To friendly hornresp guru.. |
Post Reply | Page 123 4> |
Author | |||||
hiphei
Registered User Joined: 09 December 2004 Status: Offline Points: 76 |
Post Options
Thanks(0)
Posted: 19 April 2005 at 3:43pm |
||||
As some of you might already know, i have 4 tempest horns with
KappaPro15LF-II drivers instead of original Adire tempest drivers..
They have turned out to be rather nice horns though i haven't had a chance to give 'em more juice than what single TA-2400 gives. Considering the fact that KappaPro15LF-II voice coil should be able to take 600Wrms i might be able to use amplifiers giving 1200W/bin (eg. from 2 bridged EP2500, two bins on each amp) maybe even more.. Problem is, that i don't have the slightest idea, when will i be exceeding Xmax/Xmech. If i've understood correctly it can be very easy to destroy the drivers and not hear that they are being pushed over their limits..? Is that possible to find out with hornresp? I'm also totally stumped with the idea of how to model them with it.. Is there anyone out here who would be so kind hearted who would do it for me? I don't know how can i make it up, but i'm rather good in electronics so maybe someday i can help one out with problems in that field.. Here's the link to plans: http://www.adireaudio.com/Files/TempestBassHornPlans.pdf Cheers, Jani Edit: Almost forgot! Heres a link to T/S parameters of the driver: http://editweb.iglou.com/eminence/eminence/pages/products02/ speakpro/kap15lf.htm8 Edited by hiphei |
|||||
hiphei
Registered User Joined: 09 December 2004 Status: Offline Points: 76 |
Post Options
Thanks(0)
|
||||
Of course i can type the data to hornresp and calculate it myself..
That is trivial task, but since i have no experience in horn
speakers/horn speaker design, i don't know what values to put in!
Could it be thought that there is some kinda front chamber? Where does the first segment begin? Should it be simulatetd with multiple conical segments or single hyperbolic? Is it exponential? Why there is no input field to insert Fs, Vas, Qm, Qe and Qtot to hornresp? How should i measure path length (outer or inner path or their 'square sum') and where does it begin? What to do when there is no data available to some input-fields (FR,TAL,VTC,ATC)? Can the effect of several cabinet coupling together be simulated by changing XxPi? I think someone who has designed at least some horn-speakers could answer my questions pretty easily. I just can't trust single calculation unless i know what am i doing. Some graphs that i've managed to simulate are VERY far from the measurement found on cabinet plan. I know that i have different drivers, but it's impossible to simulate the cabs with tempest parameters since adire publishes mainly Fs Vas Q, etc values that can't be entered on hornresp, thus not knowing when the result will be correct in the first place!? Anyone out there? Cheers, Jani |
|||||
mobiele eenheid
Old Croc Joined: 15 August 2004 Location: Netherlands Status: Offline Points: 1568 |
Post Options
Thanks(0)
|
||||
Yep. To be honest it would suprise me at all if the Kappa Pro 15LF wouldn't reach Xmax at just 600 Wrms in the Tempesthorn (now that I think about it, that goes for 300 Wrms aswell). A speaker reaching Xmech can be heard if you're pretty close to the box. Hearing how it exceeds Xmax would probably need superhuman hearing (talking about folded basshorns off course).
There is pretty much always a front chamber since the air in the cone can be listed as frontchamber. It's sometimes pretty tricky to find were the horn starts and the frontchamber ends. In this case I would say that the horn starts in the centre in front of the driver (at the baffle not conus) and the frontchamber the space before that. But perhaps someone else thinks/knows different? In first case you could just simulate as an exponential horn, with S1 being the area at the throat and S2 the mouth area. Dividing the horn in conical parts would give a less nice result in SPL-response tho and perhaps also in Xmax, but it would give an good idea anyway. You could always mail Adire/Dan to ask for the specs as he used it. Come to think of it, do that before you start calculating. Previous experiences from myself would be that you can get easily a 10% marge in specs if you calculated roughly yourself.
Than how the heck did I end up with the Tempest as one of the drivers in Hornresp ? Ok tell you why: If you doubleclick on the field were you put in the required T/S-parameters, you'll find that you can calculate RMS. CMS, BL and Mmd from other T/S-parameters, which you do have. Btw. You'll have to calculate in a specific order as you'll find out yourself anyhow. For lenght draw a middle line in the horn. Actual path would probably be slightly longer since bending increases hornlenght. Remember mail Dan first (say you use Hornresp), don't know if someone replies but it will save lots of time if so. VTC is Volume Throat Chamber (it's in cubic cm, so 1000 for 1 ltr), ATC is the Area at the beginning of the Throat Chamber, make it Sd in this case (my quick guess). Leave FR and TAL blank (won't matter too much). Make sure to fill in LRC (make it 20) or you'll simulated something entirely different. 1 cab at 1,0 pi would be the same as 2 cabs in 2,0 pi (just for response not power) 1,0 pi is also 1 cab on the floor and against a wall. Ussually it gets 2 Hz to 5 Hz of lower f3.
