Pickup resonance peak and Miller capacitance

[Amit]

Hi,

Do pickups form a tuned circuit with the Miller capacitance of tube amps???

I'm building a preamp/buffering system for my Tele in order to add a buffered dummy coil.
For all sorts of reasons I descided to:
1. buffer each pickup
2. buffer an extra full pickup (yup, with the magnets) as the hum pickup
3. mix the hum in parallel (and not in series like Ilitch and EB)

I want is to simulate, as much as I can, the actual loads and reactances a guitar pickup sees and respond too... "pre buffer".
There will be the known vol and tone R's, tone cap, the "amp's" input impedance setting R and grid stop R, the guitar cable capacitance and I also made a simple air coil to simulate the few pico henry that the cable have

Now what puzzeles me is the input capacitance, the Miller capacitance all tube amps have.
In most guitar amps with input triodes (99.% of them LOL) there is up to 150pF input capacitance. this capacitance is there and forms a low pass filter with the "source impedance". since it's an internal electrode capacitance (about 2.5pF) being multiplied by the tube's gain (hense the 150pF figure), I don't know if the pickups sees it like a capacitance to ground, which effects their resonance peak due to their inductance, or if the pickup's inductance is isolated from this capacitance and the low pass is actualy happening internaly in the tube.



Now why do I have a feeling this post won't generate bazillion responses?

Cheers all


p.s. I remember years ago Steve Ahola(?) and someone named Arthur(?) both made big coils like the Ilitch patent. Steve also had a picture of his coil, but I can't seem to find any post, from either guy, about how their coils performed.
Anyone else here tried that route?

[Mike Sulzer]

Hi,

Do pickups form a tuned circuit with the Miller capacitance of tube amps???

Yes, but it is not usually the dominant capacitane. Grid plate cap is about 1.7 pf in a 12AX7. The amplification factor from the Miler effect would be about 1 plus the gain of the tube, which might be 50 since the amplification factor is about 100. So this would be somewhat less than 100 pf, in the range of what you computed.

However, the cable capacitance would be expected to be around 500 pf, so it is the dominant effect. So you have the sum of pickup capacitance, cable, and Miller effect, maybe a bit under 700 pf.

The resonance would be maybe 5 KHz with some single coils. The damping of the resonant peak is important, of course.

[David Schwab]

2. buffer an extra full pickup (yup, with the magnets) as the hum pickup

Why with magnets? Are you using a pickup you had laying around? You are better off without charged magnets.

3. mix the hum in parallel (and not in series like Ilitch and EB)

One reason they did it in series because it's a passive system.

Alembic uses an active system like you are planning on using. Their dummy coil does not have a core. Having steel slugs in the core can help it pick up hum.

I want is to simulate, as much as I can, the actual loads and reactances a guitar pickup sees and respond too... "pre buffer".

Make your buffer with the input impedance a little lower. I have some JFET buffers I made, and I have an old Bartolini Hi-A TC1 preamp. The TC1 doesn't use JFETs and has a lower input impedance, and doesn't sound bright and buffered like the JFET buffers.

And of course caps and resistors can be used to tune the system.

Now what puzzeles me is the input capacitance, the Miller capacitance all tube amps have.

What Mike said about the cable capacitance.

[Amit]

Hey Mike,
Factor in stray capacitance (anything from 0.5pF to 1pF), input gain from 60 to about 75 in most guitar amps... it adds up pretty fast to a ball park of 150-180 pF. this is substential if looking on low capacitance cables, like the Klotz with 70pF per meter.

What ever the figure is, I was just not sure if that "multiplied capacitance" is isolated (or not) from the pickup's inductance. my hunch was that it should be just like you said, that it's not isolated, but I couldn't find any info to back this up.

David,
Yes I'm using a spare pickup.
I wound two test coils, one in a pickup size and one big free form, checked the hum signature with a spectrum analyzer. I found that you can't get the same hum harmonics and amplitudes without using two exact pickups+cores that are loaded (R and C) the SAME way. even the pickup shaped coil showed differences when I added a steel core and an Alnico core.
Later I'll of course demagnetize the hum pickup.

