"Variable" Impedance DI box KK pickups?

[RobKay]

Just hang a resistor across the kk pickups or if you want to be fancy get a multi position switch and put several different values so you can switch between them...its all smoke and mirrors...oh, and money

[Andy Howell]

My Headway has three impedance setting but I can't say they make a great difference!

[James May]

I know this topic is a little confusing to non-engineers. It's been stated before in other ways, here is a simplified technical explanation of the way it works:

The equivalent source capacitance of the piezo elements forms a highpass filter with the input impedance. In the case of K&Ks, the typical source capacitance is roughly 20nF-24nF.

Working into 1megohm means that the bass response will be 3dB down at 7-8Hz, and fall off at 6dB per octave below that. That's pretty dang low, well below what is needed.

Working into 100Kohms means that the bass response will be 3dB down at 70-80Hz. You'll probably notice that.

With a K&K, anything above 1megohm will likely not sound any different, as the corner frequency moves down proportionally below 7-8Hz. This is borne out by many people's observations. Thank you Doug Young.

In the case of undersaddle piezos, things get more difficult because they have much smaller equivalent series capacitance. For example, the AG series Fishman measures just 500pF. That is 44 times smaller than the K&K. So to get the same low frequency response, it requires an input impedance 44 times GREATER. That's why you see 10megohm input impedances on some devices. That would put the low freq roll off point at 32Hz.

Thus, the concept of impedance matching is not the best term to use for this issue. Better to think about the impedance being high enough to suit your pickup and low frequency requirements. To a large degree, deficiency in the bass can be exactly compensated by the right bass boosting EQ. Some onboard preamps designed for USTs have this.

In the case of ToneDexter, it does this automatically when training.

[jonfields45]

Originally Posted by James May

I know this topic is a little confusing to non-engineers. It's been stated before in other ways, here is a simplified technical explanation of the way it works:

The equivalent source capacitance of the piezo elements forms a highpass filter with the input impedance. In the case of K&Ks, the typical source capacitance is roughly 20nF-24nF.

Working into 1megohm means that the bass response will be 3dB down at 7-8Hz, and fall off at 6dB per octave below that. That's pretty dang low, well below what is needed.

Working into 100Kohms means that the bass response will be 3dB down at 70-80Hz. You'll probably notice that.

With a K&K, anything above 1megohm will likely not sound any different, as the corner frequency moves down proportionally below 7-8Hz. This is borne out by many people's observations. Thank you Doug Young.

In the case of undersaddle piezos, things get more difficult because they have much smaller equivalent series capacitance. For example, the AG series Fishman measures just 500pF. That is 44 times smaller than the K&K. So to get the same low frequency response, it requires an input impedance 44 times GREATER. That's why you see 10megohm input impedances on some devices. That would put the low freq roll off point at 32Hz.

Thus, the concept of impedance matching is not the best term to use for this issue. Better to think about the impedance being high enough to suit your pickup and low frequency requirements. To a large degree, deficiency in the bass can be exactly compensated by the right bass boosting EQ. Some onboard preamps designed for USTs have this.

In the case of ToneDexter, it does this automatically when training.

Exactly correct.

[Doug Young]

Great summary, James. Nice to have some math behind this. I think you hit the nail on the head that we need to talk about impedance being high enough as opposed to "matching", which is quite meaningless here.

I wonder if you have anything to add about the effects of poor quality long cables when used with high impedance inputs? One of my speculations to explain those who report massive bass increases when they plug K&Ks into 10M is that they may have long high-capacitance cables and are getting high frequency rolloff, which they hear as bass boost. My electronics is too rusty to do the math on this, but maybe you can clarify how this would play out with a K&K?

[James May]

Originally Posted by Doug Young

...

I wonder if you have anything to add about the effects of poor quality long cables when used with high impedance inputs? One of my speculations to explain those who report massive bass increases when they plug K&Ks into 10M is that they may have long high-capacitance cables and are getting high frequency rolloff, which they hear as bass boost. My electronics is too rusty to do the math on this, but maybe you can clarify how this would play out with a K&K?

Good question Doug, and the answer is perhaps non-intuitive.

Unlike the situation with a magnetic pickup, when working into a cable with high capacitance, a piezo pickup can only suffer signal attenuation that is flat with frequency. It can't boost bass or roll off highs. This is because the piezo capacitance and the cable capacitance form a CAPACITIVE DIVIDER which is agnostic to frequency. It's simply an attenuator.

I just measured my 15ft cable and got 600pF. With a K&K, that would give an attenuation of about 0.25dB!

With an UST, attenuation more like 5dB. But again, with no tilt to the frequency response.

This is one of the good things that come with using piezos.

With magnetic pickups, the situation is entirely different, as its source is both resistive and inductive, rather than capacitive, and that does create some havoc working into capacitive cables.