Btw. Found all this out in about a month, without help from outside, just simulating, trying and reading Help-function at least 25 times If I can do it why can't you ...... OK OK I pretty much s*ck at electronics. For now... Good luck Edited by mobiele eenheid |
|||||
hiphei
Registered User Joined: 09 December 2004 Status: Offline Points: 76 |
Post Options
Thanks(0)
|
||||
Thanks A LOT!
I think i could've found it out all by myself eventually, but considering my rather busy schedule at work, it would've easily taken quite a bit longer than one month to find out all this! Once i'm home and have time to fiddle with my computer i should already be sleeping..But isn't that the way it goes? I think You have pointed me the correct way to go! (sort of a kickstart.. ) I will also try and contact Dan at Adireaudio with buttload of questions. Supposedly my simulations were wrong, but i didn't reach Xmech with 100Vrms (about 1250W@8Ohms). Poorest/highest result was 14mm diaphragm displacement, which would be still less than Xmech (28mm total from 34mm Xmech).. Thanks again and cheers! Jani |
|||||
mobiele eenheid
Old Croc Joined: 15 August 2004 Location: Netherlands Status: Offline Points: 1568 |
Post Options
Thanks(0)
|
||||
You're talking about Xmech, I'm talking about Xmax. Normally you use a speaker within it's Xmax, sound is best that way and you have plenty headroom as safety. Beyond the Xmax less of the voicecoil is in less of the magnetic gap, so the further the speaker goes beyond Xmax the more it will loose control. Hornresp simulates as if the speaker has full control over the cone in those 14 mm. Instead beyond Xmax (so 5,5 mm orso), the speaker will have less than 100% control, perhaps only 30% at 10 mm, who know's. Well we don't, so we simulate only with the facts we do know: Xmax. Most horns I simulated reach Xmax at just 200-300 Wrm. Theoretically that is. In practice music isn't the sinewave Hornresp is using for the simulation. Depending on the kind of music and lowcut, the actual powerhandling will be about 1 - 2 times as high. The higher the lowcut is placed the higher the powerhandling. Some sinebass songs (d&B) will have close to zero tolerance on the simulated powerhandling. Normal popmusic will give a powerhandling that's about 2 times as high as the theoretical powerhandling. For tekno/techno/gabba I use about 1.4 -1.6 Cheers Edited by mobiele eenheid |
|||||
hiphei
Registered User Joined: 09 December 2004 Status: Offline Points: 76 |
Post Options
Thanks(0)
|
||||
I was after Xmech since i wanted to see how easily i can accidentally
break them if i wanted to try them with more power.. (That'd be
economical catastroph since i've recently bought an apartment..)
I won't be playing very low frequencies through them since they are for my bands live usage. It's mainly kick drum and bass guitar in standard tuning (we play some sort of metal). I made one simulation again during weekend with volumes and conditions carefully measured from the plan (i divided the horn in conical sections). I also handled the horn starting point, front chamber area and volume as you suggested and the result was even better! With all of my 4 cabs together i get LF extension to a bit lower than 50Hz. With 1250W power, the excursion stays within +-1,5mm of Xmax (6mm according to Eminence) down to about 60Hz and then it starts to rise pretty smoothly. @40Hz it was about 11-12mm. If i hipass them somewhere around 40Hz I should be completely safe with amps having such power/cab. Or what do You think? (Taken the fact into account that true excursion won't be even THAT much due to VC going away from magnetic gap and power compression) I think i still should wait a response from Adireaudio and do few more simulations before i buy new amps though.. Response isn't the smoothest available, but since i use them below 100Hz (currently -3dB at 90Hz and 24dB/oct slope), it's totally usable. -Jani- Edited by hiphei |
|||||
mobiele eenheid
Old Croc Joined: 15 August 2004 Location: Netherlands Status: Offline Points: 1568 |
Post Options
Thanks(0)
|
||||
As I tried to explain it's useless to simulated with Xmech. If the Xmax would be 7.5 mm for instance and Xmech would be 15 mm (one-way). If the speaker would still behave linear beyond Xmax it would take about 4 times more power to reach Xmech, but since the speaker looses control above Xmax it could well be that Xmech is reached with only the double power (or 6 times for that matter). You just don't have enough facts to base any serious thoughts on that simulation. I would say Xmech is a marketing-related invention. Excellent material for a manufacturer to put in big letters in an advertisement, but only something you can keep in the back of your mind while designing, not actually work with it.
