Hi again to the community. Long time lurker here.
I am a bit tired of my current speakers (ScanSpeak Revelator 18W/8531G-00 + D2905/9700) and want to build something DIY. A tiny three way. Maybe with an added floorstander bass base if the SPL is too low.
And now the twist: The woofer will be MFB controlled. I would like to play a bit more with this technology.
I feel there is a need to have an active thread here about MFB. Moreover it will be better to have 20 opinions on the matter. If I work on the speaker alone it will be only mine. Comments, ideas, remarks - all welcome!
I am a bit tired of my current speakers (ScanSpeak Revelator 18W/8531G-00 + D2905/9700) and want to build something DIY. A tiny three way. Maybe with an added floorstander bass base if the SPL is too low.
And now the twist: The woofer will be MFB controlled. I would like to play a bit more with this technology.
I feel there is a need to have an active thread here about MFB. Moreover it will be better to have 20 opinions on the matter. If I work on the speaker alone it will be only mine. Comments, ideas, remarks - all welcome!
Last edited:
I am aiming at baffle width of 180-190mm. Overall height ~850-950mm to have the tweeter at ear level.
I have Kartesian Cox120 available, and will most likely use it. At least at first. Maybe it will fit to my liking.
I will start (continue) with the woofer and the MFB first though. There are too many uncertainties there.
I have Kartesian Cox120 available, and will most likely use it. At least at first. Maybe it will fit to my liking.
I will start (continue) with the woofer and the MFB first though. There are too many uncertainties there.
Given the 4" coaxial what size and number of woofers are you considering for the tiny speaker?
Have you decided on the MFB approach?
Have you decided on the MFB approach?
I am considering a bookshelf with one 5-5,5 inch woofer, but the MFB SPL penalty may force me to add a stand with three more of the same type.
Two of those SB/SB13PFCR25-4 are sitting here in their boxes. Probably will not be used in the final speaker, but they will help to refine the MFB.
MFB sensor (accelerometer charge amplifier) was installed to a pair of TangBand W5-1138SMF, but one of them fell down from 2+ meter high shelf. 😱
The buck magnet broke to pieces and the steel frame twisted a bit. Because of that I put the speakers aside. I may return to them later.
This one is also interesting : Dayton E150HE-44
The accelerometer element is this one : https://www.murata.com/en-eu/products/productdetail?partno=PKGS-00LDP1-R
Two of those SB/SB13PFCR25-4 are sitting here in their boxes. Probably will not be used in the final speaker, but they will help to refine the MFB.
MFB sensor (accelerometer charge amplifier) was installed to a pair of TangBand W5-1138SMF, but one of them fell down from 2+ meter high shelf. 😱
The buck magnet broke to pieces and the steel frame twisted a bit. Because of that I put the speakers aside. I may return to them later.
This one is also interesting : Dayton E150HE-44
The accelerometer element is this one : https://www.murata.com/en-eu/products/productdetail?partno=PKGS-00LDP1-R
Surprised there isn't more chat. The first speakers I wanted to buy as a teenager were a pair of white servosounds from a local hi-fi shop. One of them had failed when demoing and they were going cheap. Not uncommon I believe. With the overconfidence of youth I expected I could fix them but didn't quite have enough pocket money saved up. Father, quite rightly doubting my abilities and recognising he would be pressured by mum to fix them and make good whatever I had bodged in my efforts, wisely declined to give me a loan.
What is the MFB SPL penalty?
What is the MFB SPL penalty?
Oh! Servo Sound!
During those times Servo Sound were unobtainable here and are still a mystery to me. A kind of feedback EQ/speaker modeling i guess.
Hope the thread will gain momentum.
By "penalty" i mean everything above the red horizontal line. In case this particular speaker is MFB controlled flat from 35Hz - the max SPL will be 82-83db.
I don't listen loud at home, but that seems too low.
During those times Servo Sound were unobtainable here and are still a mystery to me. A kind of feedback EQ/speaker modeling i guess.
Hope the thread will gain momentum.
By "penalty" i mean everything above the red horizontal line. In case this particular speaker is MFB controlled flat from 35Hz - the max SPL will be 82-83db.
I don't listen loud at home, but that seems too low.
It is an accelerometer sensor mounted on the loudspeaker.
https://en.wikipedia.org/wiki/Motional_Feedback
https://en.wikipedia.org/wiki/Motional_Feedback
What additional parts are required and what might need to be done differently?
A sensor to measure the motion of the cone. What options are available to DIYers?
A driver with suitable parameters or modifiable to get them. What are we looking for? Long stroke, soft suspension?, suitable dust cover?,...
