Noobman Cutting Head

[opcode66]

Exactly. Every transducer has its own fingerprint if you will.

[EpicenterBryan]

1. Moving mass of the driver (and all of the other moving elements in your design).

Mark's point is a big deal with these drivers. The springs account for most of it. With the coil, spring and those brass rivets connecting the spring to another plate of metal that the coil attaches to, it comes in at a staggering 7.9 grams.

IMG_3531.JPG

Ciuens is using these drivers and has replaced the metal spring with a plastic one (and an additional V-Spring to handle vertical load when the head is cutting). I'll be looking into that option too. At 7.9 grams of moving mass in just the driver it kind of swamps efforts in getting the rest of the mass down - like using carbon fiber and all that...

Yes, I realize there is weight associated with those wires which do not move, but you get the idea. I wasn't about to cut them off just for this photo.

Bryan

[opcode66]

To give some comparison. The entire moving mass for Bladerunner weighs less than 7 grams. This includes: 2 x bobbins with coils, 2 x moving portion of transducer springs, 2 x metal linkage, torque tube and V spring. I've reduced the size of everything to arrive at such small mass for both channels plus torque tube.

[markrob]

I guess the response curve that Bryan plotted that shows a distinct drop starting at 300 is nonsense too then?

Hi Todd,

That's the velocity response. His measuring device (a phono pickup) is velocity sensitive. If you convert that curve to display the amplitude response of the head, you'll find it is flat from DC to the system resonance range (300Hz in this case). The velocity response has to be 0 at DC because the head is not moving. However, it can be displaced to a static position if it is fed a DC current. To convert between amplitude and velocity you take the first derivative of the amplitude response. The derivative function has a +6db/oct rising response (to infinity and beyond) which is what you see in Bryan's curve in that range.

http://en.wikipedia.org/wiki/Differentiator

You can look at disk recording from a velocity or amplitude (excursion) point of view. The most common way is velocity based, both methods are really two different views of the same system. The RIAA curve we see is almost always shown from a velocity standpoint because most modern pickups are velocity responsive and most cutting feedback systems using moving coils which are also velocity responsive. If you use a piezo pickup for playback, for example, you would need to use the amplitude version of the RIAA to correctly decode recordings.

So now we get down to cutting low frequencies. Bryan's head response drops at -6db/oct as you reduce frequency (or rises +6db/oct as you increase frequency). If you look at the velocity based RIAA record curve, you will see that it matches Bryan's head in the range of 50 Hz to his system resonance at 300 hz. If he had set his resonance higher (1-2 Khz is typical), the curves would match from 500 Hz down to 50 Hz. To make the curves match below 50 Hz, you need to add some corrective EQ to the head. This a simple analog (or digital) filter that boosts the low frequencies +6db/oct from 50Hz down to the desired low frequency limit of the cutting system. Note that the RIAA curve seems to imply that the response is flat from DC to 50 Hz. In the real world, you might set a low frequency limit and let the head velocity response rolloff below some point (perhaps 10 Hz). If you were to design an RIAA EQ for cutting, you could just follow the curve as defined. However, if you understand that the the 50-500 Hz range is already provided by the head, you can eliminate the 500 Hz point and just provide the 50 Hz and 2120 Hz portions of the curve. I did that with one version of my RIAA plugin that was meant for Presto heads (these heads also match the curve at low frequencies). If you don't do this you'll end up having to flatten the head velocity response for the full RIAA curve to work correctly. You will have cascaded two complementary EQ's which would be a waste.

The upshot of all of this is Bryan's head will have no issues cutting low frequencies (other than excursion limits). I hope that clears things up a bit.

Mark

[markrob]

Mark's point is a big deal with these drivers. The springs account for most of it. With the coil, spring and those brass rivets connecting the spring to another plate of metal that the coil attaches to, it comes in at a staggering 7.9 grams.
Bryan

The CUI driver I used has about 1 gram of moving mass. I see a bunch more turns of wire than on the CUI driver. While the extra wire adds mass, it also increase the force per amp of current. So the question is if this driver is better suited to the task in comparison to the CUI driver. I have some coming and will try to test them out and see.

