I've built PA tops with relatively high tunings (60-70Hz) and the port length usually ends up being about the thickness of a 3/4" baffle.
I've noticed many generic vented box "suggested designs" from woofer manufacturers just have a hole in the baffle.
In my experience WinISD (and maybe other apps) will usually overstate the actual port length needed, so build the cabinet and vent so that you can adjust the length while taking impedance sweeps.
The optimal design will be a dance between Fb, max excursion, port velocity (keep under 20m/s) and port resonance (hopefully an octave or more above your XO)
Ive tried and been put off previously as its awkward on a mac but i will try again. I appreciate its a little crazy to be building speakers without modelling.
Finally after much faffing around I have winISD on my mac! So here is my first ever, extremely basic attempt at modelling! Hurray
Thanks tommus your words have confirmed the model, 2 vents at 55mm = 15mm depth (my front baffle will be 15mm) - i was going to confirm whether baffle depth counts as port. Port mach 0.11 is to do with air through the port not chuffing right, seem to be in the clear there.
If tuning lower you lose out on all frequencies above and if tuning higher you end up with a crazy resonance on the Fb so 60hz seems right. I dont know how to calculate/find/see port resonance though?
Other than that basically this confirms I can copy the yamaha's more or less which makes things nice and easy.
I know modelling doesnt come much simpler but that was fun and im glad ive got it working finally
Thanks tommus your words have confirmed the model, 2 vents at 55mm = 15mm depth (my front baffle will be 15mm) - i was going to confirm whether baffle depth counts as port. Port mach 0.11 is to do with air through the port not chuffing right, seem to be in the clear there.
If tuning lower you lose out on all frequencies above and if tuning higher you end up with a crazy resonance on the Fb so 60hz seems right. I dont know how to calculate/find/see port resonance though?
Other than that basically this confirms I can copy the yamaha's more or less which makes things nice and easy.
I know modelling doesnt come much simpler but that was fun and im glad ive got it working finally
Hi,
Glad you've made the effort to dive into modelling, it really does help understand how different characteristics of a design influence each other.
A couple of thoughts which I hope will help.
First, you're using one of the older versions of WinISD, and I don't think that will show you the vent resonance - the current version (0.7.0.950) does, so it may well be worth getting that.
Second, I'm not sure that two 55mm vents will really be enough. Ideally we want to keep vent airspeed below Mach 0.05, at least for hifi applications. There is a school of thought that for PA use we can get away with higher figures. However, even then, the highest figure I've seen suggested is M 0.1, and that only rarely. It would also be worth checking what power level was being simulated when your 0.11 figure came up - IIRC the default for WinISD is to model at 1 watt, which is not how PA speakers spend most of their working life 😉
What you are best to do is model at a power level that gets cone excursion up to Xmax and then recheck the vent airspeed. (I've just noticed on your screengrab that you don't even have a tab to show the excursion graph, so that must be a very old version - please do get the current one before you go any further!)
Lastly, 70l for a 12" 2-way PA box seems pretty large - there are plenty 15" boxes at that size. If you're sure you want to handle that size of a box then that's great, but it's worth checking the overall size before committing to the project. FWIW, the Yamahas you were comparing to are about 61 litres external, which, by the time we allow for the thickness of the walls, the baffle being set back from the front of the cab for grille clearance, the volume taken up by the drivers, handles, bracing, crossover etc would probably be around 35-40 litres net, which is the figure we need to use for modelling. That will result in you needing a longer vent to maintain your target 60Hz tuning though.
Hope this isn't raining too hard on your parade,
Cheers,
David.
Last edited:
Ah no not raining on my parade at all David, very grateful for the pointers! Will update the version I have (no idea how I ended up with an ancient copy) and update my methods and thinking, it’s all learning and I’m enjoying it!
Also note that many things in the program can be changed on the fly just by clicking on them. Examples of this I have circled below, change from circular ports to square/rectangular by clicking on the shape, change the units displayed in a window by clicking on it. Also note that I have previously set the input power in the Signal tab to a value that produces xmax, then the port sizes can be modeled to keep vent airspeed to an acceptable level.
Last edited:
Super helpful everyone. I have updated to the latest version, indeed there are lots more buttons and levers to play with. Just need a few noob queries tidying up as this is a whole new horizon i have discovered!
