Sorry to bowser you with that, as it is not subwoofer related, but would not MF reduce the rise in distortion in the lower end of a midbass, and make it usable lower in frequency than it would have been possible otherwise?
Case in point here is a TAD TM1201 driver with a very rigid and damped cone, but quite limited excursion-wise, and which also have a very low Qms and high Rms which make it a bit "lifeless" at low levels...
Would not MF help in this specific case?
I think with an Xmax of only 2.5mm you have a limited choice as to how low in frequency you can actually drive it. Feedback won't help if you drive it beyond those limits.
regards
david
regards
david
Temperature fluctuation
I have done some measurements of the temperature fluctuation on my system.
Test setup:
Peerless 830845 in a closed enclosure of 135 liters. Accelerometer connected as in post 144. It is glued with a generic two component epoxy.
Measurements done directly on the accel output before the capacitor feeding the opamp with a 12-bit oscilloscope @ DC.
The accelerometer is connected as in the datasheet powered from +12V and -12V. The only difference is that I am using 1Mohm instead of 2MOhm on the positive input of my TL074 opamp.
The temperature rise in the voice coil does in fact increase the DC output from the accelerometer quite a bit. 85W for 20,1 seconds increased the DC with a maximum of 93mV. However, the time constant is so big that with a suitable low pass filter it will not influence the feedback loop much. What do you think?
I can run other tests if you like.
I have done some measurements of the temperature fluctuation on my system.
Test setup:
Peerless 830845 in a closed enclosure of 135 liters. Accelerometer connected as in post 144. It is glued with a generic two component epoxy.
Measurements done directly on the accel output before the capacitor feeding the opamp with a 12-bit oscilloscope @ DC.
The accelerometer is connected as in the datasheet powered from +12V and -12V. The only difference is that I am using 1Mohm instead of 2MOhm on the positive input of my TL074 opamp.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The temperature rise in the voice coil does in fact increase the DC output from the accelerometer quite a bit. 85W for 20,1 seconds increased the DC with a maximum of 93mV. However, the time constant is so big that with a suitable low pass filter it will not influence the feedback loop much. What do you think?
I can run other tests if you like.
Last edited:
Thanks for those tests. It vindicates what I suspected about these devices 😉
It is obvious that the accel is sensitive to temperature like the spec sheet says.
It would be interesting to see how the signal varies over time with constant drive current. To do this you would need constant current sinusoidal drive so that the current remains constant in the voice coil independent of voice coil resistance. You could then see how the accelerometer gain varies with temperature. It is important that the accelerometer gain remain constant otherwise the feedback will change and so will the putput as the temperature changes with drive level. Not good 🙁
I have now come to the conclusion that the accel needs to be temperature isolated in its own cavity that is not affected by the voice coil or outside air currents. I have some ideas on how to do this and will modify my test speaker accordingly. I will post some pictures next week 😉
regards
david
It is obvious that the accel is sensitive to temperature like the spec sheet says.
It would be interesting to see how the signal varies over time with constant drive current. To do this you would need constant current sinusoidal drive so that the current remains constant in the voice coil independent of voice coil resistance. You could then see how the accelerometer gain varies with temperature. It is important that the accelerometer gain remain constant otherwise the feedback will change and so will the putput as the temperature changes with drive level. Not good 🙁
I have now come to the conclusion that the accel needs to be temperature isolated in its own cavity that is not affected by the voice coil or outside air currents. I have some ideas on how to do this and will modify my test speaker accordingly. I will post some pictures next week 😉
regards
david
Last edited:
Hmm. I see. The spec sheet says "Transient Temp Sensitivity" is 0,35g/degree(C).
My sensor is calibrated to 9,6mV/g at 25 degrees.
If the temperatur of the voice coil (and my sensor) is 75 degrees, the sensitivity will increase to:
(75deg. - 25deg.) * 0,35g/deg. + 9,6mV/g = 27,1mV/g
Have I understood this correct?
If this is the case, mounting the sensor on the voice coil as i have done is really really not good
As a test, I will add another sensor on the dustcap and compare the output from the two. I can also increase the temperature manually on the second sensor and see what happens.
EDIT:
No, if the temperature rises from 25 to 75 degrees, the g-force will increase with 50*0,35=17,5g. So my measurement earlier at 74,5g would in fact only be 74,5g-17,5g = 57g if the sensor temperature was 75 degrees. Better?
My sensor is calibrated to 9,6mV/g at 25 degrees.
If the temperatur of the voice coil (and my sensor) is 75 degrees, the sensitivity will increase to:
(75deg. - 25deg.) * 0,35g/deg. + 9,6mV/g = 27,1mV/g
Have I understood this correct?
If this is the case, mounting the sensor on the voice coil as i have done is really really not good
As a test, I will add another sensor on the dustcap and compare the output from the two. I can also increase the temperature manually on the second sensor and see what happens.
EDIT:
No, if the temperature rises from 25 to 75 degrees, the g-force will increase with 50*0,35=17,5g. So my measurement earlier at 74,5g would in fact only be 74,5g-17,5g = 57g if the sensor temperature was 75 degrees. Better?