Mark

[CHAPPERS]

Fascinating stuff gentlemen. I see a lot of effort going into removing mass, which I understand.

But can we clarify one thing?: The idea is not necessarily to make the lightest drivers, so much as
to make the most efficient and accurate right?

Like building a race car - surely it's more about power to weight ratio.

For example a Tesla P85S electric car weighs in at almost 5 tons. Does 0-60mph in 2.9 seconds!!
So moving mass becomes irrelevant when you have enough power and torque to move it.

Thoughts?

[opcode66]

So, Mark, it will take more or less energy to drive at 300 and below than at 1k? If you say more, we are saying the same thing. I never said it had zero response at 300 or lower.

The entire system is a perfect harmony. Weight, resonance, efficiency, etc. You do the best you can physically and let feedback and eq'ing take care of the rest.

[markrob]

Fascinating stuff gentlemen. I see a lot of effort going into removing mass, which I understand.

But can we clarify one thing?: The idea is not necessarily to make the lightest drivers, so much as
to make the most efficient and accurate right?

Like building a race car - surely it's more about power to weight ratio.

For example a Tesla P85S electric car weighs in at almost 5 tons. Does 0-60mph in 2.9 seconds!!
So moving mass becomes irrelevant when you have enough power and torque to move it.

Thoughts?

Hi,

Don't confuse the raw weight of the drivers with the moving mass.

This really is like building a race car. Power to weight ratio is very much a factor. There are several parameters that can be tweaked to produce a result. There isn't a single optimum design point. You have to make some choices and determine what you want to achieve within the physical laws that govern application.

Consider the system resonance. There are two variable parameters that determine the resonance point. The moving mass and stiffness (or compliance) of the suspension. If you want to raise the resonance point, you can either decrease mass or increase stiffness or some combination of both. The oscillation period of the resonance is proportional to the square root of the ratio of mass to stiffness. If you cut the mass by a factor of 4 you decrease the period by a factor of two (and increase the frequency by a factor of 2). Or, you could increase the stiffness by a factor of 4 for the same effect. What is the effect on the design of mass and stiffness?

http://en.wikipedia.org/wiki/Simple_harmonic_motion#Examples

At low frequencies, the head response is dominated by the stiffness. Beyond the resonance point, it is controlled by the mass. In the head, force is directly proportional to current. So as you increase the mass, you have to increase the current to the head to cause it to accelerate to the same levels. A the same time, power dissipation in the head is proportional to the current squared. So if you double the mass you need to dirve the head with 4 times the power to acheive the same effect. Clearly, you would like the system resonance to be as high as possible, but not at the expense of high mass and stiffness or you will need enormous power to drive the head. A the other end of the scale, if you imagine a head design with and extremely low mass and stiffness, you might be able to drive it on the bench with low power, but it would be very sensitive to the mechanical load under cutting conditions. That would make it hard to control from an EQ standpoint. In a feedback based design, you can get away with a very light system because the feedback corrects for the cutting load if it is within reason. This is just one example of the kind of trade-offs you have to make.

It helps to have good understanding of the physics involved, but you can design by the seat of your pants and come up with a head that works. The Neumann engineers were experts in the field looking to push the performance to the limits with cost no object. It is not a trivial task.

Mark

[CHAPPERS]

That all make sense, very helpful! Awesome, thanks Mark.

[Snug Music]

Hi Noobman,..
I'm not currently in my studio, but i have made my V-spring for my cutterhead with plastic from my tupper-lunchbox
I have not tried it yet, but that ought to work right? A picture of it I imagine here tomorrow purely!

Greeting Scotty

[Snug Music]

This peace i have cutted...