yours - rapid motion of the trem is generating a large low frequency 'step' function from all six piezo pickups that are overloading the pre-amp's input opamp. These pre-amps have a high impedance front end, usually something between 1 and 10 Meg-Ohms. The piezo pickups are purely capacitive, and in the case of the Graphtech product there are six sub-miniature coax cables - all in parallel. Two things to try - you can implement a simple capacitive divider to attenuate the piezo signal by wiring a small capacitor across the input coax connection to the preamp - probably best done at the 'summing' connector - try 470 pF to start with. If this 'improves' matters - you will know that the problem is most likely around the preamp input stage. The next step may be to do as you have already suggested, and try adding some form of high pass filter at the preamp input to reduce the amplitude of the very low frequency signals generated by your trem movements. You can try wiring a 1 Meg-Ohm resistor in place of your 470 pF capacitor. This should produce a noticeable increase in the lower -3 dB cutoff frequency.

[finn]

Feb 5, 2024 14:07:49 GMT -5 mikecg said:

Hello finn,

...you can implement a simple capacitive divider to attenuate the piezo signal by wiring a small capacitor across the input coax connection to the preamp - probably best done at the 'summing' connector - try 470 pF to start with. If this 'improves' matters - you will know that the problem is most likely around the preamp input stage. The next step may be to do as you have already suggested, and try adding some form of high pass filter at the preamp input to reduce the amplitude of the very low frequency signals generated by your trem movements. You can try wiring a 1 Meg-Ohm resistor in place of your 470 pF capacitor. This should produce a noticeable increase in the lower -3 dB cutoff frequency.

When you suggest connecting a capacitor "across the input coax connection", do you mean between the hot wire and ground because - if I have understood things correctly - that would be a low-pass filter and not high-pass, what probably would be needed. 🤔

I am truly not any specialist in calculating (or guessing...) component values but in order to get a high-pass filter, I have understood that the capacitor should be connected in series with the pickups to the hot wire between the pickups and the preamp input and a resistor parallel with the pickups - between the hot wire and ground - could be used to tune the cutoff frequency and reduce the overall amplitude.

Somewhere in this forum I read that a piezo might have a capacitance of ca. 20nF, which means that the 6 piezos in parallel might be approx. 120nF. I wonder whether that would be sufficient information for calculating the component values for a simple high-pass filter with a cutoff of e.g. 50Hz?

[JohnH]

I'm also thinking if input impedance. I had similar issues with my (very different) home-brew build that you linked to above.

The piezos react to pressure that they sense, and if they are feeding a very high impedance, their frequency response is huge, down to very low sub-bass frequencies. With my one, I was expecting to need the preamp to be high impedance, but this gave me crazily low frequency response. It worked fine for all the actual string vibration, but pressing on the bridge could create a huge almost dc pulse that blocked out all sound for a second or two by maxing out the input.

I needed to decrease the resistor that set my preamp input impedance, so there was enough signal draw to bleed off these unwanted low frequencies and dc transients.

My piezo was a buzzer element, same theory as a commercial piezo pickup but quite different value, with a high capacitance. I needed an input imoedance decreased to 150k, and then it solved this completely with no change to sound. The C of the piezo znd the 150k R gave me a roll off below about 50hZ and was fine for stopping the piexo getting swamped by low pulses.

I dont know if your issue is the same. I'm surprised that a commercial unit would have this issue unless the preamp is not part of the family. If there is a similar issue, it would be good to find out the capacitance of the piezos in order to judge the input impedance needed.

EDIT: I see finn has also homed in on his idea!

[mikecg]

Hello finn,

As the source impedance is purely capacitive, a capacitance to ground will create a simple capacitive divider - an attenuator, in other words.
A resistor to ground, in this topology, will increase attenuation at lower frequencies - a high pass filter with a higher low frequency cutoff, in other words.

[finn]

Feb 5, 2024 14:44:09 GMT -5 JohnH said:

I'm surprised that a commercial unit would have this issue unless the preamp is not part of the family.

The preamp and the piezos do NOT come from the same source but on the Graphtech webpage it is said "Works with any of our ghost preamp kits along with other companies' preamps.". Perhaps (probably, actually) Graphtech has somehow™ taken care of this issue in their own preamps.

[JohnH]

I looked through the links above, but didn't see specific data on piezo capacitance and preamp input impedance. But the ghost summing board does indeed seem to join them all In parallel, so they could add up to a substantial C value. So I reckon you could test with various resistors from preamp piezo input to ground. Maybe a (500k ?) pot to start with, so you can adjust ahead of deciding on a fixed value?

