C173-6-096E and 6640 project

[Face]

Why not consider the OmniMic, it's ALMOST idiot proof.

[JonMarsh]

I am willing to bet the Realtec is the weakest link here. What does everyone think of the DragonFly Black USB DAC? This would give me a 9010 Sabre DAC I can plug in to any computer, which I like a lot better than a higher end PC Express sound card for a computer I probably won't use with audio in the future.

I have experiencer with the RED, and can recommend it.

[JonMarsh]

I don't have a problem with my measurement! Or at least the main problem is output from my software, computer hardware, or amp.

At the same volume, a 10kHz sine wave outputs 0.863 volts from my amp. A 20kHz sine wave only measures 0.397 volts, I'd expect 0.4315 volts to be 6db lower, so my graph above is actually close to flat, at least closer than my output.

Now I can measure whether it is the amp or headphone jack with the issue, and replace whichever has the problem.

Did you detect this by the loopback test I recommended, or just by putting as DVM on the system? Loopback test is a fundamental check that should be done with a new test setup before any measurements- the system tests itself!

[JonMarsh]

I know Jon isn't the biggest fan, but I like Emotiva's Ego products better. I'd take Burr Brown over ESS Sabre. This would have either a PCM5142 or PCM5141 DAC.

Why?

The thing is, neither of these products is a duplex send and receive audio interface, which is what you need for testing. All the EGO or the Dragonfly's address is an output.

I think you'd a full duplex audio interface with balanced connectors much more useful...



It's USB, and 96/24 capable.

I mean, to me, it's weird that you're trying to design and build a system with $800 woofers and $500 tweeters, and use such a low end measurement setup.

Face's suggestion of the Omnimic is a good one; I'd go a step further and suggest the combo package PE offers of the Omnimic and DATS for driver measurement

http://www.parts-express.com/dayton-audio-omnimic-v2-and-dats-v2-speaker-measurement-bundle--390-800

If it's a budget thing, then just take your time to accumulate the cash and put the fundamentals together and start working when you have the tools you need.

I'm not usually this blunt, so my apologies- I'm trying to avoid channeling my inner ET. :W

[theSven]

I own a device like this that I use for recording audio for work I do on the side at work...

[bvbellomo]

It is so easy with high end audio to spend a lot without getting anything, it isn't even surprising to find a $5 sound card that outperforms a $200 one. With the Accuton, I know I got value, I am still not convinced the 6640 is an upgrade from the 9700, but if I eventually regret that purchase, I can probably sell the tweeters for close to what I paid.

I do agree money is best spent in a balanced way, but think most people would be better off spending more on drivers. I've read countless posts where someone uses a $30 driver they say is almost as good as a $60 driver, and I know the cabinet cost over $100, their woodworking equipment cost over $1000 and they have another $1000 in electronics. This is fine for a second set of speakers for the basement, for surrounds or for a friend, but my main system deserves better. I've turned down a lot of $50 / hour side jobs and spent a lot of hours building speakers.

There is also something to be said for learning the most basic way of doing something. When I was 8, I didn't understand why I should learn arithmetic. It wasn't because my parents couldn't afford to buy me a calculator.

[bvbellomo]

The Omnimic and DATS package looks like a good deal, but will this help with output problems? This looks like it only solves the input side of the problem.

I was considering buying the dats when Jon suggested the woofer tester, was trying to decide between them.

[augerpro]

Nailing down your method is free.

We still don't know if you have an individualized mic file, if it's loaded, if your loopback looks good, your measurement distance, how much clean signal you can get, what's holding the mic, what the speaker is sitting on, what the environment looks like...

[bvbellomo]

Using just the sound card, I can do loopback testing by plugging the output into the input. I haven't been able to get this to work, but it is of questionable relevance since my microphone isn't part of the loop.

I do have the microphone calibration file, and it appears loaded correctly, but doesn't change the values enough to account for the problem if it isn't.

[Evil Twin]

I can do loopback testing by plugging the output into the input. I haven't been able to get this to work, but it is of questionable relevance since my microphone isn't part of the loop.

Most definitely wrong! If you have not been able to get loopback to work, you cannot verify the fundamental functioning of your electronics chain BEFORE consideration of the microphone performance and it's calibration and calibration file.

One step at a time.... Obi Wan has NOT taught you well...

At the same volume, a 10kHz sine wave outputs 0.863 volts from my amp. A 20kHz sine wave only measures 0.397 volts, I'd expect 0.4315 volts to be 6db lower, so my graph above is actually close to flat, at least closer than my output.

