Update on Dreydels (ver. 2) *PIC*

[Jeff B.]

Provided Link: The Dreydels


For anyone who has built, or is planning to build my little Dreydels, I now have a version 2 crossover for these two drivers. When I first voiced the Dreydels I intentionally set them a little recessed in the upper midrange, with a bit of a broad shallow dip centered around 2kHz, below 500Hz the response rose gradually a little more than is typical for baffle step compensation. The result was a small speaker that could be placed well away from the walls and still sound very full.

Then I began work on my LS3/5a clones, attempting to duplicate the frequency response of this classic speaker in my own little clone. The LS3/5a’s have a wonderful midrange balance and the difference in the balance between these two speakers is that the Dreydel's had more energy in the 300 - 500 Hz range (where the LS3/5a has a dip), but the LS3/5a's were flatter above1kHz by not having any dip in this region.

After listening for quite a while to my LS3/5a clones I moved the Dreydels back in for a session. After a while I came to the conclusion that I much preferred the flatter voicing of the LS3/5a clone to the laid-back midrange of the Dreydel's balance. As I result, and because I never know when to quit tinkering, I decided to revisit the Dreydel's crossover.

I found that a very small change accomplished my goals very nicely. So, here it is:

The low pass section remains the same with no change in the values of the original crossover except that an additional .35mH inductor is placed between the crossover and the driver. This changes it to a third order circuit with some small changes in the electrical transfer function.

The addition of this inductor does three things; first, it lowers the region between300 - 400 Hz about 0.7 dB; second, it raises the output near the crossover knee by about 1.2 dB; and third it lowers the high frequency break-up mode of the RS150 by an additional 8 dB. In addition to this inductor I had already changed the original series resistor on the tweeter from 7 ohms to 6 ohms.

The combination of these two changes results in a response that measures much flatter overall. The broad dip centered at 2kHz is gone and the rise between 300Hz and 500Hz is reduced. The overall response curve is now very flat from 400Hz to 20kHz on the tweeter axis. Below 400Hz the response still rises some but not as much till you get down to 200Hz.

Below this point there remains a rise of about 2.5dB in the upper bass. Again, this was part of the design goal. This hump at centered at 125Hz coincides with the one in LS3/5a and gives the little speaker the impression that there is more bass energy than there really is. Granted the changes I have described are not very great (1.2 dB in the presence region), they are still noticable in the voicing and result in a speaker that seems a little better balanced now, and a little less laid back. Vocals are not quite as thick now.

Playing by themselves they really surprise you that a little 5 liter speaker with a 4.5" cone could sound so deceptively full-range at times, and still have a natural sounding vocal/midrange reproduction too.

After some extended listening and measurements I felt this new version 2 crossover was worth posting for anyone interested. Below is the new version 2 crossover schematic, and below that is a link to the Photobucket with several more graphics that can be viewed.

Jeff B.

[Jeff B.]

Woofer and crossover:

[Jeff B.]

The Dreydel:

[Jim Holtz]

Thanks Jeff! This is a very nice looking project that doesn't cost a fortune to build.

Jim

> For anyone who has built, or is planning to
> build my little Dreydels, I now have a
> version 2 crossover for these two drivers.
> When I first voiced the Dreydels I
> intentionally set them a little recessed in
> the upper midrange, with a bit of a broad
> shallow dip centered around 2kHz, below
> 500Hz the response rose gradually a little
> more than is typical for baffle step
> compensation. The result was a small speaker
> that could be placed well away from the
> walls and still sound very full.

> Then I began work on my LS3/5a clones,
> attempting to duplicate the frequency
> response of this classic speaker in my own
> little clone. The LS3/5a’s have a wonderful
> midrange balance and the difference in the
> balance between these two speakers is that
> the Dreydel's had more energy in the 300 -
> 500 Hz range (where the LS3/5a has a dip),
> but the LS3/5a's were flatter above1kHz by
> not having any dip in this region.

> After listening for quite a while to my
> LS3/5a clones I moved the Dreydels back in
> for a session. After a while I came to the
> conclusion that I much preferred the flatter
> voicing of the LS3/5a clone to the laid-back
> midrange of the Dreydel's balance. As I
> result, and because I never know when to
> quit tinkering, I decided to revisit the
> Dreydel's crossover.

> I found that a very small change
> accomplished my goals very nicely. So, here
> it is:

> The low pass section remains the same with
> no change in the values of the original
> crossover except that an additional .35mH
> inductor is placed between the crossover and
> the driver. This changes it to a third order
> circuit with some small changes in the
> electrical transfer function.

> The addition of this inductor does three
> things; first, it lowers the region
> between300 - 400 Hz about 0.7 dB; second, it
> raises the output near the crossover knee by
> about 1.2 dB; and third it lowers the high
> frequency break-up mode of the RS150 by an
> additional 8 dB. In addition to this
> inductor I had already changed the original
> series resistor on the tweeter from 7 ohms
> to 6 ohms.

> The combination of these two changes results
> in a response that measures much flatter
> overall. The broad dip centered at 2kHz is
> gone and the rise between 300Hz and 500Hz is
> reduced. The overall response curve is now
> very flat from 400Hz to 20kHz on the tweeter
> axis. Below 400Hz the response still rises
> some but not as much till you get down to
> 200Hz.

> Below this point there remains a rise of
> about 2.5dB in the upper bass. Again, this
> was part of the design goal. This hump at
> centered at 125Hz coincides with the one in
> LS3/5a and gives the little speaker the
> impression that there is more bass energy
> than there really is. Granted the changes I
> have described are not very great (1.2 dB in
> the presence region), they are still
> noticable in the voicing and result in a
> speaker that seems a little better balanced
> now, and a little less laid back. Vocals are
> not quite as thick now.

> Playing by themselves they really surprise
> you that a little 5 liter speaker with a
> 4.5" cone could sound so deceptively
> full-range at times, and still have a
> natural sounding vocal/midrange reproduction
> too.

> After some extended listening and
> measurements I felt this new version 2
> crossover was worth posting for anyone
> interested. Below is the new version 2
> crossover schematic, and below that is a
> link to the Photobucket with several more
> graphics that can be viewed.

> Jeff B.

[paul_n_oneal]

Looks nice and I'm sure it sounds just as nice.