That's fine, whatever you call it 😉Was thinking of using a TPA3255 for that part so I could put one coil on each channel (same signal).
Sorry, job got in the way of life🙄@Havoc so if im reading this right, instead of using the other coil actively, it passively cancels non-linearities by using it as a "sense" coil and plugging back into the circuit earlier on?
Yes, that is what I meant. Like Didden said, look to motional feedback.
It seems @jan.didden didn't mention this is quite a rabbit hole lol. Seems like there's actually a good deal of designs in use today. A few challenges seem to present itself that I need to account for:Yes, that is what I meant. Like Didden said, look to motional feedback.
1) overshoot. If the coil overcompensates it could create even more distortion on its own
2) filtering. Saw a few designs makes reference to different filter of all kind (high pass, low pass, etc.) so I need to figure out what frequencies need to be attenuated. I would assume that using it as a reciever would also mean dealing with ambient RF in some capacity so there's that too
3) sensitivity. A lot of designs used an accelerometer. I'm just wondering how well it'll track in comparison to one. I would assume we'll, as the two voice coils are on the same plane, but I don't know.
4) integration. A lot of designs have it as it's own circuit. Will I need a whole other preamp for this?
Tried this 25 years ago with an Alpine 6 x 9 Twin Driven subwoofer.
Transformer effect is minimal next to the regen voltage generated by the cone motion.
Feedback phase shift causes oscillation in the range of 70 to 200 hz depending on the cabinet.
Settled on current drive for woofer.
Transformer effect is minimal next to the regen voltage generated by the cone motion.
Feedback phase shift causes oscillation in the range of 70 to 200 hz depending on the cabinet.
Settled on current drive for woofer.
Thank you! I figured I couldn't be the only person here who thought of this! LmaoTried this 25 years ago with an Alpine 6 x 9 Twin Driven subwoofer.
Oh and for anyone who is also curious about motional feedback, I've been reading:
https://www.grimmaudio.com/wordpres...al-feedback-essentials-Rob-Munnig-Schmidt.pdf
https://audioxpress.com/article/ser...dio-from-motional-feedback-to-the-bend-sensor
https://www.audiosciencereview.com/forum/index.php?threads/motional-feedback-mfb.27956/
https://rmsacoustics.nl/papers/whitepaperMFBdesign.pdf
https://www.klippel.de/products/klippel-controlled-sound.html
https://www.rythmikaudio.com/servosurvey.html
Still need to think of a design though... if I only had the attention span to design one thing at a time 🤪
I designed a couple of versions of this in the 90s, the first time as a senior project at the University of Nebraska (I was awarded best senior project that semester) and then later at Jensen where I was an acoustical engineer for 2 1/2 years.
Linking the coils with a transformer will not work. You need to make one coil a Motional Feedback sensing coil which goes to the feedback loop of the amplifier, while the amplifier drives the other coil.
In order to make this work you have to compensate for several things and it still will have some issues.
The first thing you have to compensate for is: the powered coil directly induces voltage in the sensing coil through induction; the two naturally behave as a low quality 1:1 transformer. This coupling varies with frequency and you have to generate an equalized signal to cancel that out.
Once you've done that, you have a signal that is proportional to velocity (which is roughly proportional to the frequency response of the driver tilted downward at 6db/octave). I designed the feedback loop so that the higher my loop gain, the more the woofer behaved as a Constant Velocity device. This meant that the bulk of the negative feedback was centered around the resonant frequency of the driver, whereas ideally for distortion reduction you'd want it centered above that range because that's where the higher distortion harmonics are.
The constant velocity orientation meant that I had to pre-EQ the input signal to correct the output SPL to be flat again. (You do it somewhat differently when you use an accelerometer instead.)
Now I was still left with the problem that because of nonlinearities and hysteresis in the magnetic system, the sensing coil is an imperfect representation of the actual behavior of the cone, so when you apply feedback you are baking some forms of distortion into the design.
