Dear Experts,
I am new to speaker building. I have prepared a design based on Peerless India drivers. Woofer is 5.25" and tweeter is 26mm Silk Dome. I am attaching herewith a short ppt on the project. Please go through it and suggest improvements. If you need more information just reply to the post.
Regards
Prashant
I am new to speaker building. I have prepared a design based on Peerless India drivers. Woofer is 5.25" and tweeter is 26mm Silk Dome. I am attaching herewith a short ppt on the project. Please go through it and suggest improvements. If you need more information just reply to the post.
Regards
Prashant
Hi Prasi, I suspect in answer to your question will it sound boomy, the answer may well be yes. It looks like you have modeled with a series resistance (in line with the woofer) of 0.8 ohms. Is that based on the DCR of the coil you are planning on using? 0.8 ohms is quite a lot and will drop the efficiency (but make it model somewhat better, and with bigger volume).
See attached the step response for your modeled 20L with 52.7 hz tuning and 0.8 ohms series resistance. Then compare to step response for 7L with 66Hz tuning and 0.3 ohms series resistance. The wiggles in the step response shows how long the driver takes to return to rest, the more wiggles the more boomy I think it will sound. This is the tradeoff for getting lower extention.
I've also attached a plot comparing the 20L 52.7Hz FR curve to 7L 66Hz FR curve. You can see that the 7L the bass extension is much less, but I suspect that subjectively it will sound better. You can always build two boxes and try each and see what the difference is, no better way to learn!!
Note that I dropped the port diameter to 5cm for the 7L box model as it had terrible port resonance with 7cm port! The 7L size was the only other size I modeled as it was close to the recommended size.
ps. i thought that the quality of your hand drawing was excelent! 😉
edit: and by the way, I'm no expert!! 😉
Tony.
See attached the step response for your modeled 20L with 52.7 hz tuning and 0.8 ohms series resistance. Then compare to step response for 7L with 66Hz tuning and 0.3 ohms series resistance. The wiggles in the step response shows how long the driver takes to return to rest, the more wiggles the more boomy I think it will sound. This is the tradeoff for getting lower extention.
I've also attached a plot comparing the 20L 52.7Hz FR curve to 7L 66Hz FR curve. You can see that the 7L the bass extension is much less, but I suspect that subjectively it will sound better. You can always build two boxes and try each and see what the difference is, no better way to learn!!
Note that I dropped the port diameter to 5cm for the 7L box model as it had terrible port resonance with 7cm port! The 7L size was the only other size I modeled as it was close to the recommended size.
ps. i thought that the quality of your hand drawing was excelent! 😉
edit: and by the way, I'm no expert!! 😉
Tony.
Attachments
Last edited:
Improvements ? I dunno! Try and see.
I would put the port on the back, and maybe a total rework in the dimensions.
The woofer could be on the top side ( obviously the cabinet is 90° !) and well, oh, the tweeter ! I prefer to keep every component separated.
In your design, the tweeter could stay below the woofer.
I would put the port on the back, and maybe a total rework in the dimensions.
The woofer could be on the top side ( obviously the cabinet is 90° !) and well, oh, the tweeter ! I prefer to keep every component separated.
In your design, the tweeter could stay below the woofer.
Tony here, too, and same condition 😡
You could use a lower sensitivity tweeter. Usually for hifi the final tuning brings some attenuation in the treble, so a direct approach ( less sensitive tw ) is better.
Another option might be to go to the slightly bigger S16NI It is almost the same price as the M13H at this vendor Peerless - 6" Coated Paper Cone Woofer - S16NI and will outperform (in the low frequency stakes) the M13 hands down.
I modeled it in 20L (which is much smaller than the unibox recommendation of 66L) but that actually gives a very nice rolloff and a slightly rising response (rising as frequency decreases) It would be worth considering. I suspect that the 5" driver is targeted as being a midrange driver.
Transient response on the 16SI is good in the small enclosure too, as is it's power handling. I increased the port to 6CM for it, but you could probably use 7cm... again I just went with your volume I did not try to optimize, but I think you will see the advantage in the comparison graph from a low end point of view.
I've not heard any of these drivers, and with no Frequency response graph it is a bit hard to tell how suited to your application they are.
Anyway just another thing to consider 🙂
edit: I had the 0.8 ohms series resistance in this model not 0.3ohms that will have affected the result somewhat. with 20L and 0.3 ohms 43Hz tuning seems about the best.
Tony.
