Well.... the best one I have ever done anyway.
I have downloaded and played with lots and lots of them, and the simple ones at least, all seem to have some shortcoming. for instance, the majority of them calculate Vas by the put-the-driver-on-a-box method, which I find inconvenient. Not as many have the added mass method. Also, most use the drop in current method of finding resonance and this seems to have all these yucky intermediate steps that make my head spin.
The other Fs measurement method that uses a high value resistor to feed the speaker and measures the rise in voltage across the voice coil at resonance seems to be in the minority, and of course there is the error introduced by this method if the speaker impedance goes particularly high.
So... I sat down and wrote my own spreadsheet. It won't set the world on fire but it does just exactly what I need. Feed the speaker from an amplifier through a 1k or whatever resistor. Tell the spreadsheet the voltage coming out of the amplifier, and the voltage across the speaker etc. Any impedance measurement error both at Fs and -3dB frequencies are compensated for.
***********
For an accurate and flat response signal source I burnt a CD with 8 tracks having a linear (not log!) sweeping sinewave of 10-20 Hz, 20-30Hz etc. and each track is 10 minutes long i.e 1 minute per Hz. This means that if the frequency of interest is passed on track 4 at 8 min 30 sec then the frequency was 48.5 Hz. No frequency counter necessary!
On the CD player and looking at the output with a scope, the level varied maybe 1 part in 200 or less across the spectrum. Way better than some toy opamp oscillator. Way better than any soundcard I have too. They're a big disappointment.
I have downloaded and played with lots and lots of them, and the simple ones at least, all seem to have some shortcoming. for instance, the majority of them calculate Vas by the put-the-driver-on-a-box method, which I find inconvenient. Not as many have the added mass method. Also, most use the drop in current method of finding resonance and this seems to have all these yucky intermediate steps that make my head spin.
The other Fs measurement method that uses a high value resistor to feed the speaker and measures the rise in voltage across the voice coil at resonance seems to be in the minority, and of course there is the error introduced by this method if the speaker impedance goes particularly high. So... I sat down and wrote my own spreadsheet. It won't set the world on fire but it does just exactly what I need. Feed the speaker from an amplifier through a 1k or whatever resistor. Tell the spreadsheet the voltage coming out of the amplifier, and the voltage across the speaker etc. Any impedance measurement error both at Fs and -3dB frequencies are compensated for.
***********
For an accurate and flat response signal source I burnt a CD with 8 tracks having a linear (not log!) sweeping sinewave of 10-20 Hz, 20-30Hz etc. and each track is 10 minutes long i.e 1 minute per Hz. This means that if the frequency of interest is passed on track 4 at 8 min 30 sec then the frequency was 48.5 Hz. No frequency counter necessary!
On the CD player and looking at the output with a scope, the level varied maybe 1 part in 200 or less across the spectrum. Way better than some toy opamp oscillator. Way better than any soundcard I have too. They're a big disappointment.
In the previous post where I said you feed the speaker from an amplifier through a 1k or whatever resistor, it seems to me that this resistor would damp the impedance peak somewhat and so it should be as high a resistance as possible. Of course the higher you make it the less drive you have to the loudspeaker and after a while it gets difficult to measure with any reasonable accuracy.
What I want to know then, is there a way to estimate the damping effect of this feed resistor so that it's effects can be accounted for; so that the calculated result would be as if the resistor was infinitely large?
One other way perhaps would be to use the voltage across the speaker to control a VCA stage so that the voltage across the resistor remains virtually constant. Or maybe use a current-source amp without any resistor.
What I want to know then, is there a way to estimate the damping effect of this feed resistor so that it's effects can be accounted for; so that the calculated result would be as if the resistor was infinitely large?
One other way perhaps would be to use the voltage across the speaker to control a VCA stage so that the voltage across the resistor remains virtually constant. Or maybe use a current-source amp without any resistor.
thats it, AC curent source.....
but I think when that you use 10k, its precise enough, with 100 ohms impedance peak of the driver this is only 1% deviation.....
You need an amp with nice high voltage (100Vrms orso
) to drive the thing.... can be just an ordinary opamp circuit with 2 pp high voltage transistors running on +/- 150V...
Is there is someone with an idea of a good/simple lineair ac current source? it is a lot nicer than the high voltage, it could run on a battery....
By the way: I like the spreadsheet circlo!....
but I think when that you use 10k, its precise enough, with 100 ohms impedance peak of the driver this is only 1% deviation.....
You need an amp with nice high voltage (100Vrms orso
) to drive the thing.... can be just an ordinary opamp circuit with 2 pp high voltage transistors running on +/- 150V... Is there is someone with an idea of a good/simple lineair ac current source? it is a lot nicer than the high voltage, it could run on a battery....
By the way: I like the spreadsheet circlo!....
Some real-word stuff
Did some measurements on this el-cheapo 12" driver and the Vas result is a bit odd. The results are as follows:
Source Qes Qms Qts Fs Vas
Manuf. 2.32 4.0 1.46 32 166
Spkr W/S 1.26 2.74 0.86 34.16 141.3
me 1.19 2.84 0.84 34.1 72.1
Spkr W/S=Speaker Workshop. me=me & my spreadsheet.
The Vas calc part of the spreadsheet I grafted off someone else's so I can't be certain of it's accuracy. Can anyone give a second opinion?
Did some measurements on this el-cheapo 12" driver and the Vas result is a bit odd. The results are as follows:
Source Qes Qms Qts Fs Vas
Manuf. 2.32 4.0 1.46 32 166
Spkr W/S 1.26 2.74 0.86 34.16 141.3
me 1.19 2.84 0.84 34.1 72.1
Spkr W/S=Speaker Workshop. me=me & my spreadsheet.
The Vas calc part of the spreadsheet I grafted off someone else's so I can't be certain of it's accuracy. Can anyone give a second opinion?
Hi Circlotron,
What added mass did you use, and what was your fs1 with the added mass?
Formulas I have from Weems are:
Mass of cone M = M1 /(fs/fs1)^2 -1 where M1 is the added mass fs1 is the fs with added mass
Compliance Cms = 1/((6.28 * fs)^2*M)
effective cone area A = 3.1416 * r^2
Vas = Cms*d*C^2*A^2
where d = density of air 0.00129 g/cc and c = speed of sound 34400 cm/sec. He then says that for VAS in litres divide by 1000.
He also measures the diameter of the speaker as being from one edge (including surround) to the start of the surround on the other side (for working out r above).........
I'd be interested to know the results. When I measured my Vifa 10" drivers with SW the manufacturers spec was 130L Vas, but I got 125L with SW.
Tony.
What added mass did you use, and what was your fs1 with the added mass?
Formulas I have from Weems are:
Mass of cone M = M1 /(fs/fs1)^2 -1 where M1 is the added mass fs1 is the fs with added mass
Compliance Cms = 1/((6.28 * fs)^2*M)
effective cone area A = 3.1416 * r^2
Vas = Cms*d*C^2*A^2
where d = density of air 0.00129 g/cc and c = speed of sound 34400 cm/sec. He then says that for VAS in litres divide by 1000.
He also measures the diameter of the speaker as being from one edge (including surround) to the start of the surround on the other side (for working out r above).........
I'd be interested to know the results. When I measured my Vifa 10" drivers with SW the manufacturers spec was 130L Vas, but I got 125L with SW.
Tony.
Hi Tony. I have been changing it in the meantime of course! Your one does give better results but there was an error in the cone mass formula. To check, when you double the moving mass the Fs should decreases to root 2/2 of the original.
The formulas below, the upper one is mine, the lower is yours.
Cone mass
=G4/(G6/G7)^2-1
=G4/((G6/G7)^2-1)
Vas
=SQRT(2)*1000*(G18^2)*G20 (my new formula)
=G20*0.00129*(34400^2)*((G19/1000)^2)
Compliance
=1/(((3.14159265*2*Fs)^2)*(G17)) G17=cone mass
=1/ ((3.14159265*2*G6)^2)*G4 G4=added mass
For the compliance I have used cone mass, you have used added mass. What do you think? It is not the exact formula I used but I juggled it around to match the format of yours for easy comparison.
The Vas's give slightly fifferent results, perhaps because of the compliance stuff?
The formulas below, the upper one is mine, the lower is yours.
Cone mass
=G4/(G6/G7)^2-1
=G4/((G6/G7)^2-1)
Vas
=SQRT(2)*1000*(G18^2)*G20 (my new formula)
=G20*0.00129*(34400^2)*((G19/1000)^2)
Compliance
=1/(((3.14159265*2*Fs)^2)*(G17)) G17=cone mass
=1/ ((3.14159265*2*G6)^2)*G4 G4=added mass
For the compliance I have used cone mass, you have used added mass. What do you think? It is not the exact formula I used but I juggled it around to match the format of yours for easy comparison.
The Vas's give slightly fifferent results, perhaps because of the compliance stuff?
hi
this is a bit of the thread
i was using the woofer tester _ and now its not operating properly _ and a upgradation is long over due
i am planning on the loudspeaker lab (swedish) software + add on unit
how would you rate it , and could you recommend any other such units in the same price range , this is about 400 usd including the mic and add on unit
thanks
this is a bit of the thread
i was using the woofer tester _ and now its not operating properly _ and a upgradation is long over due
i am planning on the loudspeaker lab (swedish) software + add on unit
how would you rate it , and could you recommend any other such units in the same price range , this is about 400 usd including the mic and add on unit
thanks
I obviously shouldn't try and do formulas before having a coffee in the morning ..... Definitely should have been the cone mass in the formula.
I'm confused about the calculation of cone mass. I thought the upper one was yours and the lower one mine (well at least David B Weem's) edit: ahhh you posted again between when I read your first post and I started writing this..... I went off to watch the fireworks halfway through.....
I just went and looked at the 1st, 2nd and 3rd spread sheets and I'm pretty sure the one I put in from weem's book was the upper one. I don't know if it is "correct"....... Well I just looked it up in The loudspeaker design cookbook, and Vance Dickason agrees with your formula, seems there is a typo in my weems book, maybe time for me to check for erata!
Also your new formula for VAS is almost the same as Vance Dickasons, only difference being he multiplies by 1,000,000 not 1,000 but I suspect this has to do with him using sqaure meters or something for cone area and cubic meters for Vas.
I'm now wondering why dickason and weems have such different formulas for the Vas, and whether one is "more correct" than the other..... If I had to guess I would say maybe dickasons was an approximation, but not being good with math, or understanding the physics, who knows, I only say that because Weems is using constants like the speed of sound and the density of air, but I'm only guessing.... sqrt of 2 * 1000 comes out quite a bit different to (0.00129 * 34400^2)/1000 which is where I think the difference in the two VAS figures is comming from.
I hadn't looked at dickasons formulas before, as I didn't have the book last time I was testing speakers, but there is a hell of a lot more (complicated) detail than in my weems book. One thing he says which is a bit of a worry is that when measuring the fs with added mass you need to be acurate to within 0.1Hz.......
Looks like you have been busy adding more stuff too.
Happy New Year BTW
Tony.
..... Definitely should have been the cone mass in the formula. I'm confused about the calculation of cone mass. I thought the upper one was yours and the lower one mine (well at least David B Weem's)
edit: ahhh you posted again between when I read your first post and I started writing this..... I went off to watch the fireworks halfway through.....I just went and looked at the 1st, 2nd and 3rd spread sheets and I'm pretty sure the one I put in from weem's book was the upper one. I don't know if it is "correct"....... Well I just looked it up in The loudspeaker design cookbook, and Vance Dickason agrees with your formula
, seems there is a typo in my weems book, maybe time for me to check for erata!Also your new formula for VAS is almost the same as Vance Dickasons, only difference being he multiplies by 1,000,000 not 1,000 but I suspect this has to do with him using sqaure meters or something for cone area and cubic meters for Vas.
I'm now wondering why dickason and weems have such different formulas for the Vas, and whether one is "more correct" than the other..... If I had to guess I would say maybe dickasons was an approximation, but not being good with math, or understanding the physics, who knows, I only say that because Weems is using constants like the speed of sound and the density of air, but I'm only guessing.... sqrt of 2 * 1000 comes out quite a bit different to (0.00129 * 34400^2)/1000 which is where I think the difference in the two VAS figures is comming from.
I hadn't looked at dickasons formulas before, as I didn't have the book last time I was testing speakers, but there is a hell of a lot more (complicated) detail than in my weems book. One thing he says which is a bit of a worry is that when measuring the fs with added mass you need to be acurate to within 0.1Hz.......
Looks like you have been busy adding more stuff too.
Happy New Year BTW
Tony.
Hi,
in my HP in the download folder, there is an Excel file (Data Check) that makes all the calculation for T/S parameters and more, with all the formulas used, knowing the impedance responses with and without added mass. An explanation is visible in the SW Impedance/Verify folder of the site. Hope it helps.
Regards and Happy New Year!!!
Claudio
in my HP in the download folder, there is an Excel file (Data Check) that makes all the calculation for T/S parameters and more, with all the formulas used, knowing the impedance responses with and without added mass. An explanation is visible in the SW Impedance/Verify folder of the site. Hope it helps.
Regards and Happy New Year!!!
Claudio
Hi Claudio,
Your spreadsheet is protected and needs a password to turn the protection off I'm assuming that is why I can't see the formulas on the free air sheet.......
Tony.
edit: Oh and one other thing Circlotron, which I meant to mention before. Dickason says to use the lowest voltage possible to do the testing.
Your spreadsheet is protected and needs a password to turn the protection off
I'm assuming that is why I can't see the formulas on the free air sheet.......Tony.
edit: Oh and one other thing Circlotron, which I meant to mention before. Dickason says to use the lowest voltage possible to do the testing.
Hi Claudio,
I followed your step by step to measure driver T/S paramater and got the result.
However, when I measure Vas using added mass, in your data check.xls, the BL result showed different result, I used this formula:
Cms=Vas/(Sd^2*rho*c^2)
Mms=1/[(2*pi*Fs)^2*Cms]
BL=[(2*Pi*Fs*Re*Mms)/Qes]^0.5
your xls result 30% higher than that formula.
BTW, do you know how the easy way to measure:
Voice coil inductance (Le)
Voice coil Series (L1)
Voice coil Series (R1)
It's all for inputing Speaker Workshop driver parameters.
Thankyou.
Chris.
I followed your step by step to measure driver T/S paramater and got the result.
However, when I measure Vas using added mass, in your data check.xls, the BL result showed different result, I used this formula:
Cms=Vas/(Sd^2*rho*c^2)
Mms=1/[(2*pi*Fs)^2*Cms]
BL=[(2*Pi*Fs*Re*Mms)/Qes]^0.5
your xls result 30% higher than that formula.
BTW, do you know how the easy way to measure:
Voice coil inductance (Le)
Voice coil Series (L1)
Voice coil Series (R1)
It's all for inputing Speaker Workshop driver parameters.
Thankyou.
Chris.
I've been going over mine with a fine tooth comb lately, fiddling bits and pieces and now I'm pretty confident it's airtight. Naturally I just added a BL calc.
To everybody who downloaded an earlier version, especially the first one (and I'm amazed at how many who did, BTW
) dump it and get this one instead.