Perhaps I'm misreading this, but in my vision it looks like it would be exactly the other way around. If the VC moves out of the gap it will need less power for excursion. Powercompression wouldn't save you from overexcursion (it did totally the opposite in my experiences).
I hope that by now you will take the frequency where the Xmax is reached as your new lowcut frequency. Play it save at first, if you're unhappy with the results you can always look up to the limits. Instead of taking risks by trusting on a best case scenario straight away.
If you would try it again at a later time, chances are that you would even make more progress. Don't go over one night simulating. Let it rest for a few days and look again, changing every hornparameter and compare with the previous result with the comparisation-function.
Alright! Cheers Edited by mobiele eenheid |
|||||
Timber_MG
Registered User Joined: 27 March 2004 Location: South Africa Status: Offline Points: 1454 |
Post Options
Thanks(0)
|
||||
Xmech can indeed very usefull, eg. the Ciare 12 you use has an Xmech
very close to its Xmax. Also if the magnetic field is not symetrical,
you might arrive at Xmech in one direction sooner than in another.
However the BL product falls of fairly quickly after Xmax and hence
most Pro-sound speakers don't really come very close to Xmech, except
when put in a mode where they are in a resonant system where an
acoustical force "helps" the driver (like ar fc in a horn with a
large/no rear chamer) along a little.
|
|||||
mobiele eenheid
Old Croc Joined: 15 August 2004 Location: Netherlands Status: Offline Points: 1568 |
Post Options
Thanks(0)
|
||||
Cheers |
|||||
hiphei
Registered User Joined: 09 December 2004 Status: Offline Points: 76 |
Post Options
Thanks(0)
|
||||
I'm a bit confused now.. But then again
my conclusions were based on magnetism lessons in school abt. 10 years
ago and common sense.. Not on any knowledge on speaker theory!
See i've always thought that when the voice coil goes farther from magnetic gap the magnetic flux would be smaller thus requiring more current in coil to keep the force equal. Thus the cone excursion would be smaller and smaller flattening the obtained pressure maximum until the Xmech is reached where the driver bottoms mechanically. I've always thought that this is the main reason for 2nd order harmonic distortion. I've also concluded that 'air spring' in back chamber, spider and cone suspension would add up to that even more.. It's really hard to fit the opposite in my head. I won't buy new amps based on these simulations at least until i get some data from Adire audio.. It's still a bit fuzzy what should be interpreted as the front chamber volume and where the horn really starts. Especially now when my budget is even more tight.. -Jani- |
|||||
hiphei
Registered User Joined: 09 December 2004 Status: Offline Points: 76 |
Post Options
Thanks(0)
|
||||
I discovered an error in my last simulation! (I did quick sim at work
today and wondered why an earth i had SO different results.. I hope my
boss didn't mind )
I already had wondered why did it look so good especially considering the 'hoffman iron law' and Eminence drivers. Strangely the spl response stays the same but excursions were totally wrong. (Well maybe not so strangely.. I had comma in wrong place on mechanical suspension loss value ) |
|||||
mobiele eenheid
Old Croc Joined: 15 August 2004 Location: Netherlands Status: Offline Points: 1568 |
Post Options
Thanks(0)
|
||||
Too bad I'm not a guru, so my motivation in this case is rather based on thoughts as well. Perhaps someone else could help us out?
Agree.
Not sure that I understand this correctly (at a certain point my lack of English blurs my mind and keeps it from understanding even when read 4 times over ). But did you conclude that the VC is already moving with certain force/speed (gets kinda launched) and at the point where it normally would be forced to turn back the motor has not enough grip on the VC anymore to effectivly do so Still biggest problem would be that if the flux lowers (and so does the BL and No) the speaker parameters change, while Hornresp asumes that it will stay the same (as would be the case in Xmax). We even don't know how fast or slow BL decreases. So there really would be no use in using the Xmech into any of your simulations/calculations.
Personally I use the smallest area as the border between chamber and horn.
I had the same problem the first 2 days I used Hornresp Cheers Edited by mobiele eenheid |
|||||
Post Reply | Page 123 4> |
Tweet |
Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |