Maybe i am thick!! - (I need to explain better)
I know that behind the RTA there is a some averaging going on either by a DFT (most probably) or a FFT. I can see how to do an RTA using pink noise, but cannot see how to get the correct algorithm to use white noise without getting a rising response.
I am thinking that maybe the problem i have had is that the impedance of the sound card in the Maudio pre is forming a higher order filter with the 1st order filter capacitor. A calibration loop will sort this - just need to check if i have the right leads.
Everything rattles, you just need to shake it hard enough.
I myself use the M-Audio Mobile Pre USB soundcard and a Beringer M8000 mic. Pretty standard stuff that I run from my cheap laptop. I use the ARTA measuring suite and it has a good user manual that explains how you should set up your soundcard as well as a testing regimen that allows you to measure the soundcard itself. If you go by the user instructions step by step you will be a lot wiser about the soundcard and the setup itself.
I have now run FR measurements through all packages (Arta, TrueRTA and HolmImpulse) and i get almost identical response curves from all three. I haven't bottomed out why the bottom end is decaying away so - must be something to do with the external Maudio sound card. I was thinking i would play some pink noise via CD and use TrueRTA to measure - anybody any recommendations for some downloaded pink noise.
I have also been playing with BMS 4590 drivers as well - currently got a A to B comparision with the 2482 set up. Its difficult to tell which sounds best JBL is warmer BMS more detailed. However when you connect up the HF unit in the BMS - the point source is great - the HF is a bit ragged and bright though. ( As a former lowther user i would like point source!) A don't think the bulk standard BMS crossovers do the driver any favors - i need to get a schematic for a better balanced crossover.
Adding an open cell foam plug into the throat helps diffuses the beaming of the HF (Geddes style), more playing with different foams required.
Because the LF in the test signal is being attenuated i think by the external sound card somehow. I know if i play my tone CD i get LF so i throught i would seperate the generator from the measurement so i can fault find the cause of my measurement problem. ( I might need to remove the operator from the equation too!!!)
ok, seems sensible. But it could be the input of the sound card, also.
Everything rattles, you just need to shake it hard enough.
I don't think its the input to the sound card - as i don't hear any LF coming out of the horns when i apply the white or pink noise.
I think i may have it sorted. I have been using a 10min pink noise file i downloaded, which i play through the CD. Using TrueRTA i measure this and average it. The only issue with the file is it is limited to 10 Khz (prob 100Hz at LF to). Attached are a copy of traces for the JBL and the BMS. Before you get too excited - the scales are in 1db divisions. The JBL looks smoother at the bottom end - but does not extend as far. I used a crossover point of 275Hz via the active x-over (4th order i think) The BMS one has the HF unit coupled up (cheap BMS x-over). I know that the horn should only cover 3 octaves max - but i cannot hear any issues upto 5k with the JBL.
just realised its missed off the titles - hopefully you can guess which one is which
Just overlaid the traces for the different 2" CD's i have measured. The only difference i can see from the traces is that BMS has a small peak around 600Hz.
To make the most of the BMS driver in a coaxial mode i need to replace the crossover. Looking on the web it looks like alot of people have used the drivers with JABO horns. Various crossover posted on the web although seem to be inconsistent as have MF notch filters and some have HF notch filters. Anyone know of a good crossover for this driver?
which BMS driver are you using Steve? I have the 4592 mid driver, run it down to 350Hz and up to 3000Hz, it will run to 9000Hz but I found it sounds better at 3KHz with the Raal's from 3KHz. I also used Jabo horns before the auto-tech 200Hz horns I use now.
main system: 15" open baffle bass, 200Hz auto-tech tractrix horn with BMS 4592mid, Raal 140-15D , Amplifon wt40 6s33s push pull, J Sound labs 300B, 6l6 brimar single ended triode amp DEQX PDC, MHZS cd88, Garrard 401. Colarro 2010
Paul - the hints in the graph BMS 4590's ( i know at 6.30am i am not on the ball!!) Currently i am finding the coaxial driver appealing - the recticulated foam in the throat seems to help the HF from beaming (maybe its just attenuating it) pinched the idea from Geddes
I haven't posted but i am finding integrating a HF driver onto the midrange horn difficult - i have tried CD's ribbons AMT's and tweeters nothing seems right - just killing all imaging. The coaxial BMS drivers restore all what has been lost - just need to tame the raggedness (if that possible - million dollar question!).
I am looking forward to listening to all the horn setups at Scalford to get a reality check if what i am striving for is achievable.
Aah-ahhh! You have a woman's speakers, me lord - (Captain Redbeard Rum to Blackadder)
That trace is via a 4th order x-over, so the decay is rapid. Coincidently i carried out an FR measurement today to match the levels between bass and mid-range just using a music CD, there was a lot more lower end tailing down @~200Hz. I cannot explain why - probably to do with the averaging across the octave, but am not sure.
I will post a FR trace for the audax driver, which IMHO still has the best bottom end slam or the 3 drivers i have tested. With a bigger horn i think its virtues would win out, but in this small (68cm) horn its not that clear.
Having got the microphone calibrated ( thanks Serge) and the external sound card calibrated. I have done a few new measurements, only done the BMS drivers so far, will get around to doing the Audax and JBL's soon. As you can seen the responses don't seem to be as smooth as i was hoping for. The BMS 4592 has a Bill Wood designed crossover which is meant to flatten the response, which doesn't seem do.
(2nd Order X over at 250hz modelled in Holmimpulse).
If i use a higher order (4th) could i drop the crossover frequency from the 400 Hz i am currently using in a 220hz horn ? My logic is that the response at F(thoat) will the 12db lower. Flawed logic?
how do you compare 6.5" cone driver with 2" compression drivers? do you have appropriate horns or is it a same horn with the adapter for 6.5" to 2". not sure if this is optimal....
Horn OK Please!
Yes i do have some adaptors which take the throat down for different sizes. Whether these are optimal is another question. I have used the throat angle for the JBL's for my 2" extensions, which are matched to the 2482's but are not optimal for the BMS drivers which have a different throat configuration.
Got around today to measuring my cone drivers - Audax 6.5" and a supravox 9" in the horns. The throat arrangements are different for the drivers as the supravox is a 9" so the throat goes over the surround. The supravox has got far better extension but as a large driver beams a little more. Both drivers ran without a compression chamber on them (open baffle). IMHO i think cones offer more authority (meaty) to the sound than CD's which can sound a bit thin.
( Both drivers have a 2nd order X over @ 375 Hz- if you wonder why they are decaying @ LF
Still no closer to deciding whats best for horns in a normal room