Please help, really stuck on phase :( - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 31st January 2013, 04:06 AM   #1
diyAudio Member
 
Join Date: Dec 2003
Location: Massachusetts
Default Please help, really stuck on phase :(

Hi Folks,

I tried finding the answer to this question but couldn't and asked within another thread but didn't get an answer.
I'm really stuck on this basic question "can software (I'm using PCD7.0) accurately predict the phase response of a system".

The reason I'm asking is I designed my speaker, 2 way, Vifi tweeter and Peerless woofer, I used PCD to design a linkwitz-riley 2nd order. I got it to what I thought was the best point for a good transition, 1.33mH/5.1uF on the woofer and 10uF/0.68mH on the tweeter with a zobel on the woofer and an optional L-pad on the tweeter. (sorry I can't figure out how to print out a schematic of the crossover)

On examining the phase, reversing the phase on the tweeter seemed to give the best response, the phase was a slow slope from low to high frequency with no abrupt changes. Attached on the first screen grab is my crossover. Grey is phase sum, orange and blue are driver phases.

However I built the crossover a day or so ago and tried out the speakers, there is bass and treble but it seems like not much in the middle (maybe I should run for a week to burn in?)

Soooo I hooked up an oscilloscope to the output of the tweeter and woofer, when I run a sine wave through the system at the predicted crossover frequency the two responses are almost completely out of phase, so it looks like I'm getting destructive interference which explains why my speakers don't seem to have much 'middle'.

First of all why is this? does this mean the PCD tool cannot actually predict the correct phase response?

I've heard people on the forum talk about flipping the terminals around to try it both ways and see which sounds best, but why do this? If the software can model the response surely it is better than trial and error?

I have tried searching on here and google to find an answer on what is considered a 'good' phase response on a crossover (obviously flat is best but not usually possible) but can't see to find this. Any help is appreciated!

Chris
Attached Images
File Type: jpg 2nd order first attempt.jpg (134.7 KB, 230 views)
  Reply With Quote
Old 31st January 2013, 04:31 AM   #2
diyAudio Member
 
Inductor's Avatar
 
Join Date: Feb 2006
Location: Cascais
Quote:
"can software (I'm using PCD7.0) accurately predict the phase response of a system".
What kind of data you inputted in the simulator?
Did you used the frd/zma's?
(I don't know that version/simulator) But does it work with phase and drivers phase?
The easy answer/fix to your problem is just invert the connection to have same phase. If you want to model phase, as you should with any good simulator, you need to insert the phase data in the model and have it compound for output.
You can not simulate a good passive crossover if you are not using the differences of phase between drivers, and components, plus their time alignment and position in space/baffle in relation to each other.

Last edited by Inductor; 31st January 2013 at 04:40 AM. Reason: phase simulation
  Reply With Quote
Old 31st January 2013, 06:26 AM   #3
system7 is offline system7  United Kingdom
diyAudio Member
 
system7's Avatar
 
Join Date: Aug 2008
You don't use Zobels on paper woofers in practise. They only have a place with plastic cones and low order filter rolloff.

You've come a long way from Veeper's 1.5mH/10uF bass suggestion, never mind the tweeter circuit for a regular 6 ohm DC dome:
Veeper TM Monitor

The Peerless 830875 0.41mH Nomex woofer is a tricky unit with its 4kHz peak, but I don't think a Zobel is helping at all. It's probably rolling off the midrange too much.
PEERLESS-NOMEX-164

You'd probably do well to get more conventional here. FWIW, 2nd. order circuits on flat baffles always have the drivers in phase. That's how it is. Neither circuit below appeals to me much, but should be OK as a starting point. The object is to align the phase of the two drivers as best you can.
Attached Images
File Type: jpg bart_dood_1.JPG (22.8 KB, 218 views)
File Type: jpg bart_dood_2.JPG (23.2 KB, 218 views)
__________________
Well, there it is! Best regards from Steve in Portsmouth, UK.
  Reply With Quote
Old 31st January 2013, 08:39 AM   #4
diyAudio Member
 
Join Date: Aug 2012
What's probably happening is that it is the acoustic offset between the drivers causing the phase difference.

Unless the software has an explicit input to correct for this it assumes that both drivers are radiating from the same acoustic plane, as they are for instance for the ideal L-R characteristic.
rcw
  Reply With Quote
Old 31st January 2013, 08:55 AM   #5
AllenB is offline AllenB  Australia
diyAudio Member
 
Join Date: Oct 2008
Good phase? Start by lining the drivers up around the crossover, this is the main issue.

Phase will vary as you move around and since the drivers are not in the same location as each other they'll be 'in' or 'out' at different places. Manipulating this is really on another level though and I was just hoping to put the single 'phase plot' in to perspective. I like to move a mic up and down in front of a speaker just to double check I haven't messed anything up. (Do a search on Speaker Dave and pink noise for a non-mic method.)

