Can't the same thing be accomplished much easier via DSP?
If you've got enough parametric EQ channels available, of course.
Here's my POC #3 before and after some DSP-applied EQ to address the out-of-band response peaks:
Attachments
PEQ can flatten the peaks but I think the sound quality is better if done passively and via acoustic methods - it is an inherently better speaker and you can see that the impedance peaks are flatter and almost gone in the case of the first peak at 176Hz. You need AkAbak to sim this though - just a simple added line in the script for each stub (where node 11 is at S2):
Duct 'Stub 176Hz' node=11=0 dD=3in Len=21.6in Visc=30
I haven't heard THs with or without resonators, so I can't confirm that "sounds better" position. Certainly however there have been several cases here of successful THs that were built without resonators. Are resonators featured in all of Danley's THs?
I suspect it's going to be difficult to get them right without quite a bit of trial an error. For example, if you have a peak of 6dB and your resonator only drops it by 3dB, you'll still need to resort to EQ, or redo the resonator to be more efficient. What if you miss the target frequency?
Addressing the peaks via DSP seems much easier. And my TH sounds fine. And it's not like you'll be using it up that high anyway.
The models seem to predict the freq peak locations well, you can always measure with a mic or impedance tester to confirm location. After that it is simple to calc length based on speed of sound. Sure there is trial and error and I am not saying DSP EQ won't be needed - matter of minimizing how much PEQ one needs to apply. In JG's case, that peak was very high and sharp that getting her sub to go above 100Hz meant having that peak to deal with. The added benefit was the flattening of the sag at 50Hz. That would have required positive EQ boost which increases excursion.
Those predictions are based off an Akabak model, not the actual measured response of the TH. For all you know the passband is a lot smoother due to losses not taken into consideration in the Akabak model.
I say measure than correct with DSP if necessary. If JG's not happy with the result, then JG can try drilling holes in the box and attaching PVC ducts (which IMO are likely to break off at the earliest opportunity if they're run externally).
I need to measure asap! I got pulled away on some other ongoing projects like fixing up my previous car to sell it, but I'm almost done. Ideally I would like to test tomorrow, or the day after at the latest.
Honestly I think with the capabilities of minidsp I will just do that, I'm quite happy with the build and will most likely avoid any protrusional additions.
Brian- the minidsp 2x4 advanced has 6 PEQ slots per channel, do you think that would do? I guess I will be easy to see how many I'll need after a good test.
Also re: resonators- it appears that there is about a 4db loss at 133hz and slightly raised LC in the sim, those are two things I'm not sure if I'm willing to part with, even if the 50hz dip is flattened.. The problem is that it is flattened only from other surrounding levels brought down in efficiency. If it is EQd a couple dB higher at 50, I don't think that will max out excursion either. I have to go look at those graphs again to be sure.
Either way, gonna test. I know for sure there is a dip at 50, it will be interesting to see how much. I think I'm going to go mess around in REW tonight and prepare, I've never used it before- the quick start does seem thorough luckily
Thanks again for all the awesome help. I'll get you that measured graph asap!
Honestly I think with the capabilities of minidsp I will just do that, I'm quite happy with the build and will most likely avoid any protrusional additions.
Brian- the minidsp 2x4 advanced has 6 PEQ slots per channel, do you think that would do? I guess I will be easy to see how many I'll need after a good test.
Also re: resonators- it appears that there is about a 4db loss at 133hz and slightly raised LC in the sim, those are two things I'm not sure if I'm willing to part with, even if the 50hz dip is flattened.. The problem is that it is flattened only from other surrounding levels brought down in efficiency. If it is EQd a couple dB higher at 50, I don't think that will max out excursion either. I have to go look at those graphs again to be sure.
Either way, gonna test. I know for sure there is a dip at 50, it will be interesting to see how much. I think I'm going to go mess around in REW tonight and prepare, I've never used it before- the quick start does seem thorough luckily
Thanks again for all the awesome help. I'll get you that measured graph asap!
Okay well I realized it is rarely quiet here during the day, at night is probably the best time to do it.
I still haven't calibrated my soundcard, and was just using the lineout on my laptop for audio out. But, I was able to get the UMM working and used the calibration file from the Dayton website
Here are the results, this was 4 sweeps from 20-200hz
Here are the filters currently applied on the minidsp
Should I be testing w/ no filters applied? Something tells me that is dangerous... The mic was a rough 1m away and my laptop output level was at 25%. UMM input volume in system prefs 75%. Speaker height is approx 30" off the ground (same as truck bed) and mic was also 30" off the ground (I suppose I should raise that to listening position)
I still haven't calibrated my soundcard, and was just using the lineout on my laptop for audio out. But, I was able to get the UMM working and used the calibration file from the Dayton website
Here are the results, this was 4 sweeps from 20-200hz
Here are the filters currently applied on the minidsp
Should I be testing w/ no filters applied? Something tells me that is dangerous... The mic was a rough 1m away and my laptop output level was at 25%. UMM input volume in system prefs 75%. Speaker height is approx 30" off the ground (same as truck bed) and mic was also 30" off the ground (I suppose I should raise that to listening position)
Last edited:
Obviously some very unexpected stuff going on at 80+hz.
Some notes:
- drivers are mounted with crappy screws that are barely enough to trap the driver holes (just cheap home depot gold screws, need to get wider head screws)
- no weatherseal applied yet
- first time I've used REW, not sure if I might be doing something wrong by not calibrating my soundcard output
- lower end is measuring flatter than I expected
Curious to know your thoughts!
Some notes:
- drivers are mounted with crappy screws that are barely enough to trap the driver holes (just cheap home depot gold screws, need to get wider head screws)
- no weatherseal applied yet
- first time I've used REW, not sure if I might be doing something wrong by not calibrating my soundcard output
- lower end is measuring flatter than I expected
Curious to know your thoughts!
Calibrating the output will give you an absolute value of SPL. Not really necessary for frequency response measurements, but you'll need it if you want to do distortion.
Try taking off the low-pass filter and see how high you can run it. Looks like you'll need a notch ~120Hz, but if it all sounds fine up there, that'll make it easier to get a midbass horn to match.
Chris
Try taking off the low-pass filter and see how high you can run it. Looks like you'll need a notch ~120Hz, but if it all sounds fine up there, that'll make it easier to get a midbass horn to match.
Chris
That's good to know, Chris. Thanks! That is kind of what I figured.
The major unexpected part is at 90hz. I think there is likely a leak around one of the drivers, as I am using too small of screws and no weather stripping to seal them... I even seem to remember hearing a "pop" sound with adding a few of the screws, as if the head popped through the mounting hole on the driver (the heads of the screws I used are almost hilariously undersized) Going to fix that tomorrow and try again with PEQ bypassed!
As well I am wondering if it might have something to do with mic position, as I notice large fluctuations in sound levels within 4 or 5ft radius of the mouth- and tonight it was darn near 3ft. Being that it is late, I could not turn it up enough to stand back (I've already put my roomies through enough torture!). So, I'll try for 8-10ft away tomorrow at a more substantial volume level along with properly sealing/mounting the drivers.
Regardless, I am still very happy with the performance- as is everyone who hears this box. However I hope with these driver mount improvements & measurement changes to be at least a little bit closer to the simulation! I also wanna try your suggestion with removing LPF.. but I should be careful with how loud it gets, right?
The major unexpected part is at 90hz. I think there is likely a leak around one of the drivers, as I am using too small of screws and no weather stripping to seal them... I even seem to remember hearing a "pop" sound with adding a few of the screws, as if the head popped through the mounting hole on the driver (the heads of the screws I used are almost hilariously undersized) Going to fix that tomorrow and try again with PEQ bypassed!
As well I am wondering if it might have something to do with mic position, as I notice large fluctuations in sound levels within 4 or 5ft radius of the mouth- and tonight it was darn near 3ft. Being that it is late, I could not turn it up enough to stand back (I've already put my roomies through enough torture!). So, I'll try for 8-10ft away tomorrow at a more substantial volume level along with properly sealing/mounting the drivers.
Regardless, I am still very happy with the performance- as is everyone who hears this box. However I hope with these driver mount improvements & measurement changes to be at least a little bit closer to the simulation! I also wanna try your suggestion with removing LPF.. but I should be careful with how loud it gets, right?
Last edited:
Taking off the low-pass filter shouldn't give you any trouble - it cuts down the stuff above 100Hz.
The one to worry about is the high-pass (aka low-cut) filter, which cuts the excursion below tuning. If you're throwing any power around, make sure that's engaged.
If you're experiencing leaks, I'd expect the most output to be lost around the bottom tuning frequency. I do have limited experience with THs, though, so I'd be interested to see your before/after measurements.
Chris
The one to worry about is the high-pass (aka low-cut) filter, which cuts the excursion below tuning. If you're throwing any power around, make sure that's engaged.
