Essentially completed sealed ZRT..

[xyrium]

I assembled John's ZRT design in a sealed cab of about 1 cu ft. The only mod to the xover was the speaker pad, in which I opted for a 9 Ohm resistor instead of the 6.

Here's a shot from REW with the RS Mic about 3.3ft away, centered, and height is between the tweeter and woofer.

Edit: usual distances from front and side walls all exceeding 2ft.

[xyrium]

This shown, I don't like the dip below 100Hz, and wasn't expecting it to drop off that rapidly just because of the sealed design. Am I asking too much? I think I may be the only person on any forums that I've searched, that has done the ZRT in a sealed config, and now I'm thinking the woofers really just aren't made for it. They probably just need to be ported.

If you look at the graph, you'll notice how these drivers hit the bottom at around 55Hz, then shoot back up at 40 and 30Hz. This is extremely dangerous behavior for a driver of this size, and I believe John mentions something about this, and that the design is not a high powered one. Therefore you need to use caution when sending more than 60W to it. I have a 125W amp powering these, and with the preamp at 10:00, you can begin to hear distress as the woofers surrounds begin to flail about.

There's also a rise in the tweeter above the crossover point that's fairly significant, and was only slightly tamed by increasing the tweeter pad to 9 Ohms as mentioned.

Inside the cab (PE 1 cu ft) are damping sheets (PE stuff) on almost every wall, including the lower portion of the baffle, along with 1" foam on the walls, with OC703 at the bottom. Of course, the drivers are recessed, and have a 5/8" roundover behind the woofers.

I also think I need to pick up a mic and re-measure, because I'm hearing some harshness that might be above the RS meter's reasonable response capabilities.

[Saurav]

By "RS Mic", do you mean the SPL meter? I'm not sure what its response is like up in the midrange / treble, but I think REW might have a built-in calibration file for it. If I were you, I would double-check the driver polarity, that dip below 2kHz looks like one of the drivers may be connected backwards. I would check that, and then re-measure / listen and see what you think. The overall tweeter and midrange levels seem fairly close, within what I would expect for an uncalibrated mic / SPL meter. What you seem to have is a dip in the crossover region, not a tweeter level that's too high or rising.

Edit: Also:

If you look at the graph, you'll also notice how these drivers hit the bottom at around 55Hz, then shoot back upat 40 and 30Hz.

In my experience, when measuring with something like REW (which uses a sweep, AFAIR), down in that frequency range you're mostly measuring the room, not the speaker. You'll have to go to some pretty great lengths to accurately measure the speaker's response at those frequencies. Try moving your mic around, and I wouldn't be surprised if you see big changes in the bass response.

All IMO, of course, and I'm no expert in any of this.

[xyrium]

Hi Sauray. Yes, it is an RS mic, and the calibration file for this model was used (I believe you may see it in the graph actually). I was thinking the dip was the crossover point, as when measured individually, both speakers exhibit the same behavior. However, it's simple enough to switch connections and remeasure, so I'll certainly try that, thanks for the suggestion.

I agree. The room is definitely influencing this setup. I've measured several speakers in my basement home studio, and all exhibit a peak from 60-70, but none dipped like this below 100.

This is by far, not a criticsm of the design, just the results in my listening room, using the materials and equipment I've mentioned. For people that do not frequent this forum, your mileage will definitely vary.

By "RS Mic", do you mean the SPL meter? I'm not sure what its response is like up in the midrange / treble, but I think REW might have a built-in calibration file for it. If I were you, I would double-check the driver polarity, that dip below 2kHz looks like one of the drivers may be connected backwards. I would check that, and then re-measure / listen and see what you think.

Edit: Also:



In my experience, when measuring with something like REW (which uses a sweep, AFAIR), down in that frequency range you're mostly measuring the room, not the speaker. You'll have to go to some pretty great lengths to accurately measure the speaker's response at those frequencies. Try moving your mic around, and I wouldn't be surprised if you see big changes in the bass response.

[Saurav]

Yes, it is an RS mic, and the calibration file for this model was used (I believe you may see it in the graph actually).

That would get you in the ballpark. The unit-to-unit variations would be a few dB at least, so you don't know your individual mic's response until you send it to someone to be calibrated (or buy a calibrated mic). The expense involved makes it kinda not worth it for that SPL meter, but you can get a calibrated Dayton mic for ~$70 I think. FWIW my mic's uncalibrated as well, I did a poor man's 'calibration' by comparing my measurement against a known correct measurement and taking the difference. It's on my list of things to get to some day

I was thinking the dip was the crossover point

It probably is at the crossover frequency. The dip has all the markings of a speaker where the driver phase matching is very well done (which would be the case for any of John's designs), but one of the drivers is inverted. Again, like I said, I can't say for sure, but that's my first gut reaction.