Last edited:
My impression is that temperature effects on VCs (and maybe accelerometers) are a real non-issue in the real world of music reproduction....but jolly good fun to argue about on DIY forums. Esp. with accelerometer sitting on top of the dust cap, where it can be glued.
But might befuddle testing with long-playing test tones - which isn't much in vogue these days anyway.
Ben
Ben
But might befuddle testing with long-playing test tones - which isn't much in vogue these days anyway.
Ben
Ben
I don't plan to drive it beyond xmax, but even well within xmax limitation the distortion is already rising pretty fast with excursion...
Well, I understand midbass driver is not your target here, but I would like to try this anyway (and I think the TM1201 would be the best candidate for this type of modification, with its "almost perfect" cone and "far from perfect" motor and suspension system), so if you are still opened to suggestion, please do not restrict the upper frequency limit of the MF, and maybe also allow the use of smaller/lighter sensors for those specific applications?...
Hmm. I see. The spec sheet says "Transient Temp Sensitivity" is 0,35g/degree(C).
My sensor is calibrated to 9,6mV/g at 25 degrees.
If the temperatur of the voice coil (and my sensor) is 75 degrees, the sensitivity will increase to:
(75deg. - 25deg.) * 0,35g/deg. + 9,6mV/g = 27,1mV/g
Have I understood this correct?
If this is the case, mounting the sensor on the voice coil as i have done is really really not good
As a test, I will add another sensor on the dustcap and compare the output from the two. I can also increase the temperature manually on the second sensor and see what happens.
EDIT:
No, if the temperature rises from 25 to 75 degrees, the g-force will increase with 50*0,35=17,5g. So my measurement earlier at 74,5g would in fact only be 74,5g-17,5g = 57g if the sensor temperature was 75 degrees. Better?
I figured that for a dT = 50 °C
dV = 50°C * 0.35g/°C * 9.6mV/g = 168 mV ??
Which ever way you want to calculate it, temperature is an issue as is bass strain which requires that the accelerometer be mounted on a rigid surface.
regards
david
Last edited:
Measurement of temperature influence ACH-01-03
I know you have taken precautions regarding the temperature issue David, but I was still interested in how the sensor reacted to different temperatures.
I connected a second sensor on the dustcap of my sub. On this sensor I fitted a temperature sensor and a 530Ohm resistor using thermal glue. I also insulated the thing with foam and tape.
With this, I was able to adjust temperature from 23 degrees celcius to 80 degrees (73-176F) by applying 0-30V to the resistor. The heat generated was a maximum 1,7W and since it is thermally isolated from the voice coil it will not influence my first sensor or the output of the woofer when voltage driven.
I ran a total of 40 tests at 23, 56 and 80 degrees celsius with different output levels. The results spread a little between equal tests so I had to re-run some of them to get a good average.
DC
At 23 degrees the DC is -200mV
At 56 degrees, somewhere in the -400mV to -300mV range. This can vary.
At 80 degrees the DC is -1,2V
Senor output vs temperature
At 56 degrees the sensor output AC voltage is reduced by a factor of 0,86.
At 80 degrees the sensor output AC voltage is reduced by a factor of 0,64.
In other words, when the sensor is at 80 degrees the g-force will be only at about 64% of its correct value. In my old fashioned analogue MF system this will give a reduction in loop gain of 3,9dB. At 56 degrees the reduction is only
1,3dB.
The heat problem is definitely there, but it seems that it is not the biggest of problems after all. 🙂
Here is a scope picture from my measurements.
Another interesting thing I noticed was that the DC offset changed after a good shake even when the temperature was stable
I know you have taken precautions regarding the temperature issue David, but I was still interested in how the sensor reacted to different temperatures.
I connected a second sensor on the dustcap of my sub. On this sensor I fitted a temperature sensor and a 530Ohm resistor using thermal glue. I also insulated the thing with foam and tape.
With this, I was able to adjust temperature from 23 degrees celcius to 80 degrees (73-176F) by applying 0-30V to the resistor. The heat generated was a maximum 1,7W and since it is thermally isolated from the voice coil it will not influence my first sensor or the output of the woofer when voltage driven.
I ran a total of 40 tests at 23, 56 and 80 degrees celsius with different output levels. The results spread a little between equal tests so I had to re-run some of them to get a good average.
DC
At 23 degrees the DC is -200mV
At 56 degrees, somewhere in the -400mV to -300mV range. This can vary.
At 80 degrees the DC is -1,2V
Senor output vs temperature
At 56 degrees the sensor output AC voltage is reduced by a factor of 0,86.
At 80 degrees the sensor output AC voltage is reduced by a factor of 0,64.
In other words, when the sensor is at 80 degrees the g-force will be only at about 64% of its correct value. In my old fashioned analogue MF system this will give a reduction in loop gain of 3,9dB. At 56 degrees the reduction is only
1,3dB.
The heat problem is definitely there, but it seems that it is not the biggest of problems after all. 🙂
Here is a scope picture from my measurements.
