diyAudio (
-   Multi-Way (
-   -   Crossover testing without modelling (

Cousin Billy 30th December 2011 08:18 PM

Crossover testing without modelling
This is not a hypothetical question, but I want to know why this is a bad idea.

All observations are from the exact measurements in ARTA.

I have a tweeter. This tweeter has a bump in the 2K - 4KHz region. After this everything levels out. During my testing, and reading backwards, a capacitor of 2.0uF started the drop at 4KHz, everything levelled out till 2600Hz, at which point the slope continued down quite quickly.

My midrange also has a bump, in the 2K - 3KHz region. By applying a 1.8mH inductor, the bump was diminished, with the big drop occuring at 2400Hz.

No phase measurements have been taken.

Measurements at 1m show a slight anomoly at 2500Hz, but everything sounds great. Maybe my hearing is wrong.

My first thought is that the tweeter has a theoretical crossover starting at 4000Hz, and the midrange has it's crossover starting at 2000Hz. This leads me to believe that I have thrown modelling out the window.

Is it more correct to pick a crossover point, and use the appropriate values, or is what I have done correct?.

AllenB 30th December 2011 11:17 PM

I have a theory on each of these cases but I'd prefer to investigate the drivers before I say anything. Is there a chance you could post what they are?

Pano 30th December 2011 11:46 PM


Originally Posted by Cousin Billy (
During my testing, and reading backwards, a capacitor of 2.0uF started the drop at 4KHz, everything levelled out till 2600Hz, at which point the slope continued down quite quickly.

I don't understand what you mean by this. A shelf? You may have to post a graph.

From reading the rest, it sounds like you have a crossover point circa 2500Hz with a slight underlap. Nothing wrong with that.

Cousin Billy 2nd January 2012 06:17 PM

Sorry for the delay, just got back.


The tweeter is the Accuton BD 25-6-034. 6 Ohm version. 'In system' measurements have the bump a little wider then the manufacturers graph, but only a little.

The midrange is the Accuton C173-11-191E. 11 Ohm version. At the Accuton website the C173-6-191E is almost the same.

It is also important to note that the driving amp is an OTL with an output impedence of 2.3 Ohms. I think the driving amp has a little to do with the somewhat odd crossover component values.


The crossover point is as you said, ~2500Hz.

Am I safe in assuming there is not a problem with starting the tweeter attenuation at 4KHz, as long as the final point is good?.

I should also say that; the tweeter has an LCR for Fs, and an RC for impedence rise. The midrange has an RC as well.

Pano 2nd January 2012 07:12 PM

What really matters is the final acoustic response of the drivers. We often have to use odd-ball electrical slopes to achieve good working acoustical slopes. Without measuring it can be hard to tell, but as you aren't hearing any problems, you may have hit it.

Dave1027 2nd January 2012 08:14 PM

How about low passing the mid at 2k and high passing the tweet at 4k or something like that? There is no law that says both filters must cross at the same frequency.

Pano 2nd January 2012 08:48 PM

There are some laws that say the acoustical slopes need to cross at the (aprox) same frequency if you want the frequency responses to sum flat.

Electrical slopes are another matter.

sreten 3rd January 2012 01:18 AM


All that matters is the final acoustic slopes and acoustic c/o point. Far from
throwing modelling out of the window, this is the correct way of modelling.

Calculator/Cookbook electrical filters are not modelling, they are a gross
oversimplification of the reality, a model is only as good as how correct
it is, anything reflecting reality is a better model.

rgds, sreten.

Cousin Billy 3rd January 2012 01:40 AM

Thank you all.

It does sound great.

Can't wait till the next project.

AllenB 3rd January 2012 02:05 AM

I notice your tweeter impedance has a secondary peak in the 2-3kHz region. Your LCR may flatten it in a broadband fashion, but this small bump will still be evident. You would need a more sophisticated compensation if you had a need to flatten it.

Your tweeter's response also has a peak in the 3-4kHz region which seems to tell the rest of the story.

As far as whether it will be a problem, I don't think so. As long as you can match it to your woofer, and the capacitor is enough to keep your tweeter within its limits.

As far as the woofer is concerned, it does have a drop in response at around the point you mentioned. It also has a small impedance peak dropping off at that point. If it works for you, then go with it but you are pushing the breakup region. There could be other issues going on behind the on axis response.

All times are GMT. The time now is 01:29 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