Hi guys,
I have difficulty to understand graphs in MJK mathcad.
I quite familiar with graph #6 only!
Maybe somebody can help us or at least me, to draw how does a "perfect graph" (= perfect result) look-like for each graph, so we can design our TL speaker close to that "perfect graph".
Or maybe with some explanation for each graph what should to reach and what should to avoid.
Please guys, I desperate need this help to design "the best TL in the universe"...
MJK, would you mind...?
Thankyou.
Chris.
I have difficulty to understand graphs in MJK mathcad.
I quite familiar with graph #6 only!
Maybe somebody can help us or at least me, to draw how does a "perfect graph" (= perfect result) look-like for each graph, so we can design our TL speaker close to that "perfect graph".
Or maybe with some explanation for each graph what should to reach and what should to avoid.
Please guys, I desperate need this help to design "the best TL in the universe"...
MJK, would you mind...?
Thankyou.
Chris.
Couple of points and some background on the graphs in MathCad.
1. Perfect speaker - I don't know what the perfect speaker is supposed to look like, everything is about compromises. You trade one aspect of perfromance for another and the best balance for one person/system/room is probably not the best balance for somebody else. Sometimes you can get into analysis paralysis by trying to perfect every wiggle in the plot. At some point just jump in and build the box, you can always get more wood and try again if you are unhappy with the results. Tuning by stuffing adjustment and BSC circuits gives the designer a lot of adjustment flexibility.
The plots :
1. Acousitc Impedance plots - if you set the stuffing density to zero you can see the peaks in this plot that are the quarter wavelength resonant frequencies of your design. The first peak is the tuning frequency. I use this plot to tune the enclosure relative to the driver resonance.
2. Velocity at the Terminus plots - this plot means something to me, but if I were a user I would skip over this result.
3. Far Field TL System SPL Response plots - This is the bottom line on how extended the bass resposne will be for your design. Keep in mind that in reality you will experience a 3 - 4 dB loss of bass at 300 to 600 Hz depending on your baffle width. This is the baffle step problem that I use correction circuits to rectify.
4. Woofer anf Terminus Far Field SPL Response plots - basically this plot will help identify the causes for any ripples in the summed SPL response just above. This is a very useful plot that you can watch as you change dimensions and stuffing density and location to try and move or tame particular peaks in the terminus plot. This is probably the second most important set of plots.
5. Impedance plot - this is one nice plot to use if you want to measure your results and don't have a computer measurement program. If you use a scope, multi-meter, signal generator, and frequency counter you can get a good double check on the correlation between your simulation and actual finished speaker system. The older editions of the Loudspealer Design Cookbook used ot have a section on doing this type of testing, don't know about the latest edition.
6. Woofer Displacement plot - this is the RMS displacement of the woofer as a function of frequency. If you are somebody who worries about the Xmax of your speaker, this is an important plot. I personally don't spend much time sweating Xmax but that is a different topic.
7. System Impulse Response plot - this is the time domain pressure response due to a perfect voltage impulse. If you see a signifigant amount of ringing or reflecion pulses in this plot then your design will probably exhibit a poor bass response.
Hope that helps,
1. Perfect speaker - I don't know what the perfect speaker is supposed to look like, everything is about compromises. You trade one aspect of perfromance for another and the best balance for one person/system/room is probably not the best balance for somebody else. Sometimes you can get into analysis paralysis by trying to perfect every wiggle in the plot. At some point just jump in and build the box, you can always get more wood and try again if you are unhappy with the results. Tuning by stuffing adjustment and BSC circuits gives the designer a lot of adjustment flexibility.
The plots :
1. Acousitc Impedance plots - if you set the stuffing density to zero you can see the peaks in this plot that are the quarter wavelength resonant frequencies of your design. The first peak is the tuning frequency. I use this plot to tune the enclosure relative to the driver resonance.
2. Velocity at the Terminus plots - this plot means something to me, but if I were a user I would skip over this result.
3. Far Field TL System SPL Response plots - This is the bottom line on how extended the bass resposne will be for your design. Keep in mind that in reality you will experience a 3 - 4 dB loss of bass at 300 to 600 Hz depending on your baffle width. This is the baffle step problem that I use correction circuits to rectify.
