|
|
Post by guitfiddler on Aug 5, 2016 13:00:23 GMT -6
I can't wait to hear the progress on the 3 ways.
|
|
ericn
Temp
Balance Engineer
Posts: 16,107 
|
Post by ericn on Aug 8, 2016 7:36:34 GMT -6
I think maybe next month I might bite the bullet and build these, unless I can move some stuff I've been trying to sell. I was just Talking to a guy who built my idea of Quested 3110 influenced clones with Volt 10 , Scanspeak Mid and Morel tweet he used a pair of Hypex plate amps to Tri Amp each ! He says he just spent a afternoon demoing the real deal and with a little DSP Tuning he nailed it ! Now I think those woofers are close to your whole driver budget but still ! |
|
|
|
Post by svart on Oct 21, 2016 11:55:26 GMT -6
Just got the bread together to buy the parts for these.. Should have all the parts in a week or so. Now I need to finalize my box design.
|
|
|
|
Post by svart on Oct 24, 2016 8:49:16 GMT -6
Now that I've been doing a few mastering sessions.. I might re-purpose these as mastering monitors. I'll need to see how they turn out first and then tweak from there.
|
|
|
|
Post by svart on Oct 26, 2016 7:51:06 GMT -6
Got the drivers and parts yesterday. I'll possibly start building the boxes this weekend, not sure.
A couple notes of interest..
The 5" "woofer" I'm using as a midrange, looks more like 3.75" of actual cone, the rest is the wide lip of the basket..
The 8" woofer looks more like 8" of cone.
I think I'll make the boxes more narrow and deep to avoid diffraction off the face. I'm guessing that I'll need more than 1cuft of open space.
I'm more of a fan of grouping the drives tightly on the face so that their output mixes quickly, for a closer sweet spot, but this complicates the isolation that the midrange will need.
I usually make the boxes from hardwood, but I'm thinking MDF on these because I want to try layers of glued MDF rather than one thick panel.
|
|
|
|
Post by svart on Oct 31, 2016 7:40:40 GMT -6
Started making the boxes. Nothing really special or nice like the other ones I've made before. I'm using 3/4 MDF, around 1.8cuft to account for the midrange chamber and ending up around 1.6cuft when finished.
I'm going to try using a step offset, so I'll build out the midrange/woofer mount area so that it sticks out 3/4 from the plane the tweeter is mounted on. I'll also round the step somewhat to help with diffraction of the woofer.
I haven't even added the bracing to the inside yet and the boxes weight around 35lbs each. These things are massive! I can only imagine that with the drivers, bracing and extra step, these will weight 50+lbs each!
I ended up around 12x16x22 in outside dimensions. Quite a bit bigger than I had initially hoped, but the woofer can really go low with a large box. A bit lower than a lot of other 8" woofers it seems, on paper.
|
|
|
|
Post by svart on Nov 7, 2016 8:40:19 GMT -6
Started one design of the front panel. I did a standard line array, but with offset high and mid drivers. This was the easiest way to get the baffle step done. I kinda want to try grouping the drivers much more closely, but not sure if I can properly do a baffle step like that, or if the baffle-step compensation circuit would screw with my crossover design.  If I like the way these sound, I might build some real nice wood boxes. |
|
|
|
Post by ppa on Nov 21, 2016 21:03:54 GMT -6
Started one design of the front panel. I did a standard line array, but with offset high and mid drivers. This was the easiest way to get the baffle step done. I kinda want to try grouping the drivers much more closely, but not sure if I can properly do a baffle step like that, or if the baffle-step compensation circuit would screw with my crossover design. If I like the way these sound, I might build some real nice wood boxes. very nice, I would advice you to use an active crossover for these ones and not an passive one, the sonic performances are much increased. If you need an hand to design it I am here. |
|
|
|
Post by guitfiddler on Nov 22, 2016 20:18:10 GMT -6
Cool!!!
|
|
|
|
Post by svart on Nov 28, 2016 12:27:37 GMT -6
So I finally got a chance to poorly wire up a crossover and put everything in the box to see how one of these initially sound..  And it sounds pretty decent! The tweeter was a little hot, so I added a couple ohms in series with the tweeter. I also found the midrange a little hot too, so a few ohms in series with it as well. The bass was light, but I haven't figured out what the optimized port configuration should be. I drilled for 3" and it sounds pretty decent even without a port tube. I think there is a little room for tweaking, but seems like mostly balancing the driver levels. The tweeter does seem a little soft in the extreme top end though. |
|
|
|
Post by guitfiddler on Nov 28, 2016 19:49:41 GMT -6
That is nice Svart! Love the Desk, and the speakers too! Lol
|
|
|
|
Post by svart on Dec 1, 2016 12:20:43 GMT -6
I modeled a few changes to the crossover hoping to fix a couple issues that I heard, but they seemed to make things worse!
I changed the notch filter on the tweeter slightly because it was a little bright in the lower range of the tweeter. It's a sweeter sound from the tweeter, but it seems to have reduced the output considerably. I may recheck my wiring as this was a quick-n-dirty change.
I slightly reduced the level of the midrange, but with the tweeter change, it's still much more shouty than I want.
|
|
|
|
Post by svart on Oct 10, 2017 11:55:46 GMT -6
So after a long hiatus on these, I've decided to revisit them. One thing I've decided to do is to replace the tweeter with something else. I've chosen some soft 3/4 domes from SB acoustics. They were relatively cheap (20$ each) and had one of the flattest responses I've seen for tweeters. www.madisoundspeakerstore.com/soft-dome-tweeters-sb-acoustics/sb-acoustics-sb19st-c000-4-3/4-dome-tweeter-4-ohms/#product-page-reviewsI've decided that between my 2 ways and my 3 ways both using Dayton tweeters, the tweeter is the factor that I'm not liking as I hear similar things between both speaker designs despite different crossover designs. There is something about the Dayton tweeter that seems to sit a little strident to me. |
|
|
|
Post by svart on Oct 10, 2017 13:17:24 GMT -6
When I left off on these, I was fairly happy with the low end, although it didn't extend as much as I expected, but that could be the box tuning itself OR that I couldn't seem to get the tweeter to sit right in the spectrum.
Now that I have new tweeters coming, maybe I'll start from the bottom and work upwards this time, getting the bass where I want it, then layering in the midrange and tweeter.
We'll see how these new tweets sound. Since they're so flat, I won't need the extra filters that tame some of the Dayton tweeter's peaks and will reduce the parts count and overall complexity.
|
|
|
|
Post by jcoutu1 on Oct 10, 2017 21:03:47 GMT -6
How about an inexpensive bigger 3 way that I can soffit mount? Something like the JBL LSR6332.  |
|
|
|
Post by svart on Oct 10, 2017 21:54:19 GMT -6
These things are pretty big.. and heavy.
They're roughly 24"x30"x15" I think.
|
|
|
|
Post by jcoutu1 on Oct 11, 2017 1:16:06 GMT -6
These things are pretty big.. and heavy. They're roughly 24"x30"x15" I think. Huh, that's almost the exact size I need. You only have an 8 in there right? Seems like a lot of volume for the configuration. I've been looking at a lot of off the shelf stuff, and nothing is that big these days. |
|
|
|
Post by svart on Oct 11, 2017 9:29:18 GMT -6
These things are pretty big.. and heavy. They're roughly 24"x30"x15" I think. Huh, that's almost the exact size I need. You only have an 8 in there right? Seems like a lot of volume for the configuration. I've been looking at a lot of off the shelf stuff, and nothing is that big these days. Yeah, it's an 8" speaker in there, but so far they do go pretty low in frequency. My goal is to be able to use them without a sub, so I'm trying to get their extension as low as possible with the boxes, which is why they are so big. |
|
|
|
Post by svart on Dec 15, 2017 8:38:55 GMT -6
I actually worked on these last weekend. I did frequency sweeps and found that the midrange was excessively loud compared to the tweeter and woofer. I also found that the crossover seemed to have a little larger gap between the tweeter and midrange.
I adjusted the crossover part values and got things a lot more in line, but it still sounds a little off. It's remarkable how different in loudness each driver is compared to the simulations.
|
|
|
|
Post by svart on Dec 19, 2017 13:17:09 GMT -6
I messed with these a little more over the last few days. I have some thoughts..
I think the tweeter really works better with a 3 pole filter.
The midrange has some kind of problem with the low pass side of the crossover. I still haven't figured out what it is. Doing frequency sweeps, there is clearly a huge spike somewhere in the 2-3K range that I can't seem to get rid of. I can pad the midrange down and it gets less intense, but seems to stay fairly high in level compared to the rest of the midrange band.
In simulation, adding more poles to the midrange crossover seems to make it WORSE.. Some kind of strange interaction is happening here. I'm going to try a zobel/snubber next and see what happens.
|
|
|
|
Post by svart on Dec 20, 2017 9:56:32 GMT -6
So I'm still trying to research what might be causing my strange midrange. I've found some reviewers mention the severe lengths they had to go to crossover this specific driver (Dayton RS125-8) and the lost efficiency they had in doing so.
So my choices seem to be:
1. Keep trying to flatten the driver response at the crossover point (3.5K).
2. Move the crossover point lower, which kinda defeats the purpose of having it at 3.5K, which is getting it out of the crucial 1-3K range where voices and guitars sit..
3. Change the drivers to something else.. The one I know will work is the SEAS MCA12 or MCA15, both of which are double the cost of the midrange I'm trying to use, but come universally lauded..
I've done another crossover sim for #1, which might work better. It changes the way I match the midrange driver into the network by paralleling some more resistance across the coil. This seems to tilt the response back a little without needing another filter pole. I also added back the little notch filter for the tweeter as it is a little bright in the 10K region and needs the notch to flatten it out a little so that it can extend to 20K without rolling off by 6dB+.
#2 is not going to happen. I really want it to crossover in the 3-4K region to avoid the human voice range, as well as the detail/clarity region of 5k-10k.
#3 might happen if this next set of crossover changes don't work. I think the tweeter and the woofer are OK, although the tweeter seems a little "slow" compared to others.
I've taken more time to read *into* comments made by folks using these dayton drivers. it seems there are unforeseen gotchas with them, usually with taming some of their strange resonance points.
|
|
ericn
Temp
Balance Engineer
Posts: 16,107
|
Post by ericn on Dec 21, 2017 8:32:08 GMT -6
So I'm still trying to research what might be causing my strange midrange. I've found some reviewers mention the severe lengths they had to go to crossover this specific driver (Dayton RS125-8) and the lost efficiency they had in doing so. So my choices seem to be: 1. Keep trying to flatten the driver response at the crossover point (3.5K). 2. Move the crossover point lower, which kinda defeats the purpose of having it at 3.5K, which is getting it out of the crucial 1-3K range where voices and guitars sit.. 3. Change the drivers to something else.. The one I know will work is the SEAS MCA12 or MCA15, both of which are double the cost of the midrange I'm trying to use, but come universally lauded.. I've done another crossover sim for #1, which might work better. It changes the way I match the midrange driver into the network by paralleling some more resistance across the coil. This seems to tilt the response back a little without needing another filter pole. I also added back the little notch filter for the tweeter as it is a little bright in the 10K region and needs the notch to flatten it out a little so that it can extend to 20K without rolling off by 6dB+. #2 is not going to happen. I really want it to crossover in the 3-4K region to avoid the human voice range, as well as the detail/clarity region of 5k-10k. #3 might happen if this next set of crossover changes don't work. I think the tweeter and the woofer are OK, although the tweeter seems a little "slow" compared to others. I've taken more time to read *into* comments made by folks using these dayton drivers. it seems there are unforeseen gotchas with them, usually with taming some of their strange resonance points. Welcome to the world of inexpensive drivers, often you find that the guys who make them work are using tons of DSP to remove the warts ! In the mids it always seams it's worth it to spend more on the driver rather than fix it in the crossover but then I'm a Keep It Simple Stupid kind of guy! |
|
|
|
Post by svart on Dec 21, 2017 12:03:36 GMT -6
So I'm still trying to research what might be causing my strange midrange. I've found some reviewers mention the severe lengths they had to go to crossover this specific driver (Dayton RS125-8) and the lost efficiency they had in doing so. So my choices seem to be: 1. Keep trying to flatten the driver response at the crossover point (3.5K). 2. Move the crossover point lower, which kinda defeats the purpose of having it at 3.5K, which is getting it out of the crucial 1-3K range where voices and guitars sit.. 3. Change the drivers to something else.. The one I know will work is the SEAS MCA12 or MCA15, both of which are double the cost of the midrange I'm trying to use, but come universally lauded.. I've done another crossover sim for #1, which might work better. It changes the way I match the midrange driver into the network by paralleling some more resistance across the coil. This seems to tilt the response back a little without needing another filter pole. I also added back the little notch filter for the tweeter as it is a little bright in the 10K region and needs the notch to flatten it out a little so that it can extend to 20K without rolling off by 6dB+. #2 is not going to happen. I really want it to crossover in the 3-4K region to avoid the human voice range, as well as the detail/clarity region of 5k-10k. #3 might happen if this next set of crossover changes don't work. I think the tweeter and the woofer are OK, although the tweeter seems a little "slow" compared to others. I've taken more time to read *into* comments made by folks using these dayton drivers. it seems there are unforeseen gotchas with them, usually with taming some of their strange resonance points. Welcome to the world of inexpensive drivers, often you find that the guys who make them work are using tons of DSP to remove the warts ! In the mids it always seams it's worth it to spend more on the driver rather than fix it in the crossover but then I'm a Keep It Simple Stupid kind of guy! I'm not sure it's a "cheap" vs. "expensive" thing. The aluminum drivers are notorious for having strange breakup nodes at higher frequencies, so it might have been more of an issue with the size of the driver, and the design of the crossover, and/or the reaction between the two. While I don't see it in simulation, some designers have complained about impedance spikes causing strange resonances, so I'm hoping a little damping at the driver itself will help smooth it out. |
|
ericn
Temp
Balance Engineer
Posts: 16,107
|
Post by ericn on Dec 21, 2017 12:13:46 GMT -6
Welcome to the world of inexpensive drivers, often you find that the guys who make them work are using tons of DSP to remove the warts ! In the mids it always seams it's worth it to spend more on the driver rather than fix it in the crossover but then I'm a Keep It Simple Stupid kind of guy! I'm not sure it's a "cheap" vs. "expensive" thing. The aluminum drivers are notorious for having strange breakup nodes at higher frequencies, so it might have been more of an issue with the size of the driver, and the design of the crossover, and/or the reaction between the two. While I don't see it in simulation, some designers have complained about impedance spikes causing strange resonances, so I'm hoping a little damping at the driver itself will help smooth it out. In my experience many of the less expensive drivers fall victim to one of the following 1 victim of smoothing of the response curve 2 poor QC / high reject rate. 2 is less common with Tang Band & SB. But yes aluminum cones can be tricky Good luck ! |
|
|
|
Post by svart on Jan 2, 2018 8:42:08 GMT -6
|
|