All that said, it was not terribly difficult to add about 15dB of negative feedback and noticeably reduce distortion.
The biggest challenge with any MFB system is you HAVE to build in a compressor/limiter to keep the thing from destroying itself with strong input signals. This is tricky and comprises the bulk of the effort.
The product never saw the light of day as there was a layoff and I left to pursue other things. I do have a couple of Motional Feedback Subwoofers in my house today that I built with circuit boards and accelerometers from https://piratelogic.nl/.
Linking the coils with a transformer will not work. You need to make one coil a Motional Feedback sensing coil which goes to the feedback loop of the amplifier, while the amplifier drives the other coil.
In order to make this work you have to compensate for several things and it still will have some issues.
The first thing you have to compensate for is: the powered coil directly induces voltage in the sensing coil through induction; the two naturally behave as a low quality 1:1 transformer. This coupling varies with frequency and you have to generate an equalized signal to cancel that out.
Once you've done that, you have a signal that is proportional to velocity (which is roughly proportional to the frequency response of the driver tilted downward at 6db/octave). I designed the feedback loop so that the higher my loop gain, the more the woofer behaved as a Constant Velocity device. This meant that the bulk of the negative feedback was centered around the resonant frequency of the driver, whereas ideally for distortion reduction you'd want it centered above that range because that's where the higher distortion harmonics are.
The constant velocity orientation meant that I had to pre-EQ the input signal to correct the output SPL to be flat again. (You do it somewhat differently when you use an accelerometer instead.)
Now I was still left with the problem that because of nonlinearities and hysteresis in the magnetic system, the sensing coil is an imperfect representation of the actual behavior of the cone, so when you apply feedback you are baking some forms of distortion into the design.
All that said, it was not terribly difficult to add about 15dB of negative feedback and noticeably reduce distortion.
The biggest challenge with any MFB system is you HAVE to build in a compressor/limiter to keep the thing from destroying itself with strong input signals. This is tricky and comprises the bulk of the effort.
The product never saw the light of day as there was a layoff and I left to pursue other things. I do have a couple of Motional Feedback Subwoofers in my house today that I built with circuit boards and accelerometers from https://piratelogic.nl/.
Thank you very much for the advice 🙂 actually on my way home from work I had an idea. What if I use something like an optical or laser diode, and used the difference of the signal between it and the coil? The only issue I can imagine is if the second coil gives garbage "data" itself which would throw the whole thing out of wack... well that and everything mentioned from above lol
If you keep on trying to use the coil as the sensing element you are up to several issues as mentioned before.
None of the commercial MFB systems use that. What you want to equalize is the cone movement, so that is what you need to sense.
There's a lot on sensors and eq preamps on the mentioned piratelogic.nl website, also types of sensors you can easily put together or buy from them.
The next thing you could review is what the control loop law is? From an accelerometer to controlling the cone position requires a double integrating loop.
This is no longer audio, but control systems theory!
Jan
None of the commercial MFB systems use that. What you want to equalize is the cone movement, so that is what you need to sense.
There's a lot on sensors and eq preamps on the mentioned piratelogic.nl website, also types of sensors you can easily put together or buy from them.
The next thing you could review is what the control loop law is? From an accelerometer to controlling the cone position requires a double integrating loop.
This is no longer audio, but control systems theory!
Jan
Using a low value resistor in the negative feed of the voice coil as a correcting signal will give you most benefits of MFB with lowest cost/ complications possible. If you power the woofer active and limit the frequency range to 200-400Hz, you can realize the MFB with 3 passive parts.
Anyway, while simple it still needs a closed cabinet, a matching driver, frequency limit and a very capable, stable amp.
I don't think using MFB into the midrange is a good idea for DIYS. Best used for an active subwoofer.