I modeled it in 20L (which is much smaller than the unibox recommendation of 66L) but that actually gives a very nice rolloff and a slightly rising response (rising as frequency decreases) It would be worth considering. I suspect that the 5" driver is targeted as being a midrange driver.
Transient response on the 16SI is good in the small enclosure too, as is it's power handling. I increased the port to 6CM for it, but you could probably use 7cm... again I just went with your volume I did not try to optimize, but I think you will see the advantage in the comparison graph from a low end point of view.
I've not heard any of these drivers, and with no Frequency response graph it is a bit hard to tell how suited to your application they are.
Anyway just another thing to consider 🙂
edit: I had the 0.8 ohms series resistance in this model not 0.3ohms that will have affected the result somewhat. with 20L and 0.3 ohms 43Hz tuning seems about the best.
Tony.
Attachments
Last edited:
Dear wintermute,
Thanks a lot for your inputs. I had never understood the significance of "wiggles" in unibox. Now that you told me, I think I have screwed my design. It would have been "boom reflex box" and not a "bass reflex box"...🙂 Phew! . Thanks a lot.
Now regarding the 0.8 DCR resistance in series, somewhere I had read that one needs to put 0.8 ohms under Rs (external components).
Thanks picowallspeaker,
I will put the tweeter below the woofer. What is the reason for recommendation for putting the port at the back?. also what changes in the dimensions do you suggest. to avoid standing waves? Or simply should i go by the "golden ratio"?
Limited tweeter options available in India (Peerless Audio) and I heard that they manufacture good drivers and GR- Research (Speaker kits, loudspeaker design services, sub-woofers and amplifiers.) uses peerless India tweeters and woofers.
Also all tweeters have nearly the same sensitivity.
thank you wintermute, I see the advantage in using S20NI. Response curve is much better. As I said earlier, I am novice and I always thought that one can design a box only around unibox suggestions(60 liters is way too big for my living room).
With the 5.25 " driver, I thought to improve the low frequency extension, which is not correct as you have shown me in your post.
I will work on the simulations with S16NI and also another very good woofer (as per indian DIYers) the M16GJ (Peerless - 6.5" Glass Fiber Cone Woofer - M16GJ).
I will get back to you for your suggestions once again after preparing the entire design (Simulation, Drawing, Crossover). You have been very helpful to me and probably saved me INR 15000 (about AUD270).
I have designed the cross over based on the following link. http://www.diyaudio.com/forums/mult...designing-crossovers-without-measurement.html
Will the crossover network be of acceptable performance when built by above tutorial? Again I say that its my first project ...some imperfections are bound to be present. I am not looking to build Hi-Fi "audiophile" quality bookshelf speakers. But better than common "mass produced" Home Theaters.
P.S.
1. I will paste frequency response curves (if measured by some diy'ers)
2. what is your suggestion on the box dimensions. will they create problems (standingwaves)?
new design
Dear all,
here is my new design as per the suggestions received from experts. its based on peerless india drivers M16GJ 6.5" woofer and TL26SG tweeter. I hope this time I have done it right. Pl go through the attched pdf for details. I have box simulations in unibox, box drawing and crossover design based on rated impedances of drivers.
phew! keeping my fingers crossed so that experts will approve the design....
Dear all,
here is my new design as per the suggestions received from experts. its based on peerless india drivers M16GJ 6.5" woofer and TL26SG tweeter. I hope this time I have done it right. Pl go through the attched pdf for details. I have box simulations in unibox, box drawing and crossover design based on rated impedances of drivers.
phew! keeping my fingers crossed so that experts will approve the design....
My first bookshelf project
Dear all, I do not know where my earlier reply went with modified design using peerless india M16GJ driver. here is posting again.
reg
prasi
Dear all, I do not know where my earlier reply went with modified design using peerless india M16GJ driver. here is posting again.
reg
prasi
Hi Prasi, you can probably drop the tuning frequency down a little. I find when modelling in a small box that the flatest response tends to be when the tuning frequency is close to the F3 Higher than F3 whilst it causes a boost in the lower frequencies (and tends to give slightly better transient response) also reduces your overall power handling, as below the box tuned frequency the drivers excursion increases dramatically.
So I try to get the tuning frequency and the F3 fairly close.
With the crossover calculations it is very difficult to say without an impedance curve. It is actually very easy to measure the impedance of the drivers (once you have them 😀 ) provided you have a sound card with line in. You really only need a cable with a resistor and some software such as ARTA or REW (I prefer REW).
link to appropriate cable for measuring impedance Cables
If you haven't already, have a read of the sticky thread designing crossovers without measurement.