[jkrutke]

The dip has all the markings of a speaker where the driver phase matching is very well done (which would be the case for any of John's designs), but one of the drivers is inverted. Again, like I said, I can't say for sure, but that's my first gut reaction.

Yup. Looks like the tweeters are out of phase. Fixing that and taking a listen is probably going to drastically change the tonality in a positive way. Other than that, relative levels look good.

The RS meter doesn't work well as a measuring device. Be warned that any calibration files you may find will likely not work well because there have been about 6 versions of the meter, all different, some drastically different.

If you want to get a look at what your woofer is doing without the room, just do near field.

If your ears tell you that you need more bass, a 1 cubic foot enclosure is just big enough to vent. One of the first iterations of this design that I did was vented in 1 cubic foot. It was in a PE MTM cabinet.

[xyrium]

Thanks guys. Excuse my ignorance, but when we say the tweeter is out of phase, are we saying out of phase with the woofer, and thus, I can simply swap polarity at the terminals of either driver?

or

Are we saying that the tweeter is out of phase with the crossover, and thus the input signal itself, and I need to specifically swap the connections to the tweeter?

I performed the former by swapping the woofer terminals. I then retested using REW, and obtained pretty much the same results.

BTW, I used that exact cabinet. Though probably useless, I hit the center brace with a few hole saws to allow the woofers to breathe further back into the cabinet. I honestly have no idea if that did anything however.

I'm not exactly seeking much more bass, just flatter response below 100Hz (so, maybe you're right, and that equates to more bass ). Some of it is clearly my room, which though treated, has a hump in the 60-70Hz region. However, that falloff below 100Hz, taking the output down over 6dB seems a bit much.

Edit: I suppose all connections reference the input signal, and thus, switching polarity of either driver would resolve this anomaly. More measurements tomorrow. I'll search PE for that $70 calibrated mic tomorrow as well. If anyone has a link, please feel free...

Tks,
Paul

[davey_m]

Only $48 but out of stock ATM:

http://www.parts-express.com/pe/showdetl.cfm?Partnumber=390-801

Another option:

http://cross-spectrum.com/measurement/calibrated_behringer.html

[BobEllis]

http://www.cross-spectrum.com/measurement/calibrated_dayton.html

[Bear]

Are we saying that the tweeter is out of phase with the crossover, and thus the input signal itself, and I need to specifically swap the connections to the tweeter?

Yes.

[xyrium]

Nice, thanks guys. It appears that one can obtain the Dayton EMM-6 with a calibration file. Does anyone know if that file can be imported into REW (.cal file)?

I'll search the forum to see if there have been any comparisons between it and the ECM8000 since they are so close in price.

Tonight I'll swap the tweeter polarity. I hate dealing with those terminals on the tweeters, but I'm happy I didn't solder them at this point!

Thanks!
Paul

[Bear]

Nice, thanks guys. It appears that one can obtain the Dayton EMM-6 with a calibration file. Does anyone know if that file can be imported into REW (.cal file)?

I'll search the forum to see if there have been any comparisons between it and the ECM8000 since they are so close in price.

Tonight I'll swap the tweeter polarity. I hate dealing with those terminals on the tweeters, but I'm happy I didn't solder them at this point!

Thanks!
Paul

Look over at HTShack (home of REW) for a guy in Massachusetts doing fully-calibrated EMM6s for around $130 or so, IIRC (the EMM6 is an ECM8000 with different silkscreening). He provides both the plots and the electronic corrections for that price, which is pretty reasonable (I got one as a back-up to my Earthworks M30). REW ought to be able to use it as an input. You will also need some form of a preamp to drive the mic, so be sure to budget for that, as well (USB-based soundcards are popular).

You will want to check out this thread next:

http://www.htguide.com/forum/showthread.php4?t=35620

[xyrium]

THanks Bear. I'll probably jump on that deal for the heck of it. I believe *that* guy is *the* guy in the links above (in this thread) and he offers a basic calibration on that mic for about $70-75. I was also considering a dbx Driverack 260 or PA+ so I could cross these over at 50Hz and PEQ some of the anomalies. However, I bet that mic will work with it as well. I have an Emu 1616M which has direct mic inputs and phantom power if needed for this mic. The 1616M is a wonderful little box btw.