Another interesting thing I noticed was that the DC offset changed after a good shake even when the temperature was stable
Sure looks like it is about time to see before and after mic tests - freq sweep (to see flatness and control of resonance) and spectrum analysis (to see distortion products).
BTW, you could mount the driver in a window to create an infinite baffle. Even just mounting the driver on a board and backing it into lots of pillows would give useful results.
Seems real easy to do. And meaningful.
Ben
BTW, you could mount the driver in a window to create an infinite baffle. Even just mounting the driver on a board and backing it into lots of pillows would give useful results.
Seems real easy to do. And meaningful.
Ben
Hello Armand
I have found that air currents can generate low frequency noise in the system so it is important to isolate the device from both the temperature change of the the voice coil and the outside air. Even touching the device with my finger can send the cone wildly gyrating whilst touching the supporting bar does nothing so it is definitely responding to temperature change of my finger and not mechanical stimulus. Weird hah !!
Regards
David
I have found that air currents can generate low frequency noise in the system so it is important to isolate the device from both the temperature change of the the voice coil and the outside air. Even touching the device with my finger can send the cone wildly gyrating whilst touching the supporting bar does nothing so it is definitely responding to temperature change of my finger and not mechanical stimulus. Weird hah !!
Regards
David
Last edited:
Not "weird" but possibly bad. The gyrations mean poor feedback design because it indicates instability, one of the essential no-no's of audio.
Granted, pushing on a cone or fooling with the temperature of a sensor may be outside practical design parameters, but air currents are not (or shouldn't be).
The correct behavior of a feedback system is maintenance (or restoration) of the status quo ante. That's what feedback is all about. When you push on a MF cone, it should feel like a brick wall. No kidding. (OK, maybe a brick wall covered in cardboard.)
Any system that falls into instability - esp when expensive drivers are in the loop - can be on a short path to self-destruction.
Ben
Last edited:
It is not instability because loop gain is way below unity at these low frequencies !! It is a temperature effect of the accelerometer which is documented in the data sheet and can be simply cured by housing the accelerometer in a thermally isolated compartment or simply behind the dust cap and mounted on a thermally isolated and rigid structure.
And it is not necessary for a motional feedback system to respond down to the sub Hz region to be effective in the audio pass band. This "brick wall" response of a motional feedback system is a complete red herring and is not indicative of its real world audio reproduction capability. I know one vendor of a so called "servo system" that makes a big deal of that feature but it is really just of novelty value. It's ironical that the same vendor doesn't think that audible distortion is an issue but hand pumping a stiff cone is !!
regards
And it is not necessary for a motional feedback system to respond down to the sub Hz region to be effective in the audio pass band. This "brick wall" response of a motional feedback system is a complete red herring and is not indicative of its real world audio reproduction capability. I know one vendor of a so called "servo system" that makes a big deal of that feature but it is really just of novelty value. It's ironical that the same vendor doesn't think that audible distortion is an issue but hand pumping a stiff cone is !!
regards
Last edited:
No one could deny you have a strong conceptual grasp of the issues. But now it's time to show us a mic test too.
Ben
Ben
Wishlist:
One or even two analogue or digital temperature input(s) using one of the popular sensors? Always fun to be able to monitor the VC temperature.
One or even two analogue or digital temperature input(s) using one of the popular sensors? Always fun to be able to monitor the VC temperature.
The hybrid solution utilizing a coupling ring between isobaric clamshell mounted subs solves the temperature and air current issues- the accelerometer mounted on a tube coupling the two drivers, would be nice and far from the VCs, and one would not expect meaningful relative air movement within that cavity.
The other advantage is less demand on the feedback loop.
The other advantage is less demand on the feedback loop.
Perhaps I should speak more plainly, now that I've had a chance to think about the flow here.
1. YOU posted data showing some truly miraculous speaker improvements.
2. Turns out, you want to sell it to us and many people were impressed by your data and some were ready to queue-up to buy it.
3. Turns out, after some questioning, that your data are, umm, umm, open to re-interpretation (that's as polite as I can possibly be).
4. Despite repeated prompting to do some legitimate testing with a mic and with signal inputs more appropriate to the driver you are using, you'd rather tell me to get lost ("you are on the wrong thread") than to perform the simple mic tests to verify (or whatever is the opposite of verify) your product.
Are you now going to stomp off in a moral huff and leave this thread or are you going to test with a mic?
Ben
Last edited:
Gentlemen, keep it polite and not personal. Please read the rules.
Those are your words not mine. Read the start of the thread again. The device is not on the market yet, no price has been mentioned and it is still in prototype form and is a work in progress so I don't see why you should get so upset just because I fail to do a mic test on a speaker mounted in free air which would be kind of pointless if you get my drift 😉
As for mic testing I said I would do that when I mount the speaker in a suitable enclosure after I have fixed those pesky issues with the sensor for which someone else has already kindly taken the time to provide some feedback (pardon the pun) 😉
regards
david
- Status
- Not open for further replies.