A 5" woofer like the mentioned Dayton with +-14mm linear excursion will not be the best option to use as a 2-way with an 2500 Hz x-over. Don't do it. MFB will make it even worse. You will run into complications you don't even think of right now.
If you want something with the lowest complication and no maintanance, the Piratelogic sensor system should be the best way for DIYS. Maybe ignore their preverence for small woofers 🙂 Physics still apply, even with MFB.
Anyway, while simple it still needs a closed cabinet, a matching driver, frequency limit and a very capable, stable amp.
I don't think using MFB into the midrange is a good idea for DIYS. Best used for an active subwoofer.
A 5" woofer like the mentioned Dayton with +-14mm linear excursion will not be the best option to use as a 2-way with an 2500 Hz x-over. Don't do it. MFB will make it even worse. You will run into complications you don't even think of right now.
If you want something with the lowest complication and no maintanance, the Piratelogic sensor system should be the best way for DIYS. Maybe ignore their preverence for small woofers 🙂 Physics still apply, even with MFB.
I just feel kind of weird leaving a whole coil to do nothing. The whole reason I started this thread was because I had an idea about a way to use. I must admit though, I am really enjoying all the stuff I'm learning from you guys. Part of the reason I like this hobby so much is there's so much I do know, but so much more I don't!If you keep on trying to use the coil as the sensing element you are up to several issues as mentioned before.
In my defense, that seems to happen a lot here...This is no longer audio, but control systems theory!
That seems simple enough. The original plan was to make the system a 3 way and use it as just a subwoofer anyway so this works.If you power the woofer active and limit the frequency range to 200-400Hz, you can realize the MFB with 3 passive parts.
This also works. I've come by the name enough through all of this that there definitely has to be validity to it.If you want something with the lowest complication and no maintanance, the Piratelogic sensor system should be the best way for DIYS.
Could you elaborate on that please?A 5" woofer like the mentioned Dayton with +-14mm linear excursion will not be the best option to use as a 2-way with an 2500 Hz x-over. Don't do it.
If you push a small diameter driver to make huge excursions (they get extreme in low bass, even at low volume!) while simultainiously making the cone reproduce something like 2500 Hz, you should be able to imagine the problem by your self.
Do you really want a speaker that virtually moves forward an back from you all the time? Ever heard of Doppler shift?
If you do not make the Dayton act as a woofer with high excursion, the whole concept is worthless. Just like only powering one voice coil. This is a small subwoofer, tweaked to simulate a large woofer on the expense of very high excursion and very high power consumption, preferable by two amps, not a mid range driver.
Do you really want a speaker that virtually moves forward an back from you all the time? Ever heard of Doppler shift?
If you do not make the Dayton act as a woofer with high excursion, the whole concept is worthless. Just like only powering one voice coil. This is a small subwoofer, tweaked to simulate a large woofer on the expense of very high excursion and very high power consumption, preferable by two amps, not a mid range driver.
Great point. Definitely reinforcing the TPA3255 idea. The original crossover had it at 300Hz for the LPF so i can always go back to that.If you do not make the Dayton act as a woofer with high excursion, the whole concept is worthless. Just like only powering one voice coil. This is a small subwoofer, tweaked to simulate a large woofer on the expense of very high excursion and very high power consumption, preferable by two amps, not a mid range driver.
Like most things audio?It seems @jan.didden didn't mention this is quite a rabbit hole lol.
It isn't wrong to use a long throw driver for this and then only use a small part of the excursion. You will stay in the linear range and both coils will be always in the same field.
As for motional feedback, you could look at the article in Wireless World september 1981. They describe a complete system with speakers that have a piezo in them. Very likely not on the market anymore but the text might give you pointers. The november 1986 Elektuur (dutch version) also has an article where they put such a piezo in a speaker.
Last edited:
Well, you have a choice: do it well, or do it crappy. You can only make the choice if you know all implications.
It should be clear by now that using the second coil for your purpose is a crappy solution.