Also, doing the mixing passively is problematic. even if one "tunes" a hum pickup to match the audio pickups, once you move the vol or tone pots, that's it, the hum signatures change. you either must use dual gang pots, which is a pain with non matching hum and audio pickups, or buffer all pickups to *fix* a hum state and add the vol+tone post buffering.

As for the buffers, I'll be using same JFET opamps for distortion consistency and high input impedance, which in turn will be loaded down with (4) resistors and (3) capacitors (and a small air coil for piece of mind) to match the network a pickup sees from ground to grid. no coupling capacitors.

Btw, the Ilitch/Suhr has the farthest hum spectrum signature from a real pickup.

[Old Tele man]

...using an "active" dummy coil to achieve noise-cancellation is probably BEST done at the source (in the guitar, under the strings and pickguard), not remotely back at the preamp tube in the amplifier, separated by a lengthly double pair of cords.

[Amit]

...using an "active" dummy coil to achieve noise-cancellation is probably BEST done at the source (in the guitar, under the strings and pickguard), not remotely back at the preamp tube in the amplifier, separated by a lengthly double pair of cords.

Who said it will be anywhere else other then in the guitar?
The only thing I said is that I want to exactly mimic the whole complex impedance/reactance loads a pickup sees right up to tube preamp grid<>plate miller capacitance... and have it before the buffer(s).

[Old Tele man]

...well, because miller-capacitance is in the preamp tube in the amp.

...and the dummy coil is in the guitar, separated by the cord, that's why!

...I'm just advocating doing all the "buffering" and "noise elimination" (ie: dummy coil) as close to the noise source as possible. Feeding a buffered signal from the guitar to the amp (preamp) can greatly reduce capacitance accumulation between the pickups and the preamp, which is what you're talking about affecting the resonant peaks.

[David Schwab]

The only thing I said is that I want to exactly mimic the whole complex impedance/reactance loads a pickup sees right up to tube preamp grid<>plate miller capacitance... and have it before the buffer(s).

That's assuming someone doesn't have their guitar plugged into an effects pedal, like a Tube Screamer or something before the amp.

I think it's good enough to mimic the controls and cable before you get to the amp.

Regarding the hum canceling coil. Obviously having something exactly the same as the pickup will give you the closest match, but if you look at other systems.. Alembic's dummy coil has no core at all... just lexan. Then there is the DiMarzio Virtual Vintage stacked pickups. I've installed some of these and they sound very good.



You can see the added steel slugs in the dummy coil. The dummy is also wound with heavier wire, and has less turns than the string sensing coil.

The coils measured 8.53k on the top and 2.45k for the bottom on the Solo Pro. They are wired in series, and it works really well. It's quiet and sounds like a single coil.

Mike has a cool circuit with a dummy coil. Maybe he will post that.

[Mike Sulzer]

this is substential if looking on low capacitance cables, like the Klotz with 70pF per meter.

If you use a low capacitance cable on a passive system, you get a different sound (higher resonant frequency). If you find that OK, why are you worried about exactly allowing for the input capacitance of the amplifier?

Btw, the Ilitch/Suhr has the farthest hum spectrum signature from a real pickup.

The Suhr system is passive and series. Both pickup and dummy are modeled as a voltage source in series with an inductance and a capacitor in parallel. The impedances of the pickup and dummy add, so you want to keep the impedance of the dummy low so that the total is not much affected. Any series impedance (which includes the load on the series combination) affects both the same way. The capacitance across the coils should affect the balance only the very high frequencies (where the hum harmonics are very weak). But an addiional problem is introduced by the adjustable parallel resistor (which attempts to null the hum). Obviously its effect is frequency dependent, requiring additional correction that is only approximate.

[Amit]

Tele man,

English is not my native language so maybe I'm not explaining my self corectly... or don't understand you exactly
The issue is simple, once I put a buffer after a pickup, it will be isolated. it won't reacte to the cable capacitance, it won't reacte to the miller capacitance. now if I want to fool the pickup to think it really sees all those reactances... I need to add ALL of them before the buffer... otherwise the resonances in the pickup sound won't be the same.