A power amplifier design that includes the sensor output to better control the cone motion. What are good options? I seem to recall that conventional constant voltage is not optimal?
What else?
A sensor to measure the motion of the cone. What options are available to DIYers?
A driver with suitable parameters or modifiable to get them. What are we looking for? Long stroke, soft suspension?, suitable dust cover?,...
A power amplifier design that includes the sensor output to better control the cone motion. What are good options? I seem to recall that conventional constant voltage is not optimal?
What else?
Long stroke, but the surround should not be too soft. Otherwise there is the risk the surround will be sucked in at higher output. This is why Philips used a special, lower Cms, stiff surround variety of the 806x and 706x series woofers in their MFB systems.
The sensor can be the homemade with various piezo types or ready bought. The ACH-01 types are classics, Piratelogic supplies diy kits.
According to Bill Waslo there is a new kid on the block, the Senther part # 540A. But there is more: Audio Chiemsee also seems to produce sensors.
But it is not trivial to get a properly working sensor, that accurately tracks the woofer at ultra LF.
The sensor can be the homemade with various piezo types or ready bought. The ACH-01 types are classics, Piratelogic supplies diy kits.
According to Bill Waslo there is a new kid on the block, the Senther part # 540A. But there is more: Audio Chiemsee also seems to produce sensors.
But it is not trivial to get a properly working sensor, that accurately tracks the woofer at ultra LF.
The air trapped in a tiny speaker cabinet will likely provide all the stiffness needed and possibly more. An additional stiff suspension is likely to push the resonant frequency well into the frequency passband of the woofer. If we are not using a constant voltage amplifier is this going to cause difficulties? I cannot see the full picture so it might not. Also what was considered stiff in the days when acoustic suspension was common may not be so today?
If the woofer being modified has an adequately stiff suspension how much of a problem is the additional mass of the sensor plus it possibly not being symmetrical and driving the rocking modes more strongly?
Servo controlled subs would want to control ultra low frequencies but would this be the case for the main speakers being considered here? Wouldn't a low frequency extension to 40-50 Hz be more appropriate assuming they are mains for a room rather than a desktop?
If the woofer being modified has an adequately stiff suspension how much of a problem is the additional mass of the sensor plus it possibly not being symmetrical and driving the rocking modes more strongly?
Servo controlled subs would want to control ultra low frequencies but would this be the case for the main speakers being considered here? Wouldn't a low frequency extension to 40-50 Hz be more appropriate assuming they are mains for a room rather than a desktop?
I'm interested to see how this turns out.
Am I right in saying that, in the picture above, you've mounted the sensor to the voice coil former using the 3D printed part. And then you've mounted the preamp PCB to the pole piece?
Am I right in saying that, in the picture above, you've mounted the sensor to the voice coil former using the 3D printed part. And then you've mounted the preamp PCB to the pole piece?
Just so I get it right. The max output at 35 hz is 82 db. Does that take into account room gain?
I read here and there for lasers, additional coils, pressure sensors. Many options are available, but every approach requires a specific knowledge. I will try with the Philips classic - accelerometer.
Long stroke is a must IMO. SPL will be low anyway. At least to not make it lower. Thank you for the info about the lower Cms. I have to understand what is the reasoning behind this decision.
The sensor is home made indeed. Several versions in fact. It is a very small PCB. I know about Murata sensors from Piratelogic.
Do you mean Audio Chemgau? - https://audiochiemgau.com/en/ac-par75-75-w-dmos-audio-amplifier-with-on-board-mfb-controller
About the accelerometers. Let me find some old (and not so old photos). A picture says a thousand words. 🙂
Correct. It is a 3d printed part. Two coax wires connect the sensor to a small connection PCB. No preamp is used. The receiver is on the main board. Some pictures are attached.
Attachments
Ah now I see how this works. I was wondering how you got the signal out. You basically create a new set of tinsel leads and terminals through the cone and on the basket.
@IxnayNothing is taken into account. This is just an assumption. It could be wrong.
@andy19191
I haven't decided how low the MFB drivers will be extended. There is nothing working. The tests will show. Everyone is telling me, that tiny 5" driver and MFB cannot go together. Voltage amplifier will be used. At least for start. They are available everywhere. If there is something working I may try with current amp.
I am not good with speakers and was asking the same question myself about the added mass. At that time I used an online calculator.
The sensor is not 63 grams (7.5 grams actually). An extreme case was checked. Sadly the site is not available anymore. (mh-audio).
Against rocking ... the sensor is positioned at the top of the voice coli. At the cylinder axis. Don't know if it is enough.