This by itself clearly shows that you have a functional problem with your electronics chain... pray you find and remedy that issue, unless you wish to join Captain Needa with a similar apology...

[bvbellomo]

I am still trying to figure out what makes the Omnimic and software worth 4 times the cost UMM-6 and ARTA.

Many people on the Internet report Windows 10 issues above 10kHz. I am still researching, but am tempted to buy a better DAC or sound card, even if it doesn't fix this problem, it might be a worthwhile upgrade.

[bvbellomo]

I finally got my loopback to work:

Image not available

No sound card input with a few different versions of Realtek's driver, but the generic Microsoft driver picked it up. This is with 96kHz for both the input and output, MLS sequence length 262144, sampling at 96kHz for a 2.7 length signal. While it may not be as bad as my speaker, the 10k+ roll off is there, and the response is awfully ragged for not having any transducers.

I did try again with the speaker in line, no significant changes.

I think I need a better output device.

[augerpro]

Try 48 or 44.1 khz, 16 bit

[augerpro]

Also what gate did you use? You should be able to do 500ms for a loopback. All I see you list is signal length, which isn't referring to the gate, at least it wouldn't in Soundeasy. I think that signal length is just defined by the sequence length. Pretty meaningless to us at this point, just select the longest sequence length.

[5th element]

Even though you say that the above measurement was for 96kHz there looks to be a standard brick wall filter limit imposed around 20kHz. This implies a sample rate limitation somewhere in the measurement system.

Certainly a loopback should be a smooth and flat straight line with some possible minor deviations at the frequency extremes.

Noise could be an issue. Use averaging to help eliminate this. ARTA allows you to select how many times you wish to measure and average.

[augerpro]

yeah my experience is even if the hardware supports full duplex @ 96khz, the measurement software chokes, or windows chokes, or the software working with windows chokes. 44.1khz should be solid, and as you said, looks like one part of the process is already running there. If another isn't obviously you're going to have garbage results.

OP verify that ARTA, windows sound panel, and Realtek are all set to the same sampling, start with 44.1khz.

[bvbellomo]

Noise is why it is ragged, but that is still a problem, if not for measuring it is for listening. 44kHz or 44.1kHz don't seem to make any difference. Has anyone been able to get 20kHz+ sound out of a Windows 10 machine without a USB device? There is a brick wall filter in place, but I think it is part of the OS. The Realtek is probably the best possible engineering for a chip that costs a few pennies, but sound is not a very high priority for companies like Microsoft or ASUS.

Image not available

Here is the same MLS at 44.1kHz averaging 10 records.

[bvbellomo]

No one other than Jon addressed my questions about what to buy next. I like Dayton Dats, but if Jon likes the woofer tester he recommended first better, it is probably worth the additional money. I still maintain that I can build just as good a speaker with a multimeter, and will learn more from doing it.

To me, the OmniMic looks like a UMM-6 with a software package. Buying a complex software package from a small company, a free product from a dedicated company is almost always better than a $300 product from a company where software is not a primary focus. This is why I've stuck with ARTA and the UMM-6. The Omni Mic is definitely 'what is our spending for the month' money and not 'pick up an extra dozen eggs at the store money' but it isn't stretching the budget too far. Even if the mic shipped is better than the UMM-6, nothing I've seen indicates my mic is the weak point, and if I ever buy a serious mic that needs phantom power, the OmniMic is money wasted. While the OmniMic/DATS package saves me $30, I wouldn't be surprised to see a better deal from parts-express in the future.

I am still tempted to buy the little Ego. Even if it doesn't fix my measuring problem, it tells me for certain it is on the recording side, and I get a device that I can use in the future that is almost certainly better than whatever sound my motherboard and Microsoft give me as an after-thought.

[TEK]

Noise is why it is ragged, but that is still a problem, if not for measuring it is for listening. 44kHz or 44.1kHz don't seem to make any difference. Has anyone been able to get 20kHz+ sound out of a Windows 10 machine without a USB device? There is a brick wall filter in place, but I think it is part of the OS. The Realtek is probably the best possible engineering for a chip that costs a few pennies, but sound is not a very high priority for companies like Microsoft or ASUS.

Image not available

Here is the same MLS at 44.1kHz averaging 10 records.

You also have a big hump in the bass region - this implies that there is more going on then just some limitation in the upper frequences. Your curve almost look like a typical "house sound". I assume that you are sure there are no EQ or anything in your path?