Tony.
So I try to get the tuning frequency and the F3 fairly close.
With the crossover calculations it is very difficult to say without an impedance curve. It is actually very easy to measure the impedance of the drivers (once you have them 😀 ) provided you have a sound card with line in. You really only need a cable with a resistor and some software such as ARTA or REW (I prefer REW).
link to appropriate cable for measuring impedance Cables
If you haven't already, have a read of the sticky thread designing crossovers without measurement.
Tony.
THanks tony,
I have now made a simulation with Fb and F3 fairly close (about 3 Hz). Yes the step response has improves (2-3 wiggles only) . When you drop your Fb close to F3, it only results in increase of port length (for a given dia of port and box volume). Is this the only difference or I am doing something wrong here??
As suggested by you a have read of the cross over thread.
Thanks
Prashant
Hi Prasi, the thing to watch for when dropping the tuning frequency is that you don't get too big a port resonance. The bigger (in area) the port the lower the frequency that the port resonance occurs at. It is all a balancing act. When you optimize one parameter there is a tradeoff somewhere else. You also need to make sure that the port length is "reasonable" ie will work withing the box that you have.
The trick is to work out what tradeoffs you can live with and which you can't and try to get everything to work together to give a good result 🙂
20L is pretty much the limit for this driver I suspect, you could also look at perhaps making the cabinet a bit larger. Play around with different volumes and tunings and see what you get (unless 20L was already your maximum size).
Actually another thing, I can't remember whether you allowed for the volume of your port in the enclosure volume. You need to subtract the volume of the port from the volume of the cabinet (just like you do for the volume displaced by the driver). If you make the port longer it also reduces your volume somewhat.
Tony.
The trick is to work out what tradeoffs you can live with and which you can't and try to get everything to work together to give a good result 🙂
20L is pretty much the limit for this driver I suspect, you could also look at perhaps making the cabinet a bit larger. Play around with different volumes and tunings and see what you get (unless 20L was already your maximum size).
Actually another thing, I can't remember whether you allowed for the volume of your port in the enclosure volume. You need to subtract the volume of the port from the volume of the cabinet (just like you do for the volume displaced by the driver). If you make the port longer it also reduces your volume somewhat.
Tony.
Try also an exponential port, or a hourglass ( clepsydra) shape; other than that, some acoustic labirinth such TL or QW resonators, but you'd say that they're too big : not true!
Since you are not that newbie ( I am ! ) you could also raise the target and go for 3 way.
Then you can concentrate better on bass characteristic. Cheapish 1 $ midrange for portable radio and phenolic tweeters can render a quasi-optimal mid-treble reproduction
Since you are not that newbie ( I am ! ) you could also raise the target and go for 3 way.
Then you can concentrate better on bass characteristic. Cheapish 1 $ midrange for portable radio and phenolic tweeters can render a quasi-optimal mid-treble reproduction
Dear Tony,
I have made simulations based on your suggestions. Earlier I had not subtracted the volume of driver or port!!. Now its taken care of.. And yes 20 ltrs is about my limit.
Another question in continuation with your previous reply. In my simulation, for a 20.2 liter box, (Fb 45 Hz, F3 43 Hz, Port dia 7 cms), the unibox suggests a port length of 22.17 cms. This is too big a length to accommodate in my box. I can accommodate a max length of 6 inch (15.2 cms), considering minimum gap between rear end of port and back wall of speaker cabinet (=1 port dia minimum). 6" is the only port option available at the vendor which is suitable to me. Others are either too small or too big (in dia).
Also please have look at the following website. It says that actual port length required for a damped box is much smaller than the simulated one.
vent tuning
What is your experience or say on this...
If I have to go by the unibox simulation, I have no option but to go for normal PVC pipe with an L-bend and flush mount it on the front baffle. (No Flaring!). Will this work?
Pl suggest me a solution...
I have made simulations based on your suggestions. Earlier I had not subtracted the volume of driver or port!!. Now its taken care of.. And yes 20 ltrs is about my limit.
Another question in continuation with your previous reply. In my simulation, for a 20.2 liter box, (Fb 45 Hz, F3 43 Hz, Port dia 7 cms), the unibox suggests a port length of 22.17 cms. This is too big a length to accommodate in my box. I can accommodate a max length of 6 inch (15.2 cms), considering minimum gap between rear end of port and back wall of speaker cabinet (=1 port dia minimum). 6" is the only port option available at the vendor which is suitable to me. Others are either too small or too big (in dia).
