I've taken some of the advice about getting started designing crossovers and am going to give it a go with these two drivers.
I'm trying to follow Max's tutorial on using Jeff's FRM and have run into a little snag. I created the .frd file from Zaph's frequency response curve and SPL and imported it into FRM. I entered the T/S parameters and I get a double hump in the impedance curve instead of the single hump from Zaph's testing.
Looks like it's caused by your enclosure + vent. Notice impedance is low at your tuning point (45Hz) and rises after that.
That said, those are the default parameters. You want to change them!
Try a 10-12l box with tuning around 50 or 52 Hz. This will give a slight rise in the 100-200Hz region that will add some warmth, which, combined with the distortion profile on that driver, may reduce the need for more than 3dB of baffle step compensation, gaining you some sensitivity.
Except for adding the T/S params, I just took the defaults for everything else and then started changing the two inductance params per Max's tutorial. I didn't know what else to do at that point.
Hey John,
I tried this combination last summer, and it works pretty well. There is a little problem though, you can't place the little ND20 close enough to the 704 for a 3500 Hz XO without a little lobing (this really is minor). I went ahead and crossed mine at about 3500 Hz.
The rest of the problem (still minor) is forcing the 704 to carry on to 3500 - 4000 Hz.
I'm converting from the ND20s to Seas TD27TDFCs @ about 2500 Hz hoping to smooth the upper midrange a little. The ND28Fs might be a little better match for the 704s.
The ND20 can do a 3khz cross no problem - the somewhat low mechanical power handling of the W5-704 means that you could probably used the ND20 even at 2500hz and not run into problems. I'd aim for a 3000hz XO point with these two drivers.
Again, I'd recommend the "ND20FA" version over the "ND20FB" version, as the FA has a lower FS and much better distortion on the low end of the passband (2khz). And in reality you only gain about 1/8" in terms of CTC spacing by using the rear mount.
I love this combo, but I broke all the rules. I did a very simple 6 db at 3900 Hz with an l-pad after reading some comments on this board about the idea. I could not be happier & friends have told me they just "disappear". The Baltic Birch came out nice also - my first try with that.
Last edited by Steve Henry; 04-10-2011 at 09:34 PM.
That still results in a basically LR2 design, which is not too unusual. Lots of cheap tweeters in big-box speakers (with far lower build quality than the dayton) use only a capacitor in front of them - and the uF value is usually silly.
I've been playing around with PCD and, using the "textbook" crossover values with a 1st order network, here is the system response. The problem is the size of the cap for the tweeter is humongous! It is something like 1,562 micro farads!
I've been playing around with PCD and, using the "textbook" crossover values with a 1st order network, here is the system response. The problem is the size of the cap for the tweeter is humongous! It is something like 1,562 micro farads!
How do I get around using that big of a cap?
You might as well not be using a cap at all on the tweeter, if the one you need is that large.
You might as well not be using a cap at all on the tweeter, if the one you need is that large.
I'm not sure what crossover point you were trying, but for 3900 Hz, which is what I did, I used a 6.8 uF and a 2 ohm/12 ohm L-pad. Now I did not measure these results - I had read some ideas others had posted and tried them myself. I was very happy with the results for the $ spent.
Yes -a .33 mH - I just tried to take advantage of the natural roll offs of the drivers. It was a gamble, I know. But sometimes you just have to listen to the music.
For what it's worth Jay helped me with an XO for an active MTM using a pair of 704's and a ND20 at 3khz and these turned out quite well...no lobing that i could detect and that tweeter is amazing at 3khz and up. I may try another pair in the fall with the ND20 mounted in a baffle cut waveguide.
I'm not sure what crossover point you were trying, but for 3900 Hz, which is what I did, I used a 6.8 uF and a 2 ohm/12 ohm L-pad. Now I did not measure these results - I had read some ideas others had posted and tried them myself. I was very happy with the results for the $ spent.
Steve,
Was this with the ND20FA or ND20FB version? I have all these parts lying around and I thought this sounded like a good use for them.
I've taken some of the advice about getting started designing crossovers and am going to give it a go with these two drivers.
I'm trying to follow Max's tutorial on using Jeff's FRM and have run into a little snag. I created the .frd file from Zaph's frequency response curve and SPL and imported it into FRM. I entered the T/S parameters and I get a double hump in the impedance curve instead of the single hump from Zaph's testing.
Attached is a screen capture of this information.
How do I get there from here?
John in the graph you show the impedance response of the box and vent. You need to add the TB woofers impedance and click button below the graph that says overlay the impedance. Then adjust the Le Coeff and Le Expon to get the blue box line to match the woofers natural rising response. The TB W5 has a very rising impedance response and if you don't account for it your results could be off by quite a bit.
What the blue line shows you is the impedance effect of the box which only really occurs at lower frequencies around Fb and the port tuning frequency. At higher frequencies the woofer's impedance is controlled solely by the motor this is why it is important to overlay it and adjust the blue line to follow it. If you don't your saved impedance could be off significantly.
Dave
If you can read this, thank a teacher.
If you are reading it in English thank a Veteran.
Yes -a .33 mH - I just tried to take advantage of the natural roll offs of the drivers. It was a gamble, I know. But sometimes you just have to listen to the music.
I used it first order with an Alesis tweeter crossing at around 4000Hz and used a .85mh inductor to get enough baffle step compensation to keep it from sounding thin and shallow in the mid/bass region. It worked quite well. Unless you are mounting it on the wall I would think .33mh would be for too small to get it roll off even with it's natural rolloff. The W5 has a sharply rising impedance so unless you are compensating it .33mh won't have much affect on it's natural response at all.
Dave
If you can read this, thank a teacher.
If you are reading it in English thank a Veteran.
I used it first order with an Alesis tweeter crossing at around 4000Hz and used a .85mh inductor to get enough baffle step compensation to keep it from sounding thin and shallow in the mid/bass region. It worked quite well. Unless you are mounting it on the wall I would think .33mh would be for too small to get it roll off even with it's natural rolloff. The W5 has a sharply rising impedance so unless you are compensating it .33mh won't have much affect on it's natural response at all.
Good point, Dave. I should have been more clear that mine were used desk top against the wall.
TangBand W5-704D + Dayton ND20FB
![$vboptions[bbtitle]](images/techTalkHeader-leftGray2.jpg)
![$vboptions[bbtitle]](images/techTalkHeaderRepeatGray3.jpg)
Reply With Quote
Originally Posted by johngalt47