If you're experiencing leaks, I'd expect the most output to be lost around the bottom tuning frequency. I do have limited experience with THs, though, so I'd be interested to see your before/after measurements.
Chris
JG,
To get a baseline without floor bounce and other room effects measure with mic at exit plane of sub with very low excitation like -40dB FS on REW. Use no filters anywhere and that will let us see how it looks compared to sims. If you are putting in well under 1v don't worry about damaging it. The mic can't handle much without hitting the rails at the exit plane. I think 80Hz may be floor bounce cancellation. Put sub flat on floor to do this and we will get a better ground plane 2pi measurement. Your mic is calibrated in SPL. Your sound card only needs to be calibrated with your amp using a DMM and sine wave gen to dummy resistor load so you can correlates output dB FS with volts and hence watts.
Your sub shouldn't need a PEQ like that to work. That is making a lot of positive boost which is not ideal. The phase looks very flat though and that is good.
To get a baseline without floor bounce and other room effects measure with mic at exit plane of sub with very low excitation like -40dB FS on REW. Use no filters anywhere and that will let us see how it looks compared to sims. If you are putting in well under 1v don't worry about damaging it. The mic can't handle much without hitting the rails at the exit plane. I think 80Hz may be floor bounce cancellation. Put sub flat on floor to do this and we will get a better ground plane 2pi measurement. Your mic is calibrated in SPL. Your sound card only needs to be calibrated with your amp using a DMM and sine wave gen to dummy resistor load so you can correlates output dB FS with volts and hence watts.
Your sub shouldn't need a PEQ like that to work. That is making a lot of positive boost which is not ideal. The phase looks very flat though and that is good.
Last edited:
Also, would actually measuring my voltage at the speaker terminals be a good way to calibrate? I.e. 2.83V would equal 1W, and at 1m of distance I would be potentially ready to measure sensitivity?
That way I would also calibrate past the issue of amplifier gain setting? Probably would do this with speakers disconnected and all minidsp filters bypassed.
That way I would also calibrate past the issue of amplifier gain setting? Probably would do this with speakers disconnected and all minidsp filters bypassed.
Last edited:
The only issue I've had personally with using MiniDSP for EQ in which positive gain is applied, is the MiniDSP doesn't have any gain (0.7 in / 0.7 out). It's not usually a big issue if you can increase amplifier gain to compensate for it, (I'm not sure what the input sensitivity is like on the JL amps) but if you've got an amplifier that needs 1v of input to reach maximum output, everywhere that boost is applied will clip the MiniDSP.
I try to think of the MiniDSP boost / cut scale as if the top (+20) is zero, then every EQ setting is a reduction, requiring you to goose the amplifier gain, as the MiniDSP is effectively a loss stage by however many db your highest gain filter is set to.
This might just be retardation on my part, but MiniDSP clipping sounds pretty terrible and I usually like to run my subs pretty hot for some types of music.
I try to think of the MiniDSP boost / cut scale as if the top (+20) is zero, then every EQ setting is a reduction, requiring you to goose the amplifier gain, as the MiniDSP is effectively a loss stage by however many db your highest gain filter is set to.
This might just be retardation on my part, but MiniDSP clipping sounds pretty terrible and I usually like to run my subs pretty hot for some types of music.
That makes sense, so to convert the other way it is sqrt of power * impedance or v = sqrt(1W*2ohm) = 1.41v, and max voltage of my amp at 2ohm is v = sqrt(1200W*2ohm) = 48.99v. Neato!
DrDyna- that is a good point and really good info on the minidsp I have never looked into amidst all the years of using it. Luckily the JL amps seem to have insane amounts of gain, maybe due to the regulated power supply... Gain setting is turned up less than halfway using the 2V "low" voltage input range (between 200mV & 2V sensitivity) to reach 49V at max volume from my cell phone (according to an online source it is a max of 0.43 output voltage which does seem craptastically low).
I'm hoping the dip is just the floor bounce cancellation you speak of, xrk. Do you get that idea due to the specific length of soundwave at 80/90hz? I'd really love to run zero EQ at all. And I still think better securing and sealing the drivers will help too (maybe mainly for the LF corner which also seems a bit insensitive ~ what a jerk, ha). If y'all saw the dingbat job I did with those screws holding the speakers on you'd probably be shaking your collective heads at me. Kinda looks like the driver may wiggle itself off of them. It was really only supposed to be temporary until I painted it anyway... I have been waiting to have my top done before that step. Not going to wait on mounting properly anymore though, I need those numbers for my sanity.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.