[Doug Young]

Well that blows my theory :-)

One of the big mysteries around the impedance rumor mill is people who report a massive bass boost with 10M inputs and K&Ks. K&K techs have told people it will cause a 20db boost, if I recall, and some people confirm that they experience this. The math clearly says it can't be (and you can't haver a boost with passive circuits anyway). I suspect there's something else going on for those people - I certainly can't reproduce - but I don't know what it is.

[jonfields45]

Originally Posted by Doug Young

Well that blows my theory :-)

One of the big mysteries around the impedance rumor mill is people who report a massive bass boost with 10M inputs and K&Ks. K&K techs have told people it will cause a 20db boost, if I recall, and some people confirm that they experience this. The math clearly says it can't be (and you can't haver a boost with passive circuits anyway). I suspect there's something else going on for those people - I certainly can't reproduce - but I don't know what it is.

If the load impedance is dramatically higher than the source impedance, then the heads of a K&K will see each other (and themselves) as a more significant load than the input impedance of whatever they are plugged into. In this case you get a low pass filter. I thought I heard it once with the "guitar" mode engaged in a Mackie mixer, but it could have been something else and I no longer have that mixer or a K&K around to try it again. It was my first gig with a new K&K plugged directly into a new mixer. If I used the microphone mode there was little bass, if I engaged the "guitar" mode there was little treble. I did not have fun that night.

[James May]

Originally Posted by Doug Young

Well that blows my theory :-)

One of the big mysteries around the impedance rumor mill is people who report a massive bass boost with 10M inputs and K&Ks. K&K techs have told people it will cause a 20db boost, if I recall, and some people confirm that they experience this. The math clearly says it can't be (and you can't haver a boost with passive circuits anyway). I suspect there's something else going on for those people - I certainly can't reproduce - but I don't know what it is.

Merely speculating, I can think of a couple scenarios that might be mistaken for a boost:

1. the bass sounds louder than it did when there was a low end rolloff
2. it is working into an active preamp designed for a UST that indeed does have a low end compensatory boost

[James May]

Originally Posted by jonfields45

If the load impedance is dramatically higher than the source impedance, then the heads of a K&K will see each other (and themselves) as a more significant load than the input impedance of whatever they are plugged into. In this case you get a low pass filter. ....

I have to respectfully disagree with this. In the audio range, a very good approximation of a piezo is a voltage source in series with a capacitor, with no source resistance of any consequence. Therefore, working into a capacitive load can only make an attenuator. Working into a resistive load can only make a high pass filter. The lowpass response is not realizable.

When you put 3 of the same piezo discs in parallel, like K&K does, the result is that each contributes 1/3 of the total output level, the remaining 2/3 is attenuated by the load of its two neighbors. But somewhat magically and beautifully, the overall level is the same since each neighbor also contributes 1/3. One, two, or three of the same discs all picking up the same source and the level is unchanged! But the source capacitance is higher (3x that of a single disc), which means the source impedance is lower, and therefore it becomes harder for a capacitive cable load to cause attenuation.

[jonfields45]

Originally Posted by James May

I have to respectfully disagree with this. In the audio range, a very good approximation of a piezo is a voltage source in series with a capacitor, with no source resistance of any consequence. Therefore, working into a capacitive load can only make an attenuator. Working into a resistive load can only make a high pass filter. The lowpass response is not realizable.

When you put 3 of the same piezo discs in parallel, like K&K does, the result is that each contributes 1/3 of the total output level, the remaining 2/3 is attenuated by the load of its two neighbors. But somewhat magically and beautifully, the overall level is the same since each neighbor also contributes 1/3. One, two, or three of the same discs all picking up the same source and the level is unchanged! But the source capacitance is higher (3x that of a single disc), which means the source impedance is lower, and therefore it becomes harder for a capacitive cable load to cause attenuation.

I agree there is no issue with the cable capacitance. The issue is if the impedance of the load (preamp input) gets very large and presumably designed to be resistive, do the highs leak across the capacitance of the PUP because that is a lower impedance than preamp input. The manufacturers of these pickups seem to be certain that they won't work well into 10M ohm loads; something Doug has not reproduced and I'm not willing to testify I encountered that evening. I did make a 1M shunt resistor and used that pickup and mixer (in guitar mode with the shunt resistor) after that for several years with no issues. The simple lumped model we've been discussing might not be up to the job at this extreme point.

[Kittoon]

Well. .............that clears THAT up! ..........

[Cuki79]

Thanks guy,

I learned a lot reading that thread! I had not idea of the UST and SBT capacitance before.

James, how much are the Fishman Matrix or the Baggs element for example?

Cuki

[jonfields45]

The time constant that might be interesting in this hunt for a low pass filter is the input impedance of the preamp, resistive, times the total capacitance of the PUP.

[James May]

Originally Posted by Cuki79

Thanks guy,

I learned a lot reading that thread! I had not idea of the UST and SBT capacitance before.

James, how much are the Fishman Matrix or the Baggs element for example?

Cuki

I don't have samples handy to measure, but I suspect, based on their behavior, that they are similar to the Fishman AG UST, likely 300-500pF. The coaxial cable used in some pickups is even less at about 1nF per meter.