[bvbellomo]

Look at the scale, it isn't that big of a hump, and the measurement accuracy of mls below 100Hz isn't as accurate. I can try a sweep to see this better, but I don’t think it is enough to affect my design even if it is real. Even the 10kHz+ problem doesn't completely ruin my ability to finish my crossover.

[TEK]

Ahh, I did not notice the scale. I do however also not know how these differences in a loopback measurement, on this scale, might or might not affect the actual sound output from the system.
0.1 difference here might not make a auditable difference at all, or it might give a 10db difference. I have no clue - but I guess someone else here do ;-)

[augerpro]

I've never set my scale that fine so I'm not sure how flat my loopback really is. I suspect it should be flatter though. I currently use a simple Behringer UCA202 (win10) and it had nice flat loop back and has worked well. Will a USB card fix your issue? Dunno. I know I've went down the "throw money at the problem" before when the issue was with my setup, a setting in windows, bad test lead, etc.

I also checked out a couple other devices like the Focusrite above from Presonus and M-Audio. I ended up keeping the Presonus 22VSL but at the time win10 would not play nice at 96khz. I need to check and see if that has improved with new drivers.

You say "sweep", but when you run MLS it sounds like white noise right? You still haven't shown your gate. What gate did you use in the loopback?

[bvbellomo]

Mls sounds like white noise. Jon referred to it as a convoluted sine wave sweep, which is not incorrect.

I have a delay before the start of the signal I can window out, but I am not getting much noise on loopback when it isn't recording so it doesn't make much difference.

[augerpro]

We would still like to see how you're placing the gate, the length, shape, etc. Doesn't ARTA show all this?

[bvbellomo]

I am still not sure what you mean by gate shape. You record audio, and pick start and end points, and it does a Fourrier transform and displays the result as a graph. I don't understand where shape fits in.

[augerpro]

Have you seen a selection that says Blackman-Harris or Hanning? Just has to do with how the values are weighted over time in the FFT, so can be displayed as a curve for visual reference. Maybe ARTA doesn't do that and just uses B-H without telling you.

Anyway, post a screen shot of an impulse where you have selected the start and stop points.

[bvbellomo]

How do I decide if I need (or want) a Zobel on these? I don't have as much ringing as the 9700s, but I think there might be some, although I could be hearing other problems.

Jon measured 3.6ohms impedence at 20,000kHz. I measured 2.7 ohms. How much variation should I expect? I suspect impedence from Dayton resistors is skewing my measurement. Any way to check without buying a woofer tester (hopefully it works on tweeters) or buying better resistors (I will be hate to pay shipping without a larger order).

I've been spending time building my 2nd cabinet and have a little Ego on the way.

[bvbellomo]

Little Ego arrived today, here is a loopback 96k MLS of length 262,144 (Hanning Window):

Image not available

This is consistent with various sweeps, sample rates and other parameters. If the line/microphone input of the Realtek is the problem, I may never find it, but it won't matter as I will be recording with the UMM6. Otherwise, I am guessing I have a bad audio cable. This level of train wreck should be obvious with critical listening. I haven't had a chance to do any yet, but I should be able to blame the Ego or eliminate it as a problem pretty easily.

[bvbellomo]

A sad update. I finished my 2nd cabinet and did end up buying the DATs/ Omni mic combo. When I went to mount the 2nd Accuton, I noticed a large crack in the cone. I suppose I deserve this for trying to save a few bucks instead of dealing with a reputable vendor. I am heartbroken, I was really looking forward to hearing these.

I did email both Loudspeakerfreaks.com and Accuton and have yet to receive a reply. I am tempted to superglue the crack. Otherwise, I don't know what to do other than buy another Accuton.

[wkhanna]

that is a real bummer!
hope you get support from the vendor.

[bvbellomo]

I am really upset. I tried to contact them earlier about the fs not matching spec, and got no reply at all. On the one hand, these sound better than anything else I heard. On the other hand, I would normally never buy from a manufacturer that treats customers this way.

I do understand if I bought from a more reputable vendor, I would have another channel to possibly go through, but listing a contact email and not even replying to it when your only other contact is an International phone number is giving a big FU to all your US customers.

I will have to try calling Europe, need to figure out how that works with my cell phone.

[ergo]

Little Ego arrived today, here is a loopback 96k MLS of length 262,144 (Hanning Window):

This is consistent with various sweeps, sample rates and other parameters. If the line/microphone input of the Realtek is the problem, I may never find it, but it won't matter as I will be recording with the UMM6. Otherwise, I am guessing I have a bad audio cable. This level of train wreck should be obvious with critical listening. I haven't had a chance to do any yet, but I should be able to blame the Ego or eliminate it as a problem pretty easily.

