I need to lowering Qms of the driver, using Resistive braking - one coil shorted

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all,
I have a driver with dual voice coil.
I made a T/S param measure using this methode:
http://sound.westhost.com/tsp.htm
With voice coil in parallel (4+4=8ohm) I got the result, also with single voice coil (4ohm) and second voice coil left open,
but when I shorted the second voice coil, I couldn't get the result,
there was no voltage drop across the resistor to measure Fs, it just at the same level at all frequencies...what was wrong?

I need to lowering Qms of the driver, since it very high, using Resistive braking - one coil shorted as per this article:

"When you run a single voice coil, it has the nominal single-coil Qes of 0.8.
So, with the second coil open, you'd have a Qes of 0.8, and a Qms (unchanged) of 6.5, for a Qts of ~0.7.
Now, when you short the second (unused) coil, it essentially acts as an electric brake.
This will manifest itself by lowering the Qms to the Qes of a single coil: 0.8.
So, you'll have a driver Qes (of the coil you're driving) of 0.8, and a Qms (of the shorted coil) of 0.8, for a combined Qts of ~0.4.
One coil driven, other open, Qts = ~0.7. One coil driven, other shorted, Qts = ~0.4.

Dan Wiggins
President Avatar Audio"

taken from:
http://www.users.bigpond.com/bcolliso/driver-conversions.htm

is it wrong to measure single VC with second VC shorted? Should I use the result form single VC with second VC left open and using the theory above to get new Qms value?

thankyou,
Chris.
 

AKN

Member
Joined 2005
Paid Member
Hi christophorus,

Do you have any specific reasons for lowering just Qms?

High Qms is a sign of a low loss suspension.
If a manufacturers focus on terminating standing waves in cone a high loss surround is going to "eat" that energy and translate it to heat. Problem is that even wanted energy will be "eaten" to some degree. My personal opinion is that low Qms speakers will sound lifeless especially at low levels.
 
Hi Anders,
thank's for your reply, I need lower Qms because it seems too high, here my T/S param result:

#1-With dual voice coil in parallel (4+4=8ohm):
Fs=37,46Hz
Re=6 ohm
Qms=4,06
Qes=1,63
Qts=1,16

#2-With single voice coil (4ohm) and second VC left open:
Fs=35,36Hz
Re=3 ohm
Qms=10,33
Qes=8,28
Qts=4,69

first question:
is it normal Qts result from dual VC in parallel (#1) very different with single VC result (#2)?
second question:
with second VC shorted, it can not be tested for T/S param?
So to get it, I have to use #2 result and change Qms=Qes=8,28 and new Qts become 4,14.

Chris.
 

AKN

Member
Joined 2005
Paid Member
Hi,

#1-With dual voice coil in parallel (4+4=8ohm):
Fs=37,46Hz
Re=6 ohm
Qms=4,06
Qes=1,63
Qts=1,16



Maybe you just typed wrong, two paralleled 4Ohm coils makes 2Ohm.
Your Re indicates series connection if each coil has 4Ohm impedance.
I would say that your Qms is fine but instead Qes is way off. Shure about the correct phase to coils?

What woofer are you measuring?
 
I think your measurements are questionable.

It can be difficult to measure T/S parameters when one coil is shorted because the impedance gets so flat. I think you should measure T/S in series and parallel configurations and if Fs, Qts,Qes,Qms aren''t much the same you have made a mistake.

Assuming they are the same. use the RDO calculation in the article you have to calculate the parameters with a shorted coil.
 
Hi Anders,
you're right, it's series.

Hi Ron E,
so it must be measured in parallel or series, how about measure using one coil and the other coil left open?
And the important thing is: which ever measuring methode (parallel or series or 1 coil), the result should be not much different, right?

Hi Svante,
I did T/S param measure using digital multimeter, see my first post above.

Chris.
 

AKN

Member
Joined 2005
Paid Member
Hi,

This method is less sensitive to measurement accuracy, Rod also rightfully warns that his proposed method is sensitive to input errors, especially high Q drivers.

Regarding your multimeter and generator.
When no speaker connected, do you have same voltage reading at different frequencies, let say one octave above speaker fs, fs and one octave below fs of driver?

I'll ask once again, are you willing to share what driver you are testing or at least a picture of the driver?
 
Hi Richie,
here is picture of my driver, doesn't look too nasty, does it?
Anyway, I'm still waiting your measurement technique, as your promise.

