The ACH-01 can have some unexpected issues if not mounted in a way so as not to generate strain in the sensor unrelated to Z-axis motion. Not sure if you had seen this thread showing an odd 13Hz mode from mounting cantilevered off the VC edge. http://www.diyaudio.com/forums/subwoofers/239941-analog-servo-sub.html#post3576660
There was also a nicely documented Thesis paper identifying some other mounting issues and solutions.
http://repository.tudelft.nl/assets/uuid:dbdc165b-31c9-477a-afc6-4093045c8479/MFB_Thesis.pdf
You may recognize the name of the professor as the consultant for the MFB woofer design used in the Grimm Audio MFB subwoofer which evidently has ~30dB of loop gain.
News | GrimmAudio
Who is who | GrimmAudio
The best solution I found for mounting the ACH-01 to avoid LF anomalies and push the resonance from the mounting well above 1khz was to mount it to an extremely stiff light composite platform glued directly to the VC. Attached is a pic of the Paradigm mounting I got the idea from. Using just a bit of lag compensation to ensure stability on the low end, I could safely get 20dB loop gain for 10x reduction in 2nd and 3rd harmonic for 60Hz and below.
Attachments
Last edited:
Thanks, for even more good info from you good folks..
Well, it looks like Rythmik received their woofers sooner than expected, as I just got an e-mail from UPS about their delivery.
That means unless I go through the hassle, and the (my) expense sending them back, I'm going to be building a pair of 15" subs using their matched amps.
Couple of things: Before I complete them I'll be sealing a speaker wire where the plate amp should be on one of them, and comparing kit vs parametric+power-amp, matched EQ and level to see which one has the best sound quality. I'll both measure it (using rew, dayton audio suite, and a measurement mic), and record it using one of my studio's large diaphragm condenser mics to make a sound-clip everyone can hear for themselves, and let the chips fall where they may...
I also found a pair of donor 250W 10's for use in accelerator-based MF. I'm thinking of putting them in PPSL as described upthread, and see how they do.
Of course all of this will be coming after I build one set of sealed, and one set of ported for the Rythmik kits.
BTW, the sealed cabs are gonna be easy, take a look at the pic:
They are all 3/4" void-free birch, are perfect 3 cubic foot for the kits , and were free... can't beat that with a club..
Well, it looks like Rythmik received their woofers sooner than expected, as I just got an e-mail from UPS about their delivery.
That means unless I go through the hassle, and the (my) expense sending them back, I'm going to be building a pair of 15" subs using their matched amps.
Couple of things: Before I complete them I'll be sealing a speaker wire where the plate amp should be on one of them, and comparing kit vs parametric+power-amp, matched EQ and level to see which one has the best sound quality. I'll both measure it (using rew, dayton audio suite, and a measurement mic), and record it using one of my studio's large diaphragm condenser mics to make a sound-clip everyone can hear for themselves, and let the chips fall where they may...
I also found a pair of donor 250W 10's for use in accelerator-based MF. I'm thinking of putting them in PPSL as described upthread, and see how they do.
Of course all of this will be coming after I build one set of sealed, and one set of ported for the Rythmik kits.
BTW, the sealed cabs are gonna be easy, take a look at the pic:
They are all 3/4" void-free birch, are perfect 3 cubic foot for the kits , and were free... can't beat that with a club..
Attachments
This is a really interesting project.
Btw this might be of interest:
ServoBass - High definition bass reproduction
Btw this might be of interest:
ServoBass - High definition bass reproduction
Sounds like a good plan
What type of EQ do you have?
Can’t remember where somebody stated that Rythmik never showed any distortion reduction comparisons, but I found some. They are at the bottom of linked web page. Note that they are for 10Hz where suspension distortion dominates and with ¼” P-P excursion where VC signal is as linear as it will get. They show 2x to 3x reduction, which is in line with my experience.
Rythmik Audio • servo subwoofers and memory effects
It would be very interesting to see similar plots for, say, 40Hz – 70Hz. Often with VC MFB you wind up actually increasing the distortion slightly in this range rather than reducing it, because the distortion is dominated by magnetic related non-linearities which affect not just the driven coil, but the sensor coil as well. The Rythmik purpose-built VC sensor(much longer than the driven VC) would certainly help alleviate this problem. My velocity feedback experiments were all with standard DVC woofers.
That's not a bad haul either..Hi wreckingball,
Post #22: "...3/4" void-free birch, are perfect 3 cubic foot for the kits , and were free..."
Well, that beats my found next to road 24"Dia. sono tubes.
Regards,
Thanks tb, yeah, I'll post up pics of the project(s).
Being mostly a visual learner, I know I appreciate them while trying to follow along with others projects.
Like I said, the sealed ones will be easy, but the kit is actually a ported design.
Therefore that pair will take significantly more time.. d'oh!