The lack of those reactances "pre buffer" is why all buffered/active pickups sound un-natural. check the drawing to see what I mean.


David,

That's assuming someone doesn't have their guitar plugged into an effects pedal, like a Tube Screamer or something before the amp.

Once I have all the reactances simulated in the guitar, I don't really care if a pedal is engaged in the way to the amp. on the contrary, the pedal+guitar will sound now more natural because the pickups resonance won't change.

I think it's good enough to mimic the controls and cable before you get to the amp.

Check my sketch.
Anything that you will take out (apart from my silly coil) will change the "sound" of the pickup because the pickup doesn't see the load impedances and reactances after the buffer.

[Amit]

If you use a low capacitance cable on a passive system, you get a different sound (higher resonant frequency). If you find that OK, why are you worried about exactly allowing for the input capacitance of the amplifier?

I was just stating that to show that this capacitance can become substential and that it's actualy almost double then what you calculated.
Obviously in the simulated network (pre buffer) were able to simulate any "cable capacitance" we want. also, check my sketch, in the real world the miller capacitance and the cable capacitance effect the pickup resonance and lowpass rolloffs in a different way... you can't just add them up with a single cap to ground and expect to get the same sound.

Suhr system is passive and series. Both pickup and dummy are modeled as a voltage source in series with an inductance and a capacitor in parallel. The impedances of the pickup and dummy add, so you want to keep the impedance of the dummy low so that the total is not much affected. Any series impedance (which includes the load on the series combination) affects both the same way. The capacitance across the coils should affect the balance only the very high frequencies (where the hum harmonics are very weak). But an addiional problem is introduced by the adjustable parallel resistor (which attempts to null the hum). Obviously its effect is frequency dependent, requiring additional correction that is only approximate.

Yea I know, in a passive system this is pretty much what's needed to be in order to work... but once you see that the big coil doesn't pickup hum like a real pickup, which is the first and most important part of a "humbucking" device, then the rest of the whole idea has no meaning (even if it was flawless, which it isn't).
When I saw how the hum REALLY looks like in a spectrum analyzer, I was shocked. it has very strong harmonics up to 5kHz, and the strongest harmonics are the odd ones
It then looked reasonable to me why there are more then a couple of reviews on the net of people that installed the Suhr system and descided to pull it out. and it is why I'm taking the full active system+dummy route.

Cheers mate

[David Schwab]

The lack of those reactances "pre buffer" is why all buffered/active pickups sound un-natural. check the drawing to see what I mean.

No, they sound like the pickups. It's not unnatural at all, just brighter.

Once I have all the reactances simulated in the guitar, I don't really care if a pedal is engaged in the way to the amp. on the contrary, the pedal+guitar will sound now more natural because the pickups resonance won't change.

I think beyond the controls and cable, you wont hear much. I routinely plug directly into the hi-z input on a digital mixer, and the pickups still sound like the pickups do if I plug into a tube amp or tube preamp.

But try it and see.

[Mike Sulzer]

The magnitude of the impedance of a 100 pf cap at 5 KHz is about 300 K ohms. The 33K grid stopper does not affect the circuit very much.

I do not think you understand what I mean about the operation of the dummy circuit. Suppose you used a dummy that had the same sensitivity as the pickup aat low frequencies. The adjustment is unnecessary and the cancelation would work to past 5 KHz.

This can be achieved with a semi-active circuit in which the coil is isolated by a FET (the circuit David referred to):

"You can cancel the magnetic hum in high impedance pickups very well. I prefer to use a buffered dummy coil. A single source follower serves as the buffer, and the canceling is injected at the bottom of the single coil pickups (that is, the connection that normally goes to ground) through an adjustment pot (low R value). The dummy coil needs to have 10 or 20 % more sensitivity than the pickup so that you have some adjustment range. And the resonant frequency of the dummy coil needs to be well above that of the pickup/cable system. This method keeps the single coil sound, and has very simple electronics."

The L and C of the dummy coil are irrelevant; the voltage of the dummy cancels that of the pickup. The system fails to cancel when the pickup capacitance become important, typically well above 5 KHz with a singe coil.