Hi Anders,
When no speaker connected, it has almost the same voltage reading at different frequencies.

Picture #2 is a close up of the driver, as you can see, between spider and chassis, it has a gap (covered with perforated steel grille), how if I sealed it off so no air can in / out from there, can it reduce the Qms?

Chris
 

Attachments

  • skeleton-011a.jpg
    skeleton-011a.jpg
    11.7 KB · Views: 349
Why are you so fixated on reducing Qms?

THere is nothing wrong with a Qms of 4. There is nothing better in any way about a lower Qms. It does not need reducing for any reason other than trying to jam a woofer with a high Q into a smaller box. You can stretch a tightly woven material over the back of the driver to reduce Qms for this purpose. The typical way is to use 2" fiberglass batting stapled to the baffle. Several layers of cheesecloth may also work...

Blocking the screens will likely increase Qms and possibly cause noises during operation, not to mention reduce VC cooling, which will cause Re and Qes to rise ;).

Based on your initial comments you should not be confident enough in your measurements to even know what the actual value of Qms is. Please remeasure the voice coil in both series and parallel, if you do not get very nearly the same result in each configuration for Fs, Qts, Qes and Qms there is either something wrong with your measurements, or your driver does not have two identical coils.

The measurement technique I use (or used before using speaker workshop) consists of a smaller series resistor (5-10 ohm) along with measurements of the voltage across the resistor, the driver, and the combination. It is more work to find the parameters this way, but the technique allows extraction of impedance phase as well as magnitude, and it automatically corrects for frequency response issues in the meter, which are very common in today's digital meters.
 
Hi Bjorno,
I will try your method, it looks easy to do, thank's.

Hi Ron,
originally I want to build TL sub with this driver, but the Qts is too high even for sealed box, it needs a big box. So, I think reducing Qms is the easies way to lowering Qts. I read several time this paragraph, and now Bjorno bring it here:

"You want to use a woofer with a fairly low Qms (as compared with the Qes); for instance, you'd rather have a Qms value less than 10 times the Qes value. A low Qms means that the suspension is providing a great deal of the restoring (damping) force for the woofer, while a high Qms means that the woofer is relying on the amplifier to provide the damping electrically - which your amps will not do a good job of. If you're investigating kits or DIY, take a look at the Qms and Qes values, rather than just using the Qts."

or this:
I think the high Q will dominate the sound.
The effect will probably sound like a boom box with all bass notes sounding like they were of the same frequency.

I tend to concur with the other posters, High Q speakers were never intended for use with a cabinet.

__________________
regards Andrew T.



About blocking the screen, actually it doesn't sealed off completely, because at the middle-back of the magnet there is a hole 1" diameter, so I hope it doesn't effect VC cooling too much.

The important thing is, I should do the measurement once again to check wether it is wrong as you said.

chris.
 
christophorus,

About blocking the screen, actually it doesn't sealed off completely, because at the middle-back of the magnet there is a hole 1" diameter, so I hope it doesn't effect VC cooling too much

You’re right it doesn’t as long as you do not cover the rear of the magnet that should be of a vented type.

Se my simulation with a driver T/S you submitted in another post the effect when lower Qms from 8.36 to 1.

As long as the cone of the driver is off a stiffer type and the applied damping material is even distributed and secured not to move at all, then this method will work very well with drivers with high compliance.

If you already have an enclosure for your driver, why not practice both methods, i.e. simply first apply Qm lowering damping and if not sufficient; short one coil with resistors between 0 to about the Re value for achieving different Qm-values.

Test with 6.8, 3.9 2.2 and 1 to =0 Ohms while just listening to music on your speaker, if you don’t have measurement equipment, and of course, post your result and opinion of the results here.

b

1(1)
 

Attachments

  • lowering qm.gif
    lowering qm.gif
    83.8 KB · Views: 318
christophorus said:

"You want to use a woofer with a fairly low Qms (as compared with the Qes); for instance, you'd rather have a Qms value less than 10 times the Qes value. A low Qms means that the suspension is providing a great deal of the restoring (damping) force for the woofer, while a high Qms means that the woofer is relying on the amplifier to provide the damping electrically - which your amps will not do a good job of. If you're investigating kits or DIY, take a look at the Qms and Qes values, rather than just using the Qts."