PCD can determine phase to a point. If you use your drivers within their limits and take in to account their distances from you then you may get a usable design, but I'd rather measure them.
  Reply With Quote
Old 31st January 2013, 09:25 AM   #6
just another
diyAudio Moderator
 
wintermute's Avatar
 
Join Date: Aug 2003
Location: Sydney
Blog Entries: 22
Sure software can sim the phase. Please see below the simulated (black) and measured response (blue) for my 4th order bessel crossover. (crossover at 2.8Khz)

Note that the actual measurement was done months after the original individual driver measurements at a slightly different distance, in a different location. It was also done without as much care as the original measurements. If I were to measure the individual drivers again, and then without moving anything, measure the final crossover, and then sim and compare, the differences would be much smaller.

Tony.
Attached Images
File Type: png sim_vs_real_phase.png (24.5 KB, 199 views)
__________________
Any intelligence I may appear to have is purely artificial!
Some of my photos
  Reply With Quote
Old 31st January 2013, 02:53 PM   #7
diyAudio Member
 
Join Date: Dec 2003
Location: Massachusetts
Quote:
Originally Posted by Inductor View Post
What kind of data you inputted in the simulator?
Did you used the frd/zma's?
(I don't know that version/simulator) But does it work with phase and drivers phase?
The easy answer/fix to your problem is just invert the connection to have same phase. If you want to model phase, as you should with any good simulator, you need to insert the phase data in the model and have it compound for output.
You can not simulate a good passive crossover if you are not using the differences of phase between drivers, and components, plus their time alignment and position in space/baffle in relation to each other.
I got the driver responses from here:

Driver FRD & ZMA files
  Reply With Quote
Old 31st January 2013, 02:54 PM   #8
diyAudio Member
 
Join Date: Dec 2003
Location: Massachusetts
Quote:
Originally Posted by system7 View Post
You don't use Zobels on paper woofers in practise. They only have a place with plastic cones and low order filter rolloff.

You've come a long way from Veeper's 1.5mH/10uF bass suggestion, never mind the tweeter circuit for a regular 6 ohm DC dome:
Veeper TM Monitor

The Peerless 830875 0.41mH Nomex woofer is a tricky unit with its 4kHz peak, but I don't think a Zobel is helping at all. It's probably rolling off the midrange too much.
PEERLESS-NOMEX-164

You'd probably do well to get more conventional here. FWIW, 2nd. order circuits on flat baffles always have the drivers in phase. That's how it is. Neither circuit below appeals to me much, but should be OK as a starting point. The object is to align the phase of the two drivers as best you can.
Thanks, I'll try putting them in my sim and seeing what the output looks like

I didn't realize zobels weren't used under certain conditions, in my sim it seemed to help flatten the frequency response, but I'll take it out.

Last edited by bart_dood; 31st January 2013 at 02:58 PM.
  Reply With Quote
Old 31st January 2013, 02:57 PM   #9
diyAudio Member
 
Join Date: Dec 2003
Location: Massachusetts
Quote:
Originally Posted by AllenB View Post
Good phase? Start by lining the drivers up around the crossover, this is the main issue.

Phase will vary as you move around and since the drivers are not in the same location as each other they'll be 'in' or 'out' at different places. Manipulating this is really on another level though and I was just hoping to put the single 'phase plot' in to perspective. I like to move a mic up and down in front of a speaker just to double check I haven't messed anything up. (Do a search on Speaker Dave and pink noise for a non-mic method.)

PCD can determine phase to a point. If you use your drivers within their limits and take in to account their distances from you then you may get a usable design, but I'd rather measure them.
Yeah I can absolutely hear high and low spots. I took one of my speakers and reversed the tweeter around (in phase with the woofer) when sat quite close I can move my head around and at the crossover frequency the output goes up and down, its quite amazing how low it drops in some places.

I also made an ESP preamp mic with high quality components for measurements but I don't have one of those panasonic flat frequency response electrets for it yet, I'll get one ordered and I should be able to take some basic measurements.

Last edited by bart_dood; 31st January 2013 at 03:04 PM.
  Reply With Quote
Old 31st January 2013, 05:27 PM   #10
ra7 is offline ra7  United States
diyAudio Member
 
Join Date: Feb 2009
Blog Entries: 1
There is a place in PCD to input driver offset. Whether you put the offset there or use the measurements to capture the offset, the software needs to have it in there to predict phase correctly. What this means is that, sometimes you can capture the driver offset through the measurement. HolmImpulse lets you fix the time on one measurement and then draws the phase of subsequent measurements based on the fixed one. This is an easy and accurate way to do it. If this is not done, then you need to enter the offset in PCD.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
2 phase DDS signal generator to run 2 phase AC sync motors Giordano Analogue Source 28 4th January 2013 06:40 PM
Acoustic phase, electric phase,....phase ermes Multi-Way 6 27th November 2010 06:17 PM
Stuck in a dip ttruman Multi-Way 77 6th June 2008 12:51 PM
Amp stuck on audiobahnkid592 Car Audio 1 8th July 2007 11:31 PM


New To Site? Need Help?

All times are GMT. The time now is 01:23 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2