4. Woofer anf Terminus Far Field SPL Response plots - basically this plot will help identify the causes for any ripples in the summed SPL response just above. This is a very useful plot that you can watch as you change dimensions and stuffing density and location to try and move or tame particular peaks in the terminus plot. This is probably the second most important set of plots.
5. Impedance plot - this is one nice plot to use if you want to measure your results and don't have a computer measurement program. If you use a scope, multi-meter, signal generator, and frequency counter you can get a good double check on the correlation between your simulation and actual finished speaker system. The older editions of the Loudspealer Design Cookbook used ot have a section on doing this type of testing, don't know about the latest edition.
6. Woofer Displacement plot - this is the RMS displacement of the woofer as a function of frequency. If you are somebody who worries about the Xmax of your speaker, this is an important plot. I personally don't spend much time sweating Xmax but that is a different topic.
7. System Impulse Response plot - this is the time domain pressure response due to a perfect voltage impulse. If you see a signifigant amount of ringing or reflecion pulses in this plot then your design will probably exhibit a poor bass response.
Hope that helps,
Hi Bricolo,
Thanks for your advice.
Hi MJK,
Wow, thankyou, now I'm gonna make a perfect speaker, at least for me at specific time until I hear something better, then it becomes no longer perfect.
About your graph:
1. fully understand.
2. skip as you said.
3. how if, instead of flat, I raise it +3dB toward TL resonant freq.?
4. OK.
5. There are 2 graphs, what should to achieve and to avoid for each graph result?
6. leave it.
7. how high or low for the voltage is good and how many ringing is bad?
Almost.....
Chris.
Thanks for your advice.
Hi MJK,
Wow, thankyou, now I'm gonna make a perfect speaker, at least for me at specific time until I hear something better, then it becomes no longer perfect.
About your graph:
1. fully understand.
2. skip as you said.
3. how if, instead of flat, I raise it +3dB toward TL resonant freq.?
4. OK.
5. There are 2 graphs, what should to achieve and to avoid for each graph result?
6. leave it.
7. how high or low for the voltage is good and how many ringing is bad?
Almost.....
Chris.
Hi Chris,
3. how if, instead of flat, I raise it +3dB toward TL resonant freq.?
(MJK) You can do this but my guess is that the impulse resposne will start to ring a little. I have never tried this approach.
5. There are 2 graphs, what should to achieve and to avoid for each graph result?
(MJK) Watch the graphs as you move the woofer along the line and add or increase stuffing at different locations. The goal is to have a strong first peak in the terminus to augment the driver's rolling off bass bass, then damp out the other peaks to reduce ripples. Play with the program with an example and see what can be achieved. Other then that I don't have any firm rules. Don't get too fixated on perfection, the worksheet probably over predicts the ripple a little and your room is going to do a lot more damage to the SPL response.
7. how high or low for the voltage is good and how many ringing is bad?
(MJK) Don't worry about the absolute magnitude, look at the shape of the time trace and make sure you don't get a lot of oscillations or reflection pulses.
Hope that helps,
3. how if, instead of flat, I raise it +3dB toward TL resonant freq.?
(MJK) You can do this but my guess is that the impulse resposne will start to ring a little. I have never tried this approach.
5. There are 2 graphs, what should to achieve and to avoid for each graph result?
(MJK) Watch the graphs as you move the woofer along the line and add or increase stuffing at different locations. The goal is to have a strong first peak in the terminus to augment the driver's rolling off bass bass, then damp out the other peaks to reduce ripples. Play with the program with an example and see what can be achieved. Other then that I don't have any firm rules. Don't get too fixated on perfection, the worksheet probably over predicts the ripple a little and your room is going to do a lot more damage to the SPL response.
7. how high or low for the voltage is good and how many ringing is bad?
(MJK) Don't worry about the absolute magnitude, look at the shape of the time trace and make sure you don't get a lot of oscillations or reflection pulses.