[finn]

Feb 5, 2024 16:00:08 GMT -5 JohnH said:

So I reckon you could test with various resistors from preamp piezo input to ground. Maybe a (500k ?) pot to start with, so you can adjust ahead of deciding on a fixed value?

If a simple resistor parallel with the pickups might really solve the case, that would really be magnificent! The "summing board" with its 8 connector pairs has one "superfluous" pair of connectors, anyway. I'll probably pull some wires from the summing board to a breadboard and start testing there with some trimmers and capacitors. I just wanted to first have some ballpark figure for the capacitor value. The remaining question would still be, whether a (470pF or within which value limits?) capacitor parallel with the resistor would actually be an improvement or not.

[JohnH]

I'm really thinking it's a resistor to ground that will be best. That way you can set it up to filter say all the sub-bass that you dont want, and keep the full acoustic signal. That assumes that as it is, without moving the trem, the sound levels and knob settings are all good? But if you want to actually reduce the acoustic levels, then a cap may help.

I'm guessing resistor will be 100k to 500k and if you try a cap, maybe 10-100nF but I really don't know and doubt a cap is what you'd want.

Good luck! it's very interesting...

[mikecg]

Hello finn,

You may be interested to see a simple analysis of the simulated high pass characteristics of your 'input configuration', as follows:



C1 is my guess regarding the total capacitance of the 'isolated' piezo elements - assumed to be 6 x 50 pF = 300 pF.
C2 is my guess regarding the total capacitance of the 6 sub-miniature coax cables, again 6 x 50 pF = 300 pF
R1 is the input impedance of the pre-amp module assumed to be purely resistive in this analysis.



The plot shows the high pass response of this circuit with the value of R1 stepped through 250 k, 500k, 1 Meg, 2 Meg, 4 Meg, and 8 Meg.
Note that with C1=C2, there is an unavoidable attenuation of -6 dB resulting from the C1/C2 capacitive divider. 
The black line at -9 dB is the '3 dB down' marker line wrt to the -6 dB 'baseline'.
The intersection of this line with the red plots shows the 'corner frequencies' or '3 dB down' frequencies that characterize the high pass response(s).
Low E on the guitar has a fundamental frequency of around 82 Hz, so if we decide to pick a value for R1 that produces a 3 db drop in output at 82 Hz, from the plot, we would be looking at setting R1 somewhere between 2 an 4 Meg. This range for R1 should produce an attenuation wrt the -6 dB 'baseline' of between -36 dB (x0.016), and -44 dB (x0.006), at a 'trem frequency' of around 1 Hz.

There is one disturbing possibility that comes to mind, and that is that the output of the piezo elements are effectively saturated by the trem movement. If this is indeed the case, no amount of filtering will restore the relativity low level signal from the string vibration.

It is quite possible that the combination of a piezo pickup with a trem is just a no-no! 

Of course - the manufacturers and retailers of piezo pickups will be in no hurry to tell you this!

[sumgai]

One more thought....


A couple of decades ago, Roland brought out a new synth/modeler unit, and the GK3 hex pickup. That particular pickup demonstrated a susceptibility for sub-sonic interference, especially when used with older equipment. Richard McLish developed a small board to go between the pup and whatever followed, all it had was a sextet of high-pass filters, and he sold the board to those who needed it, for a decent price, and without fanfare.

I'm thinking that if you are having this issue, then it's possible that others might have it as well. In that case, Richard may have already constructed a "work around" for it. It won't hurt to ask, I'm sure.


rmcpickup.com



HTH






sumgai

[sumgai]

Feb 6, 2024 8:20:14 GMT -5 mikecg said:

It is quite possible that the combination of a piezo pickup with a trem is just a no-no!

Of course - the manufacturers and retailers of piezo pickups will be in no hurry to tell you this!

Absolutely not. I not only used this very setup for several years, but the Vguitarforums is absolutely full of users who aren't experiencing any issues of this sort.

HTH





sumgai

[gerinski]

I don't have the knowledge to give any valuable input here, I just wanted to say that unless I'm wrong, the Gibson Les Paul Axcess Alex Lifeson signature has a Floyd Rose and piezos.

[finn]

Feb 6, 2024 17:23:17 GMT -5 sumgai said:

Feb 6, 2024 8:20:14 GMT -5 mikecg said:

It is quite possible that the combination of a piezo pickup with a trem is just a no-no!

Of course - the manufacturers and retailers of piezo pickups will be in no hurry to tell you this!

Absolutely not. I not only used this very setup for several years, but the Vguitarforums is absolutely full of users who aren't experiencing any issues of this sort.

When you say "used this very setup", do you mean specifically piezos like the Graphtech Ghost piezos, where the tremolo causes a substantial pressure change in the piezos or do you mean such piezos where the strings just run over a bridge with piezo elements but the piezos don't carry the entire load of the (changing) string pull? In such bridges, the tremolo causes only a slight change of pressure, which is defined by the angle of the string going over the bridge.

[finn]

Feb 6, 2024 8:20:14 GMT -5 mikecg said:

Hello finn,

C1 is my guess regarding the total capacitance of the 'isolated' piezo elements - assumed to be 6 x 50 pF = 300 pF.
C2 is my guess regarding the total capacitance of the 6 sub-miniature coax cables, again 6 x 50 pF = 300 pF
R1 is the input impedance of the pre-amp module assumed to be purely resistive in this analysis.

Awesome work, mikecg! Thanks a TON!

Would guessing C1 to be e.g. 120 nF and C2 to be 100 pf instead of both being 300 pF mean that the resistance value would need to be increased or decreased?

[finn]

Feb 5, 2024 17:42:54 GMT -5 JohnH said:

I'm really thinking it's a resistor to ground that will be best. That way you can set it up to filter say all the sub-bass that you dont want, and keep the full acoustic signal. That assumes that as it is, without moving the trem, the sound levels and knob settings are all good? But if you want to actually reduce the acoustic levels, then a cap may help.

I'm guessing resistor will be 100k to 500k and if you try a cap, maybe 10-100nF but I really don't know and doubt a cap is what you'd want.

Good luck! it's very interesting...

Thanks, JohnH.

I'll probably first try to wire a 500k trimmer on a breadboard parallel to the pickups and see (well, hear, actually 🙃), how the results are with different settings. As if Graphtech had deliberately chosen to use the 8x2 connector as the summing board because after the 6 pickups and the output wire are connected, one connector pair still remains empty...   

BTW: I'm currently trying to figure out where I can get such connectors with 2mm pitch which Graphtech uses for the wiring. Ordering them or such a cable from abroad here to Finland would be ridiculously slow and expensive... Should anyone have any idea of the specific connector type, I would be thankful for a hint.

[mikecg]

Feb 6, 2024 18:42:33 GMT -5 finn said:

Feb 6, 2024 8:20:14 GMT -5 mikecg said:

Hello finn,

C1 is my guess regarding the total capacitance of the 'isolated' piezo elements - assumed to be 6 x 50 pF = 300 pF.
C2 is my guess regarding the total capacitance of the 6 sub-miniature coax cables, again 6 x 50 pF = 300 pF
R1 is the input impedance of the pre-amp module assumed to be purely resistive in this analysis.

Awesome work, mikecg! Thanks a TON!

Would guessing C1 to be e.g. 120 nF and C2 to be 100 pf instead of both being 300 pF mean that the resistance value would need to be increased or decreased?

Hello finn,

With C1 = 120 nF and C2 = 100 pf, R1 will be around 15k.

However, based on some measurements I have made on my own Fishman rod type piezo pickup, I would be very surprised if your pickup's capacitance (C1) was more than 1 nF, mine measured ~ 350 pF with ~ 50 pF of cable - that is where my value of C1 = 300 pF came from. The rod type pickups have six piezo elements all in parallel so I know that each element has a capacitance of ~ 50 pF. Your Graphtech product also has six piezo elements - and they are probably very similar to the ones in my rod pickup - so based on that assumption - I'm guessing that C1 for your pickup is also ~ 300 pF.

Some of the cheap disc type piezo pickups, used on acoustic guitars and other stringed instruments to sense the vibration of the instrument surfaces (contact piezos), have much higher values for C1, but as far as I know, nowhere near as high as 120 nF - 12 nF, maybe. 

[finn]

Feb 6, 2024 19:33:04 GMT -5 mikecg said:

Some of the cheap disc type piezo pickups, used on acoustic guitars and other stringed instruments to sense the vibration of the instrument surfaces (contact piezos), have much higher values for C1, but as far as I know, nowhere near as high as 120 nF - 12 nF, maybe. 

Ok and thanks for the clarification. As I said earlier, I just picked the 20 nF value for the piezos somewhere™ in this forum. It might, indeed, have been a value for the disc type.

("Your capacitance may vary and the real values might appear higher than they are." - SCNR)

[JohnH]

It's good to see mike's C values, I couldn't find anything similar online. So if you try this added resistor loading, I suggest to try it very simply as a quick-dirty test to see if this is in the right direction or not.

Mike's charts are helpful too. You can use them with scaling to judge different sets of values. eg, double the total capacitance and halve the resistance would end up with the same expected curve etc

20nF was probably the piezo buzzer disc that I used. Not likely anywhere near your value but let's see!