[xyrium]

I'll be swapping the tweets tonight and I ordered the EMM-6 Basic Plus mic.

Edit: Ok, I swapped the connectors one tweeter, and measured from 12 inches away, height is between woofer and tweeter. The picture is unfortunately, no prettier than before, please see attachment. Any ideas?

Here's a link to a photo of the cabinet interior:

http://www.hometheatershack.com/forums/attachments/diy-speakers/18148d1261527612-received-my-

[evilskillit]

Man, these are just giving you fits aren't they Was the previous measurement taken from the same distance with the same speaker or was it a sum of both speakers? Maybe one had a tweeter out of phase and it was the other one? Thats my guess, otherwise. I don't know, hopefully one of these guys can give you some handy advice.

Btw, those waveguides are still just chilling on my floor at the moment. I'm trying to find a good source for baltic birch in KC and get my crossover parts ordered.

[xyrium]

LOL, it has been a challenge. The room isn't the best either so... However, the original graph was a summed output of the pair. The latest is the output of only one with the terminals reversed. This said, I'll return it to its original config, and re-measure. If that doesn't influence the output, then something odd is occurring.

Meanwhile, I eagerly await my loudspeaker cookbook that's on backorder from Madisound. I have to figure out if acoustic suspension design requires special attention from the xover.

Edit: I couldn't wait. I reversed the tweeter again (back to the original config) and it measures flat. So, Matt, good assertion. It's probably the other speaker. Will post another measurement later! Have to head to work now that I'm late...

[---k---]

You can't measure a pair of speakers at the same time. The two will interfer with each other.

[xyrium]

Hey Ryan,

Thanks for your input. I never knew that their combined output would bring FR down over 12dB at the crossover point though. Man, I'm an idiot I guess. How is one to measure a summed response then?

Thanks!

[Saurav]

I never knew that their combined output would bring FR down over 12dB at the crossover point though.

In general, your mic won't be exactly equidistant from both speakers down to a fraction of an inch, so for certain frequencies, the sounds arriving from each speaker will be out of phase with each other, which makes them cancel out, and the mic shows a null. It is odd that you got a nice deep null like that though, when I've run sweeps on both speakers, I generally get hash / noise above about 1kHz. For the most part you measure speakers individually.

[xyrium]

So, here's a graph of the individual speakers, overlaid in REW. The rise begins rapidly at approx. 2.5k. I need to tame that, and it will take multiple PEQs to do that. Please note, this is quite a nearfield measurement at 12" distance.

[Saurav]

This is where I would wait till you get a calibrated mic before changing anything.

Also, look into an app called "HOLMImpulse". This is free and can take gated measurements, which would give you a clearer picture of what's going on than the REW sweeps. In the default mode it automatically figures out the start and end gating markers, so that's handy to get you off the ground without having a steep learning curve.

I think the demo version of ARTA is also free, and can take measurements (you can't save, IIRC).

Edit: Also, your RS SPL meter is set for C weighting, right?

[5th element]

I think the demo version of ARTA is also free, and can take measurements (you can't save, IIRC).

No you cannot save, but you can export the frequency response graphs you make from the impulse.

You want to place the mic on the listening axis at a minimum of about 1 meter. Take a measurement and then set the start and end points on the impulse such that the first reflections are avoided.

This sounds pretty much what Saurav said Holm will do automatically.

For what it's worth it looks, to me at least, like the tweeter output could be way too hot. I'd double check to make sure that you've wired up the crossovers correctly and with the right component values.

[xyrium]

Good poing about the SPL meter. I'll definitely cease further measurement until the calibrated unit arrives. Even then, I only really care about stuff up to about 12kHz. My ears are too old, and too worn from band practice when I was 18 years old to think I could hear anything above that range.

Hey 5th. Thanks for the suggestions as well. The xovers are straight from Madisound as the ZRT kit. One really can't screw them up. Double

[---k---]

Yeah, but how do they sound?

[xyrium]

There are definitely some holes in the response that I can hear below 200Hz, and something is harsh (just to repeat for non-frequent readers, this is my opinion, in my room, and YMMV) in the upper frequencies. The room probably needs more treatment, but at 12" measurements, you can easily see the rise in HF response almost immediately after the crossover point.

I normally don't mind PEQing stuff +/- 3dB, but if I have to go beyond 6dB, the sound starts getting more colored, and less natural than I would like, because it becomes difficult to choose a Q that doesn't impact more than one would wish for.

