Advice on speaker mod, frequency response tests

sampleton

Member
2013-02-11 5:19 pm
Hi folks, im having some confusion with my test results. my new results are showing different responses compared to old tests carried out, all testing conditions were carried out in similar conditions such as; an SPL reading off 85dB at testing mic during all tests, and all carried out in the same isolation booth.

This picture was from the first tests carried out ont he original speaker with each driver isolated and shows rough crossover points at 1k and 3k, although A weighting was mistakingly applied during testing.
originaldriversisolated_zps37accec7.png
[/URL][/IMG]

This picture was from the next set of tests on the exact same speaker under similar conditions with no frequency weighting applied. although the crossover points seem similar, the spiking pattern on the high frequency graph has me stumped, would i be correct in saying that this is comb filtering from the tweeter horn?
ORIGINALSPEAKERDRIVERSISOLATED_zps83dd6b05.png
[/URL][/IMG]

The next picture shows the frequency response of the new modified speaker, with each driver isolated and the speaker ports blocked to create a sealed box. The new crossover specs (eminence 3-way passive crossover 500/5000) say the crossover points should be 500hz and 5k, although from analyisis of this pic shows the crossover points to be around 1k and 3k similar to the old crossover. this has confused me and again i am unsure what is happening at the high end of the frequency response?
MODSPEAKERISOLATED_zps47b7bacf.png
[/URL][/IMG]

Lastly here is a picture of the both the original speaker and new modified speaker, measured at 1 meter with all drivers connected. The green line is the the new modified speaker and from the study of the responses seem that the new speaker is an un-improvement. also again i am confused with what is happening from 2k upwards?
SPEAKERORIGINAL1METERNEWMODGREEN_zps5ee670f0.png
[/URL][/IMG]

all driver were replaced for modification along with crossover
 

wintermute

Administrator
2003-08-03 11:43 am
Sydney
I'd be looking at your Mic Cable. I have had similar issues when there was a problem with the mic cable. The worst example was when mic mic cable wasn't long enough and I tried to join another (soldering and then heatshrinking rather than using connectors). I got a very similar sawtooth on the response with that. (and it is Monday here and I am at work :p)

Tony.
 

sampleton

Member
2013-02-11 5:19 pm
hi wintermute, i dont think its the xlr cable m8 :) as the comb filtering only seems to be happening in the high frequencies. have a feeling that this is due to an increased gap between the driver and the horn throat during modification and mounting, causing some bad comb filtering. also i think the inaccurate crossover points are due maybe an inaccurate crossover network, it did only cost £50 lol
 

wintermute

Administrator
2003-08-03 11:43 am
Sydney
Yes I guess you are mainly seeing it on the HF unit's measurements... Mine happened on higher frequencies but from memory it was present on measurements of all drivers past about 1Khz... but it looked very similar, and redoing the cable fixed it. I thought I may have posted about it at the time, but a check through my threads drew a blank.

Can you do a measurement of say a conventional dome tweeter in the same conditions just to get a baseline high frequency measurement for comparison, just to rule out any problems with the measurement setup?

Tony.