So you stay with that and accept a crappy, possibly even disfunctional solution, or go for 1st prize.
Or find another project more to your liking.
What you see here often is that people fight the problem and not get anywhere.
It happened to me often when I started out on this hobby.
You have an idea which seems fantastic and you start to look into it.
Then you find that the basic idea may be great but there's no way to get it implemented with good results.
Sometimes you then start 'fighting the problem'- you keep on trying to find a way to get it implemented as you originally intended, not willing to give up on a great idea. And you get nowhere, spin your wheels and waste your time and money.
I would suggest discarding the problem and find a solution with more chance of success. But that's just me.
Jan
It should be clear by now that using the second coil for your purpose is a crappy solution.
So you stay with that and accept a crappy, possibly even disfunctional solution, or go for 1st prize.
Or find another project more to your liking.
What you see here often is that people fight the problem and not get anywhere.
It happened to me often when I started out on this hobby.
You have an idea which seems fantastic and you start to look into it.
Then you find that the basic idea may be great but there's no way to get it implemented with good results.
Sometimes you then start 'fighting the problem'- you keep on trying to find a way to get it implemented as you originally intended, not willing to give up on a great idea. And you get nowhere, spin your wheels and waste your time and money.
I would suggest discarding the problem and find a solution with more chance of success. But that's just me.
Jan
@jan.didden I believe you are referring to the sunk cost fallacy. You make a good point. I need to hop off the second coil idea. Still feel weird having it do nothing, but I'll cross that bridge eventually. Could possibly bi-amp and split the power cost for it per coil. As for MFB, I still think its something worth while to pursue. The idea was to have a 3 way speaker about the size of a bookshelf speaker. My little project will probably need a separate enclosure for the electronics, but I'm ok with that. That being said, in such a small enclosure, like @Turbowatch2 said:
Having a woofer in such a small place, it would definitely behoove me to take into consideration anything that could offset any extra vibration or resonance that could occur within a small enclosure. Between grimm audio white papers (https://www.grimmaudio.com/publications/loudspeaker-technique-white-papers/) and piratelogic I think I have a good place to start in terms of design.Physics still apply, even with MFB.
Basic premise is looking for perpetual motion. Motional feedback with vc sensor needs gain.
Where tried in past (esp a trial balloon from Sony), you must have far higher voltage to come from the sensor which requires that the feedback coil have lots of turns.
Absolutely astonishing that a century of feedback evolution has still left speakers outside the feedback loop. How can you ever think of yourself as striving for fidelity without encompassing the speaker in the loop?
Where tried in past (esp a trial balloon from Sony), you must have far higher voltage to come from the sensor which requires that the feedback coil have lots of turns.
Absolutely astonishing that a century of feedback evolution has still left speakers outside the feedback loop. How can you ever think of yourself as striving for fidelity without encompassing the speaker in the loop?
Makes sense to me (not the absence of, I mean including it). In a feedback loop, isn't it most effective when done from output to input or close to?How can you ever think of yourself as striving for fidelity without encompassing the speaker in the loop?
Actually since this little conversation, I've been thinking about the concept itself. If I'm putting a subwoofer in a small enclosure... what other forces of a more mechanical nature that create some sort of distortion would be at work that would/could be canceled out? Was thinking maybe a pressure sensor in the back as well. Was also reading about how some people have tried a laser diode. Not so sure that would be a great idea when the entire enclosure is vibrating, but its a neat idea to explore.
I also decided to order an extra subwoofer for experimentation. Would love any suggestions for things to try. Am probably going to make a little test rig and maybe pull the trigger on a cheap o-scope (pun intended).
For a sensing coil, many drivers have an aluminium voice coil former with ends that do not touch. If you could attach a wire to each end you would have a single turn coil for feedback purposes. That opens up a wide range of drivers for experimentation.
- Home
- Loudspeakers
- Subwoofers
- Weird dual coil idea