My goal was to assemble a pair of acoustiic suspensions satellites that would play down to 50-60Hz relatively flat, so that I could cross a single sub (yet to be built) over nice and low, probably 60Hz. I know it's more common to dump sound over to the sub at 80, but in my room, I believe that resonates more room wall panels and excites more room modes than I can EQ out, and thus, leads me to be able to localize the sub. This is used exclusively for 2 channel stereo, and the focus is on sound quality. Sorry, that was a long explanation eh?

So, I look forward to receiving the new mic, and figuring out holmimpulse to get a better grip on what I'm hearing. Thanks!

[Saurav]

The xovers are straight from Madisound as the ZRT kit. One really can't screw them up. Double

You would think... if you can take them out and post a photo showing how the components are connected, someone here could probably check it out.

[xyrium]

Ok, Let me see if I took photos of them prior to installlation. However, it will be difficult to determine since they use boards with solder strips on them. Some of those strips may not be obvious from an aerial view IMO. Otherwise, maybe a shot from inside the cab will work. The opening on the center brace might be large enough if I take a macro shot.

Edit: This is a pretty good shot. There's just the 9 ohm (7 ohm in the original design) padding resistor hidden beneath the 30uF cap.

http://img264.imageshack.us/img264/8898/hpfilteraerial.jpg

[Saurav]

OK, yeah, that's kinda hard to tell

[xyrium]

I figured I should check the output of the woofer alone, just in case it was contributing too much to HF production. Here's the output, it looks pretty good..until John posted below.

[jkrutke]

I figured I should check the output of the woofer alone, just in case it was contributing too much to HF production. Here's the output.

Hmmm, something doesn't look right there. This system is LR4 at 1700Hz, your woofer should be a lot more rolled off by 3.5-4kHz. It only looks to be about 10 dB down at 4 kHz.

The other measurement in post 20 (single speaker measurement) looks a little weird also. If it were a tweeter level issue, there would be a lot more output in the 2-3kHz range.

What would be a nice verification here is if you could do an impedance curve to verify crossover hookup and operation. Start with a bare measurement of the woofer, no enclosure and no filter. The woofer's impedance rise and peak will verify your measurement works. Then hook everything up and get a system impedance.

[xyrium]

Hey John,

Here's a quick image or two so you guys can see how the LP filter is connected. The two wires in the center feed the woofer, while the outside wires are the input to the filter.

http://img132.imageshack.us/gal.php?g=img1787fc.jpg

Not sure how to do impedance plots though, I may or may not have the tools for that. Have a link?

[benchtester]

Hey John,

Here's a quick image or two so you guys can see how the LP filter is connected. The two wires in the center feed the woofer, while the outside wires are the input to the filter.

http://img132.imageshack.us/gal.php?g=img1787fc.jpg

Not sure how to do impedance plots though, I may or may not have the tools for that. Have a link?

I am having a little problem verifying the connections from the inductor and cap to the terminals. Could you take a picture almost straight down (as perpendicular as possible to the circuit board)?

[Saurav]

Far right says +IN, and connects to one end of the inductor. I suspect the other inductor lead is behind the inductor, which connects to the right end of the cap, and the woofer terminal (where you can see the letter W, and probably a + above that). Then I suspect the left end of the cap goes to the 2 right tabs, which are ground (-IN and -W).

That's how I expect it's laid out...

[xyrium]

Hey guys,

I'll take a better shot when I return from work. However, here's the Madisound layout (attached). In the photo of the xover in the cabinet, the positive terminals are to the right and the common negative terminals are to the left. Therefore, the two center wires are output to the woofer. The photo can be enlarged to practically full screen view (depending on monitor size and resolution of course) as it is a 1600x1200 shot, if that helps.

Rgds,
Paul

[xyrium]

Ok, here's a closeup.

http://img198.imageshack.us/img198/7994/img1794b.jpg

[Saurav]

That looks correct to me. Does the cap say something like 12uF (or 12 followed by some number of 0s) on the side?

Do you have a multi-meter? You can pick one up for cheap at RadioShack. You could try running continuity tests, to make sure all the connections are sound. For instance, measuring the resistance between some points should give you readings close to 0:

+IN to +W: Should be < 1 ohm
Right end of the cap to COM: Should be almost 0 (test this for both wires connected to COM)

Another option would be to put the woofer and tweeter back in the cabinet (so it's sealed), but connect the woofer straight to the input terminals, i.e. bypass the crossover. Post that measurement. Turn the level down, and get the SPL meter really close to the woofer dustcap. If that doesn't match what the woofer's raw response should be like... then either your drivers are bad, or your measurement setup is suspect, probably the latter.