Are you sure you have disabled this "feature" in EGO
"Headphone blend mode – The Little Ego’s sophisticated headphone blend mode helps make your headphones sound less like headphones – and more like regular speakers."

Such features are not helpful in a measurement gear at all...

[bvbellomo]

Are you sure you have disabled this "feature" in EGO
"Headphone blend mode – The Little Ego’s sophisticated headphone blend mode helps make your headphones sound less like headphones – and more like regular speakers."

Such features are not helpful in a measurement gear at all...

I haven't seen this feature and will try turning it off, but right now I am just sick about Accuton.

[Face]

Accuton should get back to you...don't hold your breath about the vendor though.

[JonMarsh]

I am really upset. I tried to contact them earlier about the fs not matching spec, and got no reply at all. On the one hand, these sound better than anything else I heard. On the other hand, I would normally never buy from a manufacturer that treats customers this way.

I do understand if I bought from a more reputable vendor, I would have another channel to possibly go through, but listing a contact email and not even replying to it when your only other contact is an International phone number is giving a big FU to all your US customers.

I will have to try calling Europe, need to figure out how that works with my cell phone.

I'm quite sorry for your problems with your purchase- Note, it is important to distinguish between the manufacturer and the distributor you worked with. Most likely they are not even getting their product from Accuton direct, but from some one else who is an Accuton distributor.

If you check the distributor lists Accuton provides online, you won't find the company you purchased from. So, expecting relief from Accuton may be naive- it really isn't any different than if you bought drivers from someone who is an Accuton dealer, and then went to sell them online, unfortunately.

This is why working with official distributors is a good idea... even when it means paying the price.

The other possible recourse you have is shipping damage, and getting the driver replaced that way.

[bvbellomo]

I would not be too disappointed with a polite response from Accuton saying I am screwed since I didn't go with a vendor on their list. This would give me confidence to buy a replacement from a vendor on the list, expensive lesson learned, but project successful.

In contrast, a lack of response means I am completely at the mercy of whatever vendor I chose, regardless of if they are on the list, and need to factor the potential of getting screwed again in to any future projects when I decided which driver to choose.

I am interested in what you said about shipping the driver. Where would I ship it?

Compare this with the excellent level of service from PartsExpress/Dayton. I realize they aren't the same market, but should Dayton ever release anything to compete at this level, I can say who I'd buy from.

Jon - do you think I should try to glue this cone? Since the cone is still in one piece, this could prevent it falling apart, and it might have the same stiffness. But it could also void any chance of a warranty claim, and I am not sure how easily I can take this apart.

[bvbellomo]

I was too quick to jump out against Accuton. They did get back to me, and in retrospect, not in an unreasonable amount of time. They also answered my measurement issue:

Regarding the difference between our data and your measurements we’ve been alarmed and made new measurements to find the status of our drivers. And we found that materials from our suppliers changed. The actual data we will issue in our new site which is headed to be published this month. Here’s the data

Re = 6.7676 ohms
Fs = 49.8579 Hz
Zmax = 141.1786 ohms
Qes = 0.2356
Qms = 4.6789
Qts = 0.2243
Le = 0.3207 mH (at 1 kHz)
Diam = 128.6550 mm ( 5.0652 in )
Sd = 12999.9993 mm^2( 20.1500 in^2)
Vas = 16.9887 L ( 0.6000 ft^3)
BL = 11.3837 N/A
Mms = 14.4000 g
Cms = 707.6342 uM/N
Kms = 1413.1595 N/M
Rms = 0.9641 R mechanical
Efficiency = 0.8398 %
Sensitivity= 91.2596 dB @1W/1m
Sensitivity= 91.9862 dB @2.83Vrms/1m
Krm = 77.781E-06 ohms Freq dependent resistance
Erm = 1.096E+00 Rem=Krm*(2*pi*f)^Erm
Kxm = 10.744E-03 Henries Freq dependent reactance
Exm = 652.297E-03 Xem=Kxm*(2*pi*f)^Exm


This is close to my measurements. The best part is they will offer at least discounted replacement if not an outright replacement. So still possibly an expensive lesson, but worst case, I end up with the drivers I wanted very close to what I would have paid from a more reputable vendor.

[bvbellomo]

I've been going back and forth with Accuton. Today they asked for my shipping address, which is a good sign.

