diyAudio (
-   Multi-Way (
-   -   New 3-way project - "The French Touch" (

yov 7th February 2013 01:47 PM

1 Attachment(s)
I added the bafle step attenuation and also sensitivity correction to the tweeter and the midrange.
I added 6 Ohm to the tweeter and 8 Ohm to the mid.

pguerin 7th February 2013 01:52 PM

1 Attachment(s)
Hi yov,

To design a crossover I think it's better to simulate it.
You can find measures of the Davis 16 on part express :

Davis 16gklv6m 6-1/2" Kevlar Midbass 297-568

A simulation tool wich can be good as a first try is PCD :

Loudspeaker Design Software

You should also take a look at the Davis MV15 crossover :

I see one problem on your design, with a passive crossover the woofer is not enough sensitive.

dumptruck 7th February 2013 03:41 PM

Since you have engineering training, you will not have too much trouble learning how to use simulation software and measurement hardware/software. Since you are using expensive drivers, I think you can afford a calibrated microphone and USB interface/preamp (~130 Euro). It will be best if you also make a cable or buy a device to measure impedance, although you can trace the mfg. curves with the spltrace program for impedance instead and do okay (the ported midrange would be a problem, though). I suggest you do this right ;).

PeteMcK 7th February 2013 10:03 PM

re: - "a very complex computation only possible to make through simulation software" - personally, I think simulation is over-rated. All that is needed is an impedance plot of the driver & the knowledge of how to use it...

PeteMcK 7th February 2013 10:08 PM

hmmm, the kevlar driver is smoother than I thought it would be, if this graph is to be trusted you may get away with 1st order:

dumptruck 7th February 2013 10:21 PM


Originally Posted by PeteMcK (
hmmm, the kevlar driver is smoother than I thought it would be, if this graph is to be trusted you may get away with 1st order:

Based on those impedance plots you were just talking about ;), the Davis driver starts to break up around 1.2kHz, and without heavy smoothing on the graph, it probably looks a bit nasty above that.

tvrgeek 8th February 2013 01:15 AM

Yup, you waste a lot of power padding the mid and tweeter. That's how it works.

You are going in the right direction. Remember, the most valuable measurement tool is your ears. All the rest of our equipment is to make it easier and quicker to identify problems. Sims can be helpful, but the don't do the work for you. You still have to understand what they are telling you and make the decisions yourself. There is only one right answer in speaker building: "It depends". At least when we mess up a speaker metric to inch measurement, we don't miss Mars, and when we totally mess up a crossover, the roof does not collapse. We enjoy "safe" engineering.

I prefer sealed enclosed for my mids because it helps the transient response. If you could do CSD (Cumulative Spectral Decay) measurements, you will see that sealed chambers damp the driver at and below it's resonance far better. Plug the mid into WinISD and see what it tells you for a sealed chamber.

I was not really sure about the breakup from the plots, it has one bad one high, but it looks like it may have some issues lower. This is a real difficult problem, finding a mid that get up to where you want it.

I am sure it varies for everyone, but it takes me many months to do a new speaker. From idea, to getting the drivers and measuring then, models,mock ups, and final voicing. Every set I set out to learn something new. The last few have been about the trade offs caused by position of the drivers on the baffle and how that effects the Q of the initial diffraction ripple. I was quite surprised how big a difference 3/4 inch vs 1/2 inch radius makes, and the additional smoothness by doing a radius on the back edges.

yov 8th February 2013 06:36 AM

3 Attachment(s)
I have changed again the filter...
I want to use a lower xover frequency for the sub. I used 250 Hz instead of 300 Hz because I think the sub cannot really go much higher than 200 Hz and because the mid can do much lower than 300 Hz in a vented chamber.
I changed the high cutoff frequency from 5000 to 4000 Hz because the mid doesn't go as high as I thought.
I also used a 3db Attenuation for the mid and tweeter instead of 6dB because it's too frustrating to loose so much power in the resistors (power is money !)... Also the frequency range of the sub is quite short (30-250Hz).

yov 8th February 2013 06:45 AM

4 Attachment(s)
Ok, vented vs closed on the mid.
This is what I get in WinISD.

- Using a vented chamber I have a cut-off frequency of 70 Hz.
- Using a closed chamber I have a cut-off frequency of 150 Hz.

yov 8th February 2013 08:25 AM

3 Attachment(s)
Ok, did the same with the sub...
I think the curves talk for themselves...
Day and night !

All times are GMT. The time now is 02:24 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 17.65%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio