
Home  Forums  Rules  Articles  The diyAudio Store  Gallery  Blogs  Register  Donations  FAQ  Calendar  Search  Today's Posts  Mark Forums Read  Search 
Software Tools SPICE, PCB CAD, speaker design and measurement software, calculators 

Please consider donating to help us continue to serve you.
Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving 

Thread Tools  Search this Thread 
17th March 2010, 09:15 AM  #1 
diyAudio Member
Join Date: Nov 2009
Location: Pune

Using LIMP
Hi,
I am currently trying to use LIMP for measurement of Thiele/Small parameters using added mass method. I got some queries regarding the method and the tool. 1) Reference resistor value. 2) Calibration method 3) correction files for calibration. LIMP has recomended resistor of 100E so I used it as a reference resistor. Also as other try, I used 1K resistor. But I got different huge difference in inductance (Le) values. What could be the reason for this? How must it be affecting while calculating real and imaginary part of it? For calibration of the system I used a 10E 1% resistor instead of speaker. Then to fine tune the setup altered the value of reference resistor in the setting menu of LIMP. Such that impedance graph shows me 10E exact. This method is used for any value of reference resistor. I had selected 10E value as this is close to 8E or 4E speakers generally used. Please know if the method is correct or any adverse effects on measurements. I am using my Desktop card for measurement. The problem is frequencies below 10Hz are not linier. This needs to be corrected, but to my knowledge LIMP do not support this calibration and correction. Pls guide me. The BL value Vas changes with different mass added. So what is right way or weight to be added? Some of the measurement made with ref resi of 100E are given below. Parameter 10g mass 30g mass Units Fs 32.04 32.04 Hz Re 3.5 3.5 ohms[dc] Le 2155.67 2155.67 uH L2 4218.47 4218.47 uH R2 3.51 3.51 ohms Qt 0.37 0.37  Qes 0.39 0.39  Qms 5.55 5.55  Mms 58.78 64.91 grams Rms 2.131052 2.35336 kg/s Cms 0.00042 0.00038 m/N Vas 58.18 52.68 liters Sd 314.16 314.16 cm^2 Bl 10.25345 10.775 Tm ETA 0.47 0.42 % Lp(2.83V/1m)92.39 91.96 dB 
17th March 2010, 11:21 AM  #2 
diyAudio Member
Join Date: Nov 2009
Location: Pune

Parameter 10g Dim
Fs 32.04 Hz Re 3.5 ohms[dc] Le 2155.67 uH L2 4218.47 uH R2 3.51 ohms Qt 0.37  Qes 0.39  Qms 5.55  Mms 58.78 grams Rms 2.131052 kg/s Cms 0.00042 m/N Vas 58.18 liters Sd 314.16 cm^2 Bl 10.253448 Tm ETA 0.47 % Lp(2.83V/1m) 92.39 dB Parameter 30g Dim Fs 32.04 Hz Re 3.5 ohms[dc] Le 2155.67 uH L2 4218.47 uH R2 3.51 ohms Qt 0.37  Qes 0.39  Qms 5.55  Mms 64.91 grams Rms 2.35336 kg/s Cms 0.00038 m/N Vas 52.68 liters Sd 314.16 cm^2 Bl 10.774996 Tm ETA 0.42 % Lp(2.83V/1m) 91.96 dB the table looked messed in above post so.. .added again 
17th March 2010, 12:57 PM  #3 
diyAudio Member
Join Date: Oct 2005

Using the soundcard output for LIMP is probably not ideal for precise measurements. Due to the limited power you must use a resistor of 50100 ohm, which is large compared to the measured object. Small errors will thus be magnified.
It is best to use a proper amplifier and a series resistor of say 27 ohm. Also the leads to the speaker should be thick enough not to add any resistance. In principle the frequency curve should not matter to LIMP since it compares the voltage applied to the voltage over the device_under_test. A calibration file is therefore not applicable. Precise balance between the two input channels is however important. Your calibration method will calibrate OK for 10 Ohm, but it may still give an error at frequencies where the impedance is considerably higher. Using the addedmass method, the mass must be large enough to measure a significant change in Fs. Measurements tend to vary a little depending on the sound level used. I do not know what level might be the industry norm. Svein. (not an expert by far) 
17th March 2010, 06:33 PM  #4 
diyAudio Member
Join Date: Nov 2009
Location: Pune

Thanks Sevin for your reply.I managed thick cable and could get better precision. I am using weights that will give frequency shift of 20 % or so.
its like LIMP says add weight that shifts fq by 20% to 50% now suppose I use x grams to get a deviation by 20% which gives me a set of result. but when a weight of y grams is used that will shift fq by 50% ... ideally the results should be same. but it is not so. so.. got stuck. what is the correct measurement? 
18th March 2010, 07:56 AM  #5  
diyAudio Member
Join Date: Oct 2006
Location: delhi

Quote:
hi Aucosticraft, i feed both input channel the same signal for calibration or channel difference. as you increase the mass vas and fs will drop.i used to measure with 1 volt and 10 ohm series resistor.using a 10watt amp between pc and driver.many have suggested on this forum to use 1 volt for t/s measurements.for vas measurements first i measure with weight equivalent to cone/vc/spider. (i dont know know the actual weight just a guess work ).then calculate the t/s parameters.this gives you the idea of moving mass.then i weigh the same amount of weight and restart from the first step.. regards ravs be a vegeterian 

18th March 2010, 08:14 AM  #6 
diyAudio Member
Join Date: Oct 2005

Both your measurements could be correct  although different.
It is probably a reflection of the fact that the mechanical compliance is not really a constant (as assumed in the model), but varies with excursion and frequency. Some explanations by those who know: Bjorn Ideland at Seas: Help with Seas w15cy001 T/S parameters The LienarX forum: Thiel Small parameter testing  LinearX Forums Svein. 
19th March 2010, 08:15 AM  #7  
diyAudio Member
Join Date: Nov 2009
Location: Pune

Quote:
__________________
There is always a first time.... 

19th March 2010, 08:18 AM  #8 
diyAudio Member
Join Date: Nov 2009
Location: Pune

Svein, I am going through the links... and some other ref. manual too. so.. would be silent for sometime.. thanks...
__________________
There is always a first time.... 
20th March 2010, 07:35 AM  #9  
diyAudio Member
Join Date: Oct 2006
Location: delhi

Quote:
hi Aucosticraft, its only guess work and by look of cone typically for 5' woofers weight of the moving mass 1014g for 10' woofers its 3550g for 12' its 6080g and so on first test gives you the rough estimation of real moving mass.now get that new mass shown in the t/s parameters window.do the whole process one more time with new mass. regards ravs be a vegeterian...... 

6th April 2010, 07:36 AM  #10 
diyAudio Member
Join Date: Nov 2009
Location: Pune

Hey Friends... I made small mistake .. that resulted in big error... I measured membrane diameter as 20cm where it was 18cm. but after this correction.. the readings were accurate. i used weight that would give frequency drift of approx 30%.
__________________
There is always a first time.... 
Thread Tools  Search this Thread 


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
LIMP bug?  crazyhub  Software Tools  2  16th September 2009 04:43 PM 
New To Site?  Need Help? 