To everybody who downloaded an earlier version, especially the first one (and I'm amazed at how many who did, BTW
) dump it and get this one instead.
Hi Ciclotron,
looks like mine XLS helped you, isn't it?
Check your EFF. CONE DIAMETER, it shoul be mm. not cm.
Chris,
BL=[(2*Pi*Fs*Re*Mms)/Qes]^0.5 is the same of BL=2,51*�ã(Fs*Re*Mms/Qes) that is the one I used.
Now your BL value differs of 30% from what? The manufacturer one?
SW doesn't calculate BxL, so in my spreadsheet I calculate its value using the parameters founded with SW, and compare it with the manufacturer value.
About Voice coil inductance (Le), Voice coil Series (L1), Voice coil Series (R1), Speaker Workshop calculate them , you don't need to input them.
Regards
Claudio
looks like mine XLS helped you, isn't it?
Check your EFF. CONE DIAMETER, it shoul be mm. not cm.
Chris,
BL=[(2*Pi*Fs*Re*Mms)/Qes]^0.5 is the same of BL=2,51*�ã(Fs*Re*Mms/Qes) that is the one I used.
Now your BL value differs of 30% from what? The manufacturer one?
SW doesn't calculate BxL, so in my spreadsheet I calculate its value using the parameters founded with SW, and compare it with the manufacturer value.
About Voice coil inductance (Le), Voice coil Series (L1), Voice coil Series (R1), Speaker Workshop calculate them , you don't need to input them.
Regards
Claudio
Hi Claudio,
the different is from Mms formula you use:
Mms=added mass weight/[(fs/Fsm)^2-1]
I use:
Mms=1/[(2*pi*Fs)^2*Cms]
Cms came from:
Cms=Vas/(Sd^2*rho*c^2)
maybe T/S parameter test I did wasn't very accurate, maybe the mass weight I added not accurate, so Fsm result showed wrong too, so I use the formula with input just from Fs, Vas and Sd read from SW result and the different is ~30%.
If T/S test very accurate, it must give the same result, right?
I wonder, can it be as a sanity check, if Mms calc. using added mass formula not the same as Mms using Cms formula, then the mass weight added is wrong?
Chris.
the different is from Mms formula you use:
Mms=added mass weight/[(fs/Fsm)^2-1]
I use:
Mms=1/[(2*pi*Fs)^2*Cms]
Cms came from:
Cms=Vas/(Sd^2*rho*c^2)
maybe T/S parameter test I did wasn't very accurate, maybe the mass weight I added not accurate, so Fsm result showed wrong too, so I use the formula with input just from Fs, Vas and Sd read from SW result and the different is ~30%.
If T/S test very accurate, it must give the same result, right?
I wonder, can it be as a sanity check, if Mms calc. using added mass formula not the same as Mms using Cms formula, then the mass weight added is wrong?
Chris.
Hi Chris,
of course if you input not exact values, you get not exact results; the added mass value has to be very precise, try a jewelery or pharmacy, they have good balance; was Fsm lower of at least 25% in comparison to the Fs?
The purpose of my speradsheet, is to verify that the values measured through SW are similar to the one obtained using the known formulas, always using the T/S fundamental parameters given by SW; moreover it supplies other parameters that SW don't give. If something is well out of range, means something is wrong, and need an investigation. But first you have to be sure on the input you gave.
regards
Claudio
of course if you input not exact values, you get not exact results; the added mass value has to be very precise, try a jewelery or pharmacy, they have good balance; was Fsm lower of at least 25% in comparison to the Fs?
The purpose of my speradsheet, is to verify that the values measured through SW are similar to the one obtained using the known formulas, always using the T/S fundamental parameters given by SW; moreover it supplies other parameters that SW don't give. If something is well out of range, means something is wrong, and need an investigation. But first you have to be sure on the input you gave.
regards
Claudio
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.