The entire quoted argument above is spurious. Nearly all drivers have a Qes that dominates, even when an amp has high output impedance. Furthermore, Qms doesn't describe a "restoring" force as stated in the quote above - it is purely a dissipative element. Qms has NO place in box design, IMO. In your case, shorting a coil will never reduce Qts below the level it had when series or parallel. Basically when you short one coil, Qes doubles and Qms ~= Qes original and Qts stays relatively unchanged.

If you want to modify response, a linkwitz transform would be more appropriate, you can use a LT in a TL ;) You'd want to add an infrasonic filter, though. You could also just use a notch filter at resonance...

christophorus said:
or this:
I think the high Q will dominate the sound.
The effect will probably sound like a boom box with all bass notes sounding like they were of the same frequency.

About blocking the screen, actually it doesn't sealed off completely, because at the middle-back of the magnet there is a hole 1" diameter, so I hope it doesn't effect VC cooling too much.

The screen is an added cost, they wouldn't likely use it unless it were necessary. VC cooling is not the only issue, there can be air rushing noises caused by the narrow gap. Whatever you do to block the screens, make sure it is reversible ;)

All this silly gymnastics could be avoided if one just used a suitable driver in the first place... Sell it to some dumb kid who doesn't know any better.
 

AKN

Member
Joined 2005
Paid Member
hi christophorus,

In all this I want to question your driver and specifically the manufacturer.
Has the manufacturer published any specs, especially TS parameters for your woofer, if so what specs?
If sold in retail and no specs avalible there really is a "nasty" driver.
Often we can find drivers that are really cheep. Usually the manufacturer slab the same motor assembly on every size in a series of woofers resulting in increasing Qts as size goes up. The small ones may be useful but the larger ones has Q values totally out of range.
 
Hi Bjorno,
thank's for your very helpfull simulation, it's look promising, but with your acoustic damper, could it really reduce Qms from 8.36 to 1? because when I sealed off the screen, Qms only drop to 7.83 (~ -0.5), but Qes goes up! around +0,4 , so overal Qts didn't change too much...
Does sealed of the screen not enough?

Would you mind giving me a drawing/sketch of your acoustic damper, maybe 1 cross section of the driver can explain better.

Hi Ron,
shorting a coil will never reduce Qts below the level it had when series or parallel. Basically when you short one coil, Qes doubles and Qms ~= Qes original and Qts stays relatively unchanged.
now, I confuse with your last statement above, it's opposite with what Dan Wiggins said as on my first post.

Really? Linkwitz transform can use for TL? Doesn't it for sealed box only?

Maybe you're right, maybe it's better for garage door stopper (quite heavy) :) , but before it goes there, I want to know how far this thing can do with some treatment to achieve what I want.

Hi Anders,
there is no T/S spec, only freq. range, magnet weight and bla bla blaa...
 

AKN

Member
Joined 2005
Paid Member
Hi christophorus,

I agree with Ron E here regarding Q values.

I have myself never tried shorting one coil, anyway I did a quick test setup to verify.
Clio as measurement tool using internal amp.
Object: Sinus 16W10.3 6.5" dual VC 8Ohm midbass in freespace.

Connection Qms Qes Qts
Series 2.78 0.38 0.34
Parallell 2.79 0.38 0.34
One open 2.82 0.80 0.62
One shorted 0.87 1.36 0.53
 

AKN

Member
Joined 2005
Paid Member
Hi,

One more comment.

In your quote from Dan Wiggins:

"When you run a single voice coil, it has the nominal single-coil Qes of 0.8.
So, with the second coil open, you'd have a Qes of 0.8, and a Qms (unchanged) of 6.5, for a Qts of ~0.7.
Now, when you short the second (unused) coil, it essentially acts as an electric brake.
This will manifest itself by lowering the Qms to the Qes of a single coil: 0.8.
So, you'll have a driver Qes (of the coil you're driving) of 0.8, and a Qms (of the shorted coil) of 0.8, for a combined Qts of ~0.4.
One coil driven, other open, Qts = ~0.7. One coil driven, other shorted, Qts = ~0.4."

He never compares both coils connected to amp with one connected and one shorted.
Intead he states what will happen when you connect one coil and short the other coil. So you will not lower Qts with one shorted coil beyond what you will have with both coils connected.
Correct though that Qms will be lower, but for what use when Qes goes up?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.