Hope that helps,
-------------------------------------------------------------------------------
1. Acoustic Impedance plots - if you set the stuffing density to zero you can see the peaks in this plot that are the quarter wavelength resonant frequencies of your design. The first peak is the tuning frequency. I use this plot to tune the enclosure relative to the driver resonance.
---------------------------------------------------------------------------------
Hello MJK,
I have a question regarding the tuning frequency.
I have a driver with fc=60 Hz. Is any problem if my enclosure is tuned at 40Hz? SPL look very linear and I used your latest Ported Box sheet.
I'm asking this because I saw in this thread that you recommend tuning the enclosure close to the driver fc (if I understood correctly).
Thanks,
tda
1. Acoustic Impedance plots - if you set the stuffing density to zero you can see the peaks in this plot that are the quarter wavelength resonant frequencies of your design. The first peak is the tuning frequency. I use this plot to tune the enclosure relative to the driver resonance.
---------------------------------------------------------------------------------
Hello MJK,
I have a question regarding the tuning frequency.
I have a driver with fc=60 Hz. Is any problem if my enclosure is tuned at 40Hz? SPL look very linear and I used your latest Ported Box sheet.
I'm asking this because I saw in this thread that you recommend tuning the enclosure close to the driver fc (if I understood correctly).
Thanks,
tda
Hi tda,
"I have a driver with fc=60 Hz. Is any problem if my enclosure is tuned at 40Hz? SPL look very linear and I used your latest Ported Box sheet.
I'm asking this because I saw in this thread that you recommend tuning the enclosure close to the driver fc (if I understood correctly)."
In most cases tuning near the driver fs works well. However, nothing says that you cannot tune above or below fs. If the SPL response looks good and nothing looks way off in the other graphs, then I don't see any reason not to use the design you describe.
"I have a driver with fc=60 Hz. Is any problem if my enclosure is tuned at 40Hz? SPL look very linear and I used your latest Ported Box sheet.
I'm asking this because I saw in this thread that you recommend tuning the enclosure close to the driver fc (if I understood correctly)."
In most cases tuning near the driver fs works well. However, nothing says that you cannot tune above or below fs. If the SPL response looks good and nothing looks way off in the other graphs, then I don't see any reason not to use the design you describe.
-------------------------------------------------------------------------------
If the SPL response looks good and nothing looks way off in the other graphs, then I don't see any reason not to use the design you describe.
-------------------------------------------------------------------------------
What do you mean: "nothing looks way off in the other graphs”?
Beside SPL graph what else should I check to see if everything is right?
Thanks
If the SPL response looks good and nothing looks way off in the other graphs, then I don't see any reason not to use the design you describe.
-------------------------------------------------------------------------------
What do you mean: "nothing looks way off in the other graphs”?
Beside SPL graph what else should I check to see if everything is right?
Thanks
"What do you mean: "nothing looks way off in the other graphs”?
Beside SPL graph what else should I check to see if everything is right?"
tda,
You asked a very general question without any real information. I don't know what the driver is or what your design looks like. So I gave you a very general answer, yes you can tune below fs and it will work fine if everything else looks OK. I cannot give a definitive response based on your question. Please recognize that I really don't want to be in the roll of reviewing other people's designs and pronouncing them good or bad. I am not that smart.
I am starting to see a trend in the recent discussions that people want to achieve a "perfect" result or have an absolute answer as to which potential TL design is best. The MathCad worksheets will go a long way towards producing a good TL design but they cannot by themselves produce a great result. The selection of the drivers, the enclosure construction details, the quality and accuracy of the T/S paramaters input, the skill and experience of the DIY in creating the simulation (remember GIGO), and the final room/system are just as important. Please do not get too focused on tweaking the MathCad simulation to produce exact "perfection" at the expense of these other issues. At some point, the improvements in the MathCad simulation result are small compared to all of the other issues that deserve more attention.
If you have not measured the T/S parameters for your drivers, then the errors in your simulation are already larger then the improvements being made by continuing to tweak the paper design. This issue is not unique to the MathCad worksheets but is true for any simulation program that you might use to design an enclosure..