When adding input power to find the xmax limit, is this going to absolute or will it be affected by other parameters.? For instance, I give it 325w and it just kisses Xmax at 89hz. But as driver is rated by manufacturers as 400w (a.e.s) / 1600w peak - even if thats a little optimistic of them, 1600w gets it more than double beyond Xmax (in this kind of box?). I am aware id probably not be powering each channel with that much headroom but just theoretically wonder how these relate, whether to ignore power ratings as the model will be more accurate etc.
Much more to look at and try and understand in this version, very much juicier and fascinating to see how it all relates.
Regarding the ports, 2 ports 50mm x 45mm long seems to produce a flat 0 all the way up the range at xmax signal. Does this sound plausible? With a 1st port resonance (?) of 3819.25 hz - which is not the most desirably place for any sort of resonance right 4khz is where harshness lives?
When adding input power to find the xmax limit, is this going to absolute or will it be affected by other parameters.? For instance, I give it 325w and it just kisses Xmax at 89hz. But as driver is rated by manufacturers as 400w (a.e.s) / 1600w peak - even if thats a little optimistic of them, 1600w gets it more than double beyond Xmax (in this kind of box?). I am aware id probably not be powering each channel with that much headroom but just theoretically wonder how these relate, whether to ignore power ratings as the model will be more accurate etc.
Much more to look at and try and understand in this version, very much juicier and fascinating to see how it all relates.
Regarding the ports, 2 ports 50mm x 45mm long seems to produce a flat 0 all the way up the range at xmax signal. Does this sound plausible? With a 1st port resonance (?) of 3819.25 hz - which is not the most desirably place for any sort of resonance right 4khz is where harshness lives?
Ah ha I added some filters to emulate the signal processing theyll go through and its helped with the Xmax issue a bit, i can see theres more to play with on that front.
The absolute power rating of the woofer is sort of irrelevant, it's up to you not to drive the cabinet at power that will bottom out the woofer and make you ears bleed from tweeter distortion.
If you design for the edge of the envelope and then use EQ to boost the bass even a little you can go past the Xmax pretty easily. +3db will double the power input at that frequency. A 2nd or 3rd order HP filter below the tuning frequency can give you a lot more headroom without affecting the response much at all. But you can easily use this headroom up with EQ boost. If you are using a digital HP crossover at say 80Hz or 100Hz then the equation changes dramatically. If I'm using DSP processing, I use the "force flat response" and apply the ideal filters that I want to see in the final response for more perspective. I find this is a grey area of the simulation process and just make sure I have plenty of headroom at reasonable drive levels. Typical usage probably won't see more than a few hundred watts peak into the woofer.
Port resonance at 3800Hz is good if the woofer is crossed over at say 1500Hz, especially with a 4th order slope. Any hash will be -24db down. Realistically you probably won't get the resonance higher without messing up other parameters. Judicious damping in the cabinet will help some but that's another can of worms.
If you design for the edge of the envelope and then use EQ to boost the bass even a little you can go past the Xmax pretty easily. +3db will double the power input at that frequency. A 2nd or 3rd order HP filter below the tuning frequency can give you a lot more headroom without affecting the response much at all. But you can easily use this headroom up with EQ boost. If you are using a digital HP crossover at say 80Hz or 100Hz then the equation changes dramatically. If I'm using DSP processing, I use the "force flat response" and apply the ideal filters that I want to see in the final response for more perspective. I find this is a grey area of the simulation process and just make sure I have plenty of headroom at reasonable drive levels. Typical usage probably won't see more than a few hundred watts peak into the woofer.
Port resonance at 3800Hz is good if the woofer is crossed over at say 1500Hz, especially with a 4th order slope. Any hash will be -24db down. Realistically you probably won't get the resonance higher without messing up other parameters. Judicious damping in the cabinet will help some but that's another can of worms.
Ok well after a little bit more playing around im beginning to see the trade-offs i have to figure out... i appreciate this is probably the shallow end of the pool for you guys but ill keep spouting my workings as its been a fun and fast learning curve with your combined feedback, much appreciated
In reality these cabs will usually be used over kicks and sub with crossovers, and the modelling reveals that after most commonly applied HP filters they all level out to more or less exactly the same, so thats reassuring. However, with no filtering at all (with which they might be used occasionally direct from an amp, as monitors in a near-field setting) i suppose its worth doing the excersize just to understand the principles and perhaps keep them a little safer from over excursion in future
So its probably nitpicking but ive crunched some numbers anyway
I have modelled 45/50/55/60hz enclosure tunings and the lower tunings exceed Xmax 70-90hz at 300w, so that rules them out i guess. Higher tuning frequencies give higher port resonances (safely away from woofer range) and less change of exceeding Xmax, but also a big bump around 100hz (+6.5db for a 60hz enclosure) which i suppose I would EQ out?