[sumgai]

Feb 6, 2024 18:21:36 GMT -5 finn said:

Feb 6, 2024 17:23:17 GMT -5 sumgai said:

Absolutely not. I not only used this very setup for several years, but the Vguitarforums is absolutely full of users who aren't experiencing any issues of this sort.

When you say "used this very setup", do you mean specifically piezos like the Graphtech Ghost piezos....

I mean that I had a 97 MIJ Strat that had a Ghost setup including the Hexpander. I went up to Canada (Delta, not all that far from the US border) and consulted with Leigh and David, the chief techie and owner, respectively.  An otherwise stock vibrato unit, set to float because I like to 'lean' on it to raise the pitch as well as go lower.  And strings made by Dean Markley, their stainless steel set of 10-46 gauge.   Never a problem.

I've installed several sets of hex pickups from various companies, on various guitars, for various customers, and they're pretty much all the same.  So easy that a rookie modder should be able to do it without errors.  But that does not mean that there can't be problems.  Like the  piezos themselves could be bad, or the lead coming out of them, ditto.  Companies try their best to ship error-free products, but then again, that's why they have warranties.

I've always said that the best method of troubleshooting is "substitution with a known good part".  The problem with that little dictum is that it runs up against the trade-off of either cost, time, effort, or any combo thereof.  But when push comes to shove, that's the bottom line,   All that's left is to determine priorities - do you take the easy route, the cheap route, the quickest route, or do you just bite the bullet and plug in a known good part, and see what happens.  Of course, that part is up to you, the modder.

HTH





sumgai

[mikecg]

Hello sumgai and finn, many thanks sumgai, for confirming that these Graphtech pickups and trems are a compatible and trouble free combination, when installed using the standard Graphtech parts and accessories. It is helpful to hear that we are not 'flogging a dead horse', in this particular case.
I have given this problem some more thought - and based partly on what you have said, have come to the conclusion that the problem here is that finn has used a non-Graphtech preamp that does not appear to be completely compatible with his Graphtech piezo pickup. He describes his preamp as a "Fishman-esque" (chinese) preamp, and I am familiar with this preamp type. They are designed for use with the standard under saddle rod piezos used on acoustic guitars, and subsequently do not need to deal with the additional input level 'excursions' imposed by the use of a trem. My guess is that the input stage opamp in question is not protected against rail to rail input excursions, and goes into 'latch-up' mode as a result. And once in latch-up, it will take some significant time to recover. In this case, preamp input sensitivity can be reduced by wiring a small capacitor in parallel with the preamp input, I  have suggested a trial with a 470 pF capacitor. The final value should be selected so that it is just large enough to prevent latch-up for all normal trem movements. 

[JohnH]

How about trying R and also C for this? If the actual characteristics of the preamp and pickup are suitable, adding R will bleed away the problem ar source so long as R doesn't get so low as to start cutting into bass notes too much. Or maybe some C and some R in parallel.

[finn]

Feb 7, 2024 14:20:43 GMT -5 JohnH said:

How about trying R and also C for this? If the actual characteristics of the preamp and pickup are suitable, adding R will bleed away the problem ar source so long as R doesn't get so low as to start cutting into bass notes too much. Or maybe some C and some R in parallel.

That's my plan, JohnH. I just bought a bunch of caps (220pF, 470pF, 330pF, 1uF) and resistors (390K, 470K, 510K, 560K, 680K, 1M) and will start experimenting, probably starting with the above mentioned 470pF and 510K, as soon as I find out where here in Finland I can find a suitable connector with 2mm pitch (like this) to get the breadboard connected to the summing board in parallel with the pickups. I still haven't managed to find the exact series/type of that connector...

[mikecg]

Hello finn,

You can order a suitable cable with connectors, from Poland and they will ship within the EU for a small charge:


guitarpartscenter.eu/en_US/searchquery/pe-5021-00+ghost+dual+connector+cable/1/phot/5?url=pe-5021-00,ghost,dual,connector,cable&msg=ec029179d2


Alternatively, you may be able to find the same part in Finland?

[finn]

Feb 7, 2024 18:26:40 GMT -5 mikecg said:

You can order a suitable cable with connectors, from Poland and they will ship within the EU for a small charge:

guitarpartscenter.eu/en_US/searchquery/pe-5021-00+ghost+dual+connector+cable/1/phot/5?url=pe-5021-00,ghost,dual,connector,cable&msg=ec029179d2

Alternatively, you may be able to find the same part in Finland?

Unfortunately I haven't been able to find any source for those cables in Finland and also not for such plain connectors. I got so impatient that I just placed my order to Graphtech. I will, however, bookmark the guitarpartscenter.eu site. Hadn't heard anything about them, so far. Thanks!