You know, a lot of folks love this DIY stuff, and I do to an extent, but mostly for me it's just because I have limousine taste and a go-kart budget..
Sounds like a good plan
What type of EQ do you have?
Can’t remember where somebody stated that Rythmik never showed any distortion reduction comparisons, but I found some. They are at the bottom of linked web page. Note that they are for 10Hz where suspension distortion dominates and with ¼” P-P excursion where VC signal is as linear as it will get. They show 2x to 3x reduction, which is in line with my experience.
Rythmik Audio • servo subwoofers and memory effects
It would be very interesting to see similar plots for, say, 40Hz – 70Hz. Often with VC MFB you wind up actually increasing the distortion slightly in this range rather than reducing it, because the distortion is dominated by magnetic related non-linearities which affect not just the driven coil, but the sensor coil as well. The Rythmik purpose-built VC sensor(much longer than the driven VC) would certainly help alleviate this problem. My velocity feedback experiments were all with standard DVC woofers.
Thanks bolserst,
Yes I did see those plots as well, that's another reason why I was curious to do a little experimenting before everything is put together.
I have 3 EQs I can try out, one is a dual 31-band Yamaha (pro-audio), the second is a Alesis PEQ-450 that I picked up second hand on the cheap, and the other option is dsp eq inside my Emotiva Umc-1 (probably not a good option).
I saw you mention a mini DSP in the other thread, and given the fact that I could probably use that with any upcoming accelerometer-based project, I might pick one of those up anyway.
There is that little uncertainty though, that is I do not know what the built-in EQ of the Rythmiks are...
I s'pose a sweep with a measurement mic and program would reveal that though?Do you happen to know which mini DSP would be the best option if dual analog servos might be in my future?
Thanks
Oh, it wasn't a mini DSP. You talked about a Behringer EQ.
I think I'll be able to do a ton more with the mini DSP though, specifically the 2x4 HD. That seems to be the most widely used with folks over at AVS.
Just got some more great news, budget-wise.
My friend who can no longer woodwork, is gifting me several sheets of 5'x5' 18mm baltic birch!.. cha-ching! Thanks buddy!..
I think I'll be able to do a ton more with the mini DSP though, specifically the 2x4 HD. That seems to be the most widely used with folks over at AVS.
Just got some more great news, budget-wise.
My friend who can no longer woodwork, is gifting me several sheets of 5'x5' 18mm baltic birch!.. cha-ching! Thanks buddy!..
miniDSP 2x4 HD would be perfect. It can handle biquad-programming, so makes it much easier to match a response that you want. Also, the HD version of the 2x4 is powerful enough to handle FIR filters, so should be able to experiment with phase linearization if that interests you.
You would measure once with Rythmik MFB electronics in place, then again with just a plain amp. Then you would use REW EQ wizard or the available mini-DSP spreadsheet to design some filter settings to make the response with the plain amp match what you measured with the rythmik MFB setup.
I don't know if the latency is low enough to perform processing for an accelerometer-based motional feedback loop though.
Have you heard of anybody doing that ?
Last edited:
Unfortunately, I do not have a good answer for your question right now...
I know that a miniDSP (2x8) was used as a 'pre filter' in the "Analog Servo Sub" thread.
That's where I got the idea...
Much more research is needed it seems..
I should have an answer by week's end, that is, if my Labrador doesn't decide that 'his human' is shirking his duties and needs to devote more time...lol
Anyway, thanks for all your help..
Last edited:
Someone tell me I'm wrong...
I don't think that's going to happen. MDF is heavy, weak and poisonous. Even if there were no chemicals the dust is so fine it gets in EVERYTHING including your lungs. And if you get it wet it's garbage. I was once working with mdf, a drop of sweat fell from my face and landed on the panel. Before I could wipe it off it blistered up like chicken pox. Not all mdf is equal but the stuff at big box stores is usually junk.
Anyway, about the mini dsp, I've been eyeing up the non HD 2x4 (because it's cheap) but I was holding off because I was led to believe it couldn't do delay. But the 2x4 advanced plugin demo video I just watched seems to indicate that you can delay.
I can't find the plug in comparison summary page anywhere even though it's referred to several times and links to it just link back to the plug in page. There is a small summary graph on the 2x4 HD product page but it doesn't compare much.
So other than the FIR filter thing, is there any good reason to spend the extra $100 on the HD vs the 2x4 non HD? An extra $100 is double the price. The HD can put out 2V, which is higher than the non HD which is nice, but the non HD is usb powered (not wired to a plug like the HD) which is super cool.
Last edited:
A little Housekeeping:
Of all of the links to accelerator-based MF that have been posted so far, only one has product (kits) on the horizon, and that was posted in #8 of this thread (if anyone is keeping track)
Couple of links: Motional Feedback Audio. ..and a link to a thread here at DIY audio: http://www.diyaudio.com/forums/vendors-bazaar/242760-piratelogic-2013-01-3ch-mfb-module.html
Both of those are (as far as I can tell) 2 or 3-way configurations. So far, I've not seen sub-woofer only kits listed.
Side note: I'm starting to think that this type of MF would be extremely useful in the 100Hz-15k+ region ..as opposed to the sub-bass region, due to the ears' sensitivity being more prominent in the former areas. (IMHO)
On to the miniDSP: the latency figures I have seen on this vary wildly (putting it mildly). Anywhere from the nanosecond range in a thread over at AVS, to a FAQ quote from the manufacturer stating 1.5-2 ms.
Even if the 1.5-2ms is accurate, that amount of delay is comparable to 64 samples in 'studio speak' . And I can tell you from my musician's experience, and having to play along to that latency, that amount of delay is a non issue.
*edit* Not to open up a whole 'nother can of worms, but the phase relationship between the delayed, and the non-delayed output would have to be figured out by someone a whole lot smarter that me..lol..
Now onto the max output voltage of the MDSP vs the HD version. One would have too look up the input voltage specs of their particular amp.
The nonHD version, as correctly stated by JAG, is .9v, whereas the HD version puts out 2v.
In my particular case, I am going to need the HD version for my amp to reach reference level.
Steve
Of all of the links to accelerator-based MF that have been posted so far, only one has product (kits) on the horizon, and that was posted in #8 of this thread (if anyone is keeping track)
Couple of links: Motional Feedback Audio. ..and a link to a thread here at DIY audio: http://www.diyaudio.com/forums/vendors-bazaar/242760-piratelogic-2013-01-3ch-mfb-module.html
Both of those are (as far as I can tell) 2 or 3-way configurations. So far, I've not seen sub-woofer only kits listed.
Side note: I'm starting to think that this type of MF would be extremely useful in the 100Hz-15k+ region ..as opposed to the sub-bass region, due to the ears' sensitivity being more prominent in the former areas. (IMHO)
On to the miniDSP: the latency figures I have seen on this vary wildly (putting it mildly). Anywhere from the nanosecond range in a thread over at AVS, to a FAQ quote from the manufacturer stating 1.5-2 ms.
Even if the 1.5-2ms is accurate, that amount of delay is comparable to 64 samples in 'studio speak' . And I can tell you from my musician's experience, and having to play along to that latency, that amount of delay is a non issue.
*edit* Not to open up a whole 'nother can of worms, but the phase relationship between the delayed, and the non-delayed output would have to be figured out by someone a whole lot smarter that me..lol..
Now onto the max output voltage of the MDSP vs the HD version. One would have too look up the input voltage specs of their particular amp.
The nonHD version, as correctly stated by JAG, is .9v, whereas the HD version puts out 2v.
In my particular case, I am going to need the HD version for my amp to reach reference level.
Steve
Last edited:
2 ms (if it really is that quick) is still 76 degrees at 100 Hz, if I have that right. Plus you need to add in all kinds of other delays everywhere in the loop. Two feet.
Ben
Meanwhile, back in the real world.... Yes, the last frontier after the last frontier may be MF applied to cones over 200 Hz. Of course, lousy enough we have to use crude Rice-Kellogg drivers for subs, the future will surely see the demise of cones for higher frequencies.
But my point is this: to make credible bass despite using drivers playing in their resonant range (a peculiar idea), MF is called for. At least for the present time, MF's benefits to higher frequency drivers is less substantial and the challenges of many acoustic sorts to applying MF far more forbidding.
wreckingball - many of us are overjoyed at the prospect of hearing how your trial goes. Thanks a lot.
Ben
Last edited:
From my experience, 1 - 2ms sounds about right for the standard miniDSP 2x4...it depends on exactly what plugins/filters you are using. I'm guessing a good portion of that time is for the ADC at the input and DAC out the output. bentoronto is right that you will need something faster...I'd suggest at least 100uS if you plan to have more than about 10dB - 15dB of loop gain. I have not had the opportunity to test the HD version of the miniDSP 2x4 yet, so can't comment on it's latency.
This seems exceptionally well conceived. Big feedback factor yet, I guess, doesn't blow up. Thanks for link.
Interesting that he shows bandpassed tone bursts for comparison rather than the unfiltered ones you usually see. Not a lot different but a good move. The bandpassed tone bursts are closer to what it is fair for a real woofer to produce.
Ah, but there is an interesting falsehood in the data, one you see a lot. He/she shows very big distortion reduction (which goes along with big feedback). But what we are looking at is testimony of the accelerometer (ACH01-03). Not legitimate! That's the sensor! We need to see testimony of a mic.
Of course the results of the accelerometer will be looking good with feedback applied. Duh. That would be true even if the accelerometer were terribly distorted itself (and I am not convinced about ACH01's quite yet).
It's practical to do R&D while eyeballing the accelerometer. Bit performance should be based on a mic.
Ben
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.