LMK if ive missed anything stoopid or ill start doing my drawings of the boxes... thanks all 🙂
all with same filter applied @ 60hz
@350w
In reality these cabs will usually be used over kicks and sub with crossovers, and the modelling reveals that after most commonly applied HP filters they all level out to more or less exactly the same, so thats reassuring. However, with no filtering at all (with which they might be used occasionally direct from an amp, as monitors in a near-field setting) i suppose its worth doing the excersize just to understand the principles and perhaps keep them a little safer from over excursion in future
So its probably nitpicking but ive crunched some numbers anyway
I have modelled 45/50/55/60hz enclosure tunings and the lower tunings exceed Xmax 70-90hz at 300w, so that rules them out i guess. Higher tuning frequencies give higher port resonances (safely away from woofer range) and less change of exceeding Xmax, but also a big bump around 100hz (+6.5db for a 60hz enclosure) which i suppose I would EQ out?
LMK if ive missed anything stoopid or ill start doing my drawings of the boxes... thanks all 🙂
all with same filter applied @ 60hz
@350w
if you are going to run them above kick bins (you did say that at the beginning I see now) make plugs to seal the cabinet.
no midrange resonances. remove when you need to run them full range.
no midrange resonances. remove when you need to run them full range.
Hi again,
Those peaks of 4-6dB don't seem right to me, are we still talking about the same PD.123 driver in something around 35-40 litres?
I wonder if, in reloading the newer version of WinISD, some units have got mixed up, throwing off the model? Also, the SPL doesn't seem right for 350W - that looks more like a couple of Watts only.
Here's my effort in 35l tuned to 60Hz @ 350W with no filters - you can see it doesn't even hit the 7mm of the manufacturer's Xmax in-band:
SPL:
Excursion:
Airspeed with two 10cm diameter vents - this would be no problem after an HPF is put in place - 24dB Butterworth @ 50Hz would be ideal here IMO.:
I'm also attaching my WinISD project file in the zip in case you want to compare with yours to see how the differences have arisen.
Last edited:
Oh thanks very kind of you to check that. I hope your model is right, I’ll double check my parameters didn’t get messed up, the only other major difference I can see is that your ports are much bigger and longer than mine, perhaps that’s a factor. I will check in the new year when my head regains balance
No they won't actually. This is a bass box simulation program and as such it does not simulate the fullrange response of the system... just the lowend section up to 150-200hz or so. To get an idea of the fullrange response of the driver look at the manufacurers published response and again ignore everything below ~150hz or so as the box it is installed in determines that.
Published driver power handling can be very misleading, as you are seeing the enclosure type introduces resonant frequencies where excursion peaks and that ultimately becomes a liminting factor, so the amount of power a driver can handle will depend on how it is loaded and the frequency range it is exposed to. The very same driver could be limited to a few hunderd watts in one situation but handle many times more in another.
Also in the real world it may not be as bad as it would appear because music is dynamic and you're not likely to experience sustained tones at any one frequency.. unless you're into Dubstep of course.😢
Ive loaded our projects next to one another David - the SPL difference was because I had left mine set at 8m distance (i was just playing around with parameters and forgot to change it back - oopsie) and I had also gotten confused and put 8 ohms into the series resistance box on signal settings, which gave the humps. Lines now closely tracking your model, thanks.
Conanski - cant promise the occasional dubstep number wont make its way through them (there are actually some quite sophisticated bassy numbers out there, not all that saw-synth nightmare noise) but itll mainly be boogie music from across the genres/continents 🙂 Thanks for the tips though re: limitations of the software and patching together the sources, didnt realise that 150-200hz was the point at which the enclosure kicks in more so, interesting!
Think im ready for my build now thanks everyone 🙂
Conanski - cant promise the occasional dubstep number wont make its way through them (there are actually some quite sophisticated bassy numbers out there, not all that saw-synth nightmare noise) but itll mainly be boogie music from across the genres/continents 🙂 Thanks for the tips though re: limitations of the software and patching together the sources, didnt realise that 150-200hz was the point at which the enclosure kicks in more so, interesting!
Think im ready for my build now thanks everyone 🙂