Analysis-paralysis can kill a project. Jump in and build something. Measure the results and compare against the simulation predictions. Make intelligent changes to the simulation to try and correlate against the measured data and learn what was important and what you missed in the original computer model. You can only learn from experience.
Beside SPL graph what else should I check to see if everything is right?"
tda,
You asked a very general question without any real information. I don't know what the driver is or what your design looks like. So I gave you a very general answer, yes you can tune below fs and it will work fine if everything else looks OK. I cannot give a definitive response based on your question. Please recognize that I really don't want to be in the roll of reviewing other people's designs and pronouncing them good or bad. I am not that smart.
I am starting to see a trend in the recent discussions that people want to achieve a "perfect" result or have an absolute answer as to which potential TL design is best. The MathCad worksheets will go a long way towards producing a good TL design but they cannot by themselves produce a great result. The selection of the drivers, the enclosure construction details, the quality and accuracy of the T/S paramaters input, the skill and experience of the DIY in creating the simulation (remember GIGO), and the final room/system are just as important. Please do not get too focused on tweaking the MathCad simulation to produce exact "perfection" at the expense of these other issues. At some point, the improvements in the MathCad simulation result are small compared to all of the other issues that deserve more attention.
If you have not measured the T/S parameters for your drivers, then the errors in your simulation are already larger then the improvements being made by continuing to tweak the paper design. This issue is not unique to the MathCad worksheets but is true for any simulation program that you might use to design an enclosure..
Analysis-paralysis can kill a project. Jump in and build something. Measure the results and compare against the simulation predictions. Make intelligent changes to the simulation to try and correlate against the measured data and learn what was important and what you missed in the original computer model. You can only learn from experience.
Hi MJK,
I hope you're not tired to answer my questions,
5. (MJK) The goal is to have a strong first peak in the terminus to augment the driver's rolling off bass...
There are 2 graphs, which graph did you pointed, phase or impedance?
I made a simulation, I made TL resonant freq. 20% below Fs and the SPL response looks good, however TL box total volume= 4,5x Vas! Is it still OK, or too big?
Is there any optimum number for TL box total volume agains Vas?
Thankyou,
Chris.
I hope you're not tired to answer my questions,
5. (MJK) The goal is to have a strong first peak in the terminus to augment the driver's rolling off bass...
There are 2 graphs, which graph did you pointed, phase or impedance?
I made a simulation, I made TL resonant freq. 20% below Fs and the SPL response looks good, however TL box total volume= 4,5x Vas! Is it still OK, or too big?
Is there any optimum number for TL box total volume agains Vas?
Thankyou,
Chris.
Martin,
This is the SPL graph I get with a SEAS H1217 in a TL tuned to 35Hz with your alignment table (very nice tool!
)
And here are the input parameters (I used your ofset driver worksheet)
How can I tune this to cancel the resonances?
I tried to add some stuffing, but I had to add 0.6 or more, and this gave a SPL curve I didn't like (very close to the IB)
PS: if I make a excell sheet for your alignment table, do you allow me to post it here?
This is the SPL graph I get with a SEAS H1217 in a TL tuned to 35Hz with your alignment table (very nice tool!
)An externally hosted image should be here but it was not working when we last tested it.
And here are the input parameters (I used your ofset driver worksheet)
An externally hosted image should be here but it was not working when we last tested it.
How can I tune this to cancel the resonances?
I tried to add some stuffing, but I had to add 0.6 or more, and this gave a SPL curve I didn't like (very close to the IB)
PS: if I make a excell sheet for your alignment table, do you allow me to post it here?
Hi Chris,
5. (MJK) The goal is to have a strong first peak in the terminus to augment the driver's rolling off bass...
There are 2 graphs, which graph did you pointed, phase or impedance?
(MJK) I was refering to the SPL graph, the lower of the two. The terminus response is the blue curve. If you set the stuffing to zero in the "TL Offset Driver" worksheet you will see a hump in the terminus blue curve at approx 48 Hz and a deep dip in the driver red curve at approx 42 Hz.
I made a simulation, I made TL resonant freq. 20% below Fs and the SPL response looks good, however TL box total volume= 4,5x Vas! Is it still OK, or too big?
Is there any optimum number for TL box total volume agains Vas?
(MJK) I have never tracked the volume of the TL compared to Vas so I am not sure.
5. (MJK) The goal is to have a strong first peak in the terminus to augment the driver's rolling off bass...
There are 2 graphs, which graph did you pointed, phase or impedance?
(MJK) I was refering to the SPL graph, the lower of the two. The terminus response is the blue curve. If you set the stuffing to zero in the "TL Offset Driver" worksheet you will see a hump in the terminus blue curve at approx 48 Hz and a deep dip in the driver red curve at approx 42 Hz.
I made a simulation, I made TL resonant freq. 20% below Fs and the SPL response looks good, however TL box total volume= 4,5x Vas! Is it still OK, or too big?
Is there any optimum number for TL box total volume agains Vas?
(MJK) I have never tracked the volume of the TL compared to Vas so I am not sure.
Hi Bricolo,
How can I tune this to cancel the resonances?
I tried to add some stuffing, but I had to add 0.6 or more, and this gave a SPL curve I didn't like (very close to the IB)
(MJK) Your pictures do not show up on my PC so I am not exactly sure what the curves look like. But my guess is that you have stuffed the entire length of the TL. If you only stuff the first 2/3 of the TL you should be able to attenuate the unwanted peaks and maintain the bass response.
PS: if I make a excell sheet for your alignment table, do you allow me to post it here?
(MJK) I have no problem with people turning the alignment tables into an Excel worksheet and using that for DIY projects. I know of one version already. The only thing I would ask is that the original article be refereced (or linked) in the worksheet. If you want to distribute the spreadsheet for DIYers to use that would also be acceptable.
How can I tune this to cancel the resonances?
I tried to add some stuffing, but I had to add 0.6 or more, and this gave a SPL curve I didn't like (very close to the IB)
(MJK) Your pictures do not show up on my PC so I am not exactly sure what the curves look like. But my guess is that you have stuffed the entire length of the TL. If you only stuff the first 2/3 of the TL you should be able to attenuate the unwanted peaks and maintain the bass response.
PS: if I make a excell sheet for your alignment table, do you allow me to post it here?
(MJK) I have no problem with people turning the alignment tables into an Excel worksheet and using that for DIY projects. I know of one version already. The only thing I would ask is that the original article be refereced (or linked) in the worksheet. If you want to distribute the spreadsheet for DIYers to use that would also be acceptable.
Chris,
I had one other thought. If you have not already done this, you should read Bob Brines' entire site. Bob spent a lot of time running simulations using the MathCad worksheets and has many examples that demonstrate the behavior of quarter-wave pipes and how to design and optimize them. He has done a much better job convering these details then I have on my site. His site is linked from mine.
I had one other thought. If you have not already done this, you should read Bob Brines' entire site. Bob spent a lot of time running simulations using the MathCad worksheets and has many examples that demonstrate the behavior of quarter-wave pipes and how to design and optimize them. He has done a much better job convering these details then I have on my site. His site is linked from mine.
Oops! I've just realized that the alignment tables have been updated (the 10/16/2003).
I used the first ones, and the new ones have some big differences. The equations have also some changes.
I'll update my excell sheet, and port a new one.
Martin: the nex tables gives me different results, is this normal?
For the same driver, tuned to 35Hz, I got
Property Old New
Effective length 76in 76.3in
ksi 24.2 24.4
S0 6.28.Sd 4.15.Sd
SL 2.1.Sd 1.38.Sd
Is this normal?
I used the first ones, and the new ones have some big differences. The equations have also some changes.
I'll update my excell sheet, and port a new one.
Martin: the nex tables gives me different results, is this normal?
For the same driver, tuned to 35Hz, I got
Property Old New
Effective length 76in 76.3in
ksi 24.2 24.4
S0 6.28.Sd 4.15.Sd
SL 2.1.Sd 1.38.Sd
Is this normal?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.