On the advice of people on the DIYAudio forums, I added an LCR in parallel with the 6640. It is a pretty big improvement.

[JonMarsh]

This issue that they got caught unawares by changes in supplier material is disconcerting- they should have incoming and outgoing QC that would catch that.

[bvbellomo]

This issue that they got caught unawares by changes in supplier material is disconcerting- they should have incoming and outgoing QC that would catch that.

I agree, although I'd prefer the lower FS, can you image if someone who had no interest in measurement or electrical design bought all the parts to build Wavecore Ardents, actually built them, and found out his C90-9-079 had an FS of 200Hz (actually not bad where you crossed it, but you get my point) and sensitivity of 90db @ 2.83V/meter? Or worse, never realized it was off spec, and decided they didn't like Ardents because they are too warm.

[cochinada]

If you do a search on the net you will find lots of people complaining with issues with loudspeakerfreaks.com. I don't know if they have improved recently but I've done some business with them in the pass and I can also tell you from my experience that they NEVER EVER answer emails or phone and it doesn't matter if you're in the USA or in Europe.

That being said, I cannot stop endorsing this dealer I've discovered in Belgium since some years ago (he sells Accuton). True that the site is not top of the art by no means but Fabrice (the owner) told me once that he's working alone and simply hasn't the time to work on it as he prefers to focus on his customers and the orders he receives. There is a strong chance that he might get stuff that is simply not on his site.
I have nothing but praise to say to his small operation. He's truly an honest guy and his prices are often among the lowest you will see out there as he's also not greedy.
I'm talking about http://www.compacbel.be and if I'm doing him publicity is because I think he fully deserves it and also because you can all gain from knowing it.

[bvbellomo]

Here is where I am at with the OmniMic:

Image not available

this is unsmoothed, with the short sweep.

I am crossing this 60Hz or higher to my subwoofer, so I am not worried about anything below that.
The peak from 250Hz to 550Hz is real, I can confirm it with long sweeps in RoomEq wizard.
Then it dips to about 1500Hz and gets pretty ragged after that, with a big dip at 2500Hz.
I still don't trust my 10k+ measurements on this PC, but am less worried about that, as my crossover won't affect that anyway.

I also have frequency for individual drivers (with and without crossover), and impedance data for the whole system, as well as individual drivers with and without crossovers, and my crossover design.

I can and will add some of this as time allows, but if anyone is willing to help me, it will all work better if I add what you want to see, instead of trying to add every possible way of measuring.


BTW - I highly recommend Dayton's DATS. Very easy to use, and does a number of useful things. I am less happy with the OmniMic - for 4 times the cost of the UM6 I get software that is easier to use than ARTA and RoomEQ wizard but so far doesn't do anything I can't with those or do anything better.

[JonMarsh]

Do near field (3-6" ) checks on each of of the drivers before jumping to any conclusions- your 250-550 Hz issue could be a consistent boundary effect where you're measuring them, OR, it could be a real issue with the interaction of the drivers and cabinet design. You can easily estimate/calculate the BSC effect for your enclosure *(use EDGE if you want to get fancy about it) and separate the wheat from the chaff...

[JonMarsh]

If you do a search on the net you will find lots of people complaining with issues with loudspeakerfreaks.com. I don't know if they have improved recently but I've done some business with them in the pass and I can also tell you from my experience that they NEVER EVER answer emails or phone and it doesn't matter if you're in the USA or in Europe.

That being said, I cannot stop endorsing this dealer I've discovered in Belgium since some years ago (he sells Accuton). True that the site is not top of the art by no means but Fabrice (the owner) told me once that he's working alone and simply hasn't the time to work on it as he prefers to focus on his customers and the orders he receives. There is a strong chance that he might get stuff that is simply not on his site.
I have nothing but praise to say to his small operation. He's truly an honest guy and his prices are often among the lowest you will see out there as he's also not greedy.
I'm talking about http://www.compacbel.be and if I'm doing him publicity is because I think he fully deserves it and also because you can all gain from knowing it.

Appreciate you bringing them to our attention again. I have used LoudspeakerFreaks twice with good results, but I chose what I bought carefully... (hard to mess up a Jantzen waveguide)- they came through with good pricing and low cost but well packed delivery.

[bvbellomo]

Maybe I'm calculating wrong, but shouldn't a 7 inch baffle have a 6db drop at 650Hz?

[fish fingers]

pretty sure loudspeakerfreaks are the infamous europe audio under another name- they only delivered my goods after 3 months when i made a claim back through the bank. With the amount of complaints on various forums, best avoid