Also please have look at the following website. It says that actual port length required for a damped box is much smaller than the simulated one.
vent tuning
What is your experience or say on this...
If I have to go by the unibox simulation, I have no option but to go for normal PVC pipe with an L-bend and flush mount it on the front baffle. (No Flaring!). Will this work?
Pl suggest me a solution...
Dear Tony,
One more thing, I have designed my crossover based on the sticky thread suggested by you. Hope it works well.
One more thing, I have designed my crossover based on the sticky thread suggested by you. Hope it works well.
Hi Prasi, Troels has way more experience than me! I'd take his word over mine any day 😉 I've only made two bass reflex enclosures. The first is so long ago that I'm not 100% sure whether my sim matched the final result or not as I don't have documentation of what I simmed. I think it was pretty close (but it was a 70L cabinet).
My recent box was a tiny one. 2.5L and the sim was for a 70Hz tuning. I've just had a look at the measurements of the finished box and it shows a tuning freq of about 65 Hz so this tends to fit with Troel's observations that port length does not need to be as long as the simulation predicts. http://www.diyaudio.com/forums/full...-4-full-range-build-thread-4.html#post3598778
I used a slot port for this speaker rather than a tube.
With respect to the length of the port, the tradeoff for going for a smaller diameter is that you will get more port noise when playing at higher volumes. If you don't intend to play at high volume levels then it probably isn't an issue.
Using an L bend with the PVC pipe is definitely an option, I've not done it myself but have seen other projects where this has been done.
Going on Troel's rule of thumb, perhaps multiplying the simmed vent length by 0.7 would be the way to go 🙂 I'll have to measure the length of the vents in my old 3 ways and work out whether the tuning is lower than what a sim would have predicted!
Tony.
My recent box was a tiny one. 2.5L and the sim was for a 70Hz tuning. I've just had a look at the measurements of the finished box and it shows a tuning freq of about 65 Hz so this tends to fit with Troel's observations that port length does not need to be as long as the simulation predicts. http://www.diyaudio.com/forums/full...-4-full-range-build-thread-4.html#post3598778
I used a slot port for this speaker rather than a tube.
With respect to the length of the port, the tradeoff for going for a smaller diameter is that you will get more port noise when playing at higher volumes. If you don't intend to play at high volume levels then it probably isn't an issue.
Using an L bend with the PVC pipe is definitely an option, I've not done it myself but have seen other projects where this has been done.
Going on Troel's rule of thumb, perhaps multiplying the simmed vent length by 0.7 would be the way to go 🙂 I'll have to measure the length of the vents in my old 3 ways and work out whether the tuning is lower than what a sim would have predicted!
Tony.
Hi Prassi, I measured my vent's actual length and diameter for one of my old three ways. Put in the T/S params of my woofer and modeled.
70L box 2 X 63mm id vents 203mm long. simulation shows a box tuning of 32.6 Hz with those dimentions.
Actual box tuning is 30Hz (as determined from impedance measurement) so it is lower than expected from the sim.
30Hz tuning in the sim says 247mm port length. So in my case multiplier is 0.82 However I am not 100% certain about the cabinet volume, it may be a bit more or a bit less than 70L. But this does demonstrate that my builds have followed what troels study shows, ie the box tuning ends up lower than predicted by the sim for a given port length.
Tony.
70L box 2 X 63mm id vents 203mm long. simulation shows a box tuning of 32.6 Hz with those dimentions.
Actual box tuning is 30Hz (as determined from impedance measurement) so it is lower than expected from the sim.
30Hz tuning in the sim says 247mm port length. So in my case multiplier is 0.82 However I am not 100% certain about the cabinet volume, it may be a bit more or a bit less than 70L. But this does demonstrate that my builds have followed what troels study shows, ie the box tuning ends up lower than predicted by the sim for a given port length.
Tony.
Dear Tony,
Thanks for your inputs. So this does prove Troels observation. I will go ahead with shorter port length for now at least, with a multiplier of 0.7.
Then make additional front baffles with provision to flush mount PVC pipes of different lengths and dia's.
Thanks for your inputs. So this does prove Troels observation. I will go ahead with shorter port length for now at least, with a multiplier of 0.7.
Then make additional front baffles with provision to flush mount PVC pipes of different lengths and dia's.
- Status
- Not open for further replies.