That's all I can think of in terms of further debugging steps

[xyrium]

Hi Saurav,

Not sure about cap value, they managed to glue it face down. I can measure continuity with my MM, but I'm thinking that not much could possibly have gone wrong in a two component filter.

Yeah, I'm doubting that as well. Maybe I simply shouldn't take a ported design and out it into a sealed box. I'm thinking that the additional output from the port may bring the woofer more in line with the tweeter output, but when sealed, that changes the balance.

Any ideas on how to do this as John specified?

"...do an impedance curve to verify crossover hookup and operation. Start with a bare measurement of the woofer, no enclosure and no filter. The woofer's impedance rise and peak will verify your measurement works. Then hook everything up and get a system impedance."

This said, if there's anyone in NJ that knows what they're doing, and that could go for a nice glass of Scotch, beer, wine, etc., while helping me figure this out, drop me a line!

[Mark K]

Can you post a picture of your exact measurement setup-i.e. where the mic is, where the unit is, where the walls are. Also, do you have an impulse response you can post? I'm having a bit of trouble interpreting your measurements.

You had mentioned distances to side walls of two feet. Are you saying there are reflecting surfaces two feet away from the unit/mic?

[---k---]

I would use a woofer tester to measure the impedance. That is probably the simplest way. Other than that, probably can do it with Speaker Workshop. Free software, just requires a little learning curve.

No offense, but I don't trust your measurements. I doubt there is anything wrong with the speakers. Probably just your personal taste.

Zaph has on his website 4 different tweeter level options and options for changing the baffle step. You might want to check these out. Very easy and cheap changes. Zaph also states that they will work easily sealed.

[xyrium]

Mark, yes, easily done. Will post shortly. The last measurement had the meter literally a few inches from the woofer, and previous measurements were as indicated in each post, either 3.3 ft or 12 inches from the source. Here's the setup, ignore the mess while I'm doing all of this prep work.

http://img413.imageshack.us/img413/5658/img1799.jpg

Ryan. Thanks for the suggestions. No offense taken. However, I don't think these measurements are lying, vs. my taste. I have already tried varied implementations of those crossovers. I used the 2.0mH inductor on the woofer, but went back to the original 2.7 because the lower mids were too forward. Both boards assembled by Madisound. For the tweeter, I used a 9 Ohm resistor pad instead of the 7 Ohm. I thought about ordering a WT3, and maybe I should just suck it up and do it. By the time I'm done with everything, I'll have a lab that I won't know what to do with... :rofl:

[---k---]

According to Thomas, you should order a Woofer Tester 2, not 3. Based upon using my WT3 again today, i would have to agree. One minute the results were garbage, the next they were fine, then back to garbage. I couldn't figure out why.

[xyrium]

Ok, thanks Ryan. I'll check to see what they PE has in stock. I suppose you'll be sending yours back... Are we thinking software, or hardware problem with WT3?

Edit: No luck. It appears only WT3 is available. I typed "woofer tester" and "WT2" in their search option, and only the WT3 is available. Noooooooooooooooooo!!!!!!!!

[benchtester]

Ok, here's a closeup.

http://img198.imageshack.us/img198/7994/img1794b.jpg

My concern is resolved, I thought I saw wires soldered to the left terminal, which they are. However, it is jumpered back to common. So it appears to be wired correctly.

[Saurav]

You're probably getting reflections off the camera tripod, which would be skewing the results, though it's hard to say by how much. My measurement setup looks somewhat similar, I need to build myself some kind of boom mic stand. Try propping a cushion between the baffle and the mic stand, or wrapping a towel or a sheet of foam around the tripod.

Also, that looks like the digital SPL meter. Not sure how much the response differs from the analog one, which is what I expect REW has calibration files for.

You could try Speaker Workshop. It's free, and has a pretty bad learning curve, but you should be able to use that to get an impulse response, and that will make it a lot easier to get a clearer picture of what's going on. Search for "Speaker Workshop tutorial", there are a couple that can help you get up and running without too much frustration. Unless you want to measure impedance using SW, skip those parts, that's even harder than acoustic measurements IMO.

[evilskillit]

The woofer tester 2 isn't carried by PE anymore. But you can get it here.

http://www.woofertester.com/

It's a bit more but I guess from what everybody says, if you're going to get one, its worth it.

Btw. I milled those waveguides you sent me down today. Getting ready to post some pics. Glad I got those spares from you... I needed 4 to get 2 good ones :E