SevenSixTwo - InDIYana 2018 Coax Design

[KEtheredge87]

Happy Saturday Everyone!

Today is a conflicted post for me. I finally got around to installing the curly maple veneer wrap on the test cabinet, which has been a long time coming. As the saying "Experience is what you get when you didn't get what you wanted" goes, I got plenty of experience today. I am now experienced enough to know I don't like the way veneer looks if it's only sitting on part of the cabinet. I am also now experienced enough to know that curly maple veneer doesn't always like getting massaged over a 3/4" roundover! I had to pull up the veneer on my first go-round since I didn't get things lined up correctly. That caused me to take even more time to make some repairs to that roundover... at least I got to practice the tomzarbo veneer seaming method twice on this cabinet! Overall I think I need a thinner utility knife blade. The one I used left a clean, if a bit wide seam.

After all this learning I'm also conflicted about how to best proceed with the aesthetics of the cabinet design. I will not wrap the veneer around the other two "real deal" cabinets, but I do intend to have the veneer side panels still. To avoid the appearance of the veneer sitting on top of the cabinet like I created today, I'll set up a router jig to carve out the surface on the side panels so the veneer sits more flush with the cabinet wall. That brings me to the choice of paint color...the painted portions of this cabinet were going to be satin black, but I'm concerned it will be too much black without the veneer wrap. I'll still have the black from the driver cones, so I'll need to make a new paint color choice to regain some visual interest.

This also means I'm not sure about the 1/4" x 1/8" channels I was going to cut on the cabinet to trim the veneer... Honestly, after today's adventure I'm thinking I might want to rip the veneer back off the test cabinet so I can still have 3 total cabinets that have the same visual styling. I imagine that will be a royal PITA.

[KEtheredge87]

I can definitely see why you need the LCR at 5kHz!

Yep! Glad my eyes aren't deceiving me there. Even better to hear someone else agree with my assessment. Thanks for the feedback Craig!

[PWR RYD]

I can definitely see why you need the LCR at 5kHz!

[KEtheredge87]

Hello again everyone,

More fiddling with XO stuff tonight. I got brave (relatively speaking) by taking out both of the parallel notch filters that I had in the tweeter section, then gave things a listen. It took me a while to recognize it, but there was definitely a noticeable increase in energy at the 5 kHz range. Honestly, I am not sure I heard it during any of the sample tracks I've been using (InDIYana 2017 house mix, if your curious), but I heard it loud and clear during each sine sweep taking frequency response graphs. The attached FRD plots at 0,15,30,45, and 60 degrees bear this out. I also think that leaving the 5 kHz range untouched does a disservice when the 7k-10k range filled in nicely off-axis. Using that notch filter to reign in the 5k range seems to keep everything fairly consistent until 11 kHz and above.

Having said all that about the 5k filter, what I did NOT notice was any meaningful contribution from the second notch filter that was trying to improve the ~11 kHz to 18 kHz range. All the music sounded good to my ears whether that extra notch was included or not. I'm pretty certain I can remove these three components and buy a few more resistors! The attached XO diagram doesn't show it, but several of those resistors will be combinations of series and parallel groups for the power consumption reasons that I've worried about ad nauseum to this point.

I'll take a few more measurements in a couple days when the 5mH inductor and 50ohm resistor groups arrive. Until then I really need to get off my butt and get back to the garage for some more woodworking. That veneer isn't going to install itself on the test cabinet, and I'm pretty sure I've got a few bondo places to clean up first!

I'm not sure I can ever say this enough, but seriously guys... thank you all for following along and providing helpful hints and experience. I'm sure this would be more of a "squawk box" than a legitimate attempt at DIY Hi-Fi without the guidance, and I am very thankful you're all here!

[KEtheredge87]

+1, with more practical experience you can temper the data. Same thought goes with all the parts you have in the crossover.

Alright John... You've got me thinking again. Thanks for the push (intended or not!) I'll be experimenting more tonight. More details later! -- Sent from my SM-N900V using Tapatalk

[KEtheredge87]

+1, with more practical experience you can temper the data. Same thought goes with all the parts you have in the crossover.

I'll take that as a vote of confidence that one day I won't be showing a 30+ part XO design for a 3 way speaker

I'm sure there's better ways to do what I've done... This HF section required everything I knew how to throw at it just to calm all the fidgets up to 20K. I should probably A/B Test with and without those LCRs just to see if I notice them missing or not.

[jhollander]

+1, with more practical experience you can temper the data. Same thought goes with all the parts you have in the crossover.

[PWR RYD]

Check out this thread (it's what Wolf was referring to):

http://techtalk.parts-express.com/fo...sovers-get-hot

[KEtheredge87]

Don't you do have the crossover prototyped externally? Just crank'm up as loud as you ever expect them to play. Let them play all sorts of different music that loud for an hour. Then check all the parts with your trusty fingers to see if anything is too hot to touch. If

Thanks Craig! Yep. I do have everything prototyped externally. My engineering OCD is likely getting the best of me here. I'm always trying to reconcile what I see from predictions like XSim vs. real world experience from experienced folks like yourself and others on the forum. At the end of the day, I just want to do all I can to avoid burning up any resistors (or worst case, somehow start a fire!)

I'm sure I'll end up doing what you recommend with playing music for an hour with high volume.

[PWR RYD]

Don't you do have the crossover prototyped externally? Just crank'm up as loud as you ever expect them to play. Let them play all sorts of different music that loud for an hour. Then check all the parts with your trusty fingers to see if anything is too hot to touch. If

[KEtheredge87]

Hi Guys, I appreciate the comments and help today. Although the more I look at it, the more I find myself stumped by resistors of all things... I feel like the safest bet is to combine resistors to get the values I need while still getting the XSim predicted power handling for each resistor under the rated values. That approach is leading me to unattractive ideas like splitting up a few single resistors in favor of series and/or paralleled groups to achieve the power handling side of things. That could actually lead me to break up one 10 ohm / 25W panel mount resister into a pile of five 2 ohm / 10W sand cast jobbies. That doesn't seem like an elegant solution to me.

I'm sure there's a realistic / "happy medium" approach here since we keep talking about music vs. pure sine wave inputs, but it's not clear to me what I can, and cannot get away with. Can someone try (once again) to set me straight? I think I get the concerns on the woofer circuit, but shouldn't I have the same concern about cooked resistors in the mid and tweeter path? Those two drivers needed a significant amount of padding to bring them down to the level of the Esoteric driver.

Check this mess... is this really how I ought to be doing things (Better organized on the diagram, of course)

[Wolf]

I second the 50 ohm across the woofer if the higher DCR does not suffice.
Later,
Wolf

[KEtheredge87]

...As long as you're not going as loud as Pellegrene did during a Queen DVD, you'll likely be okay...

I feel like there's a good story attached to this one that you'll have to share sometime!

Based on the concerns folks have shared here (Thanks a million!) I think I'll pick up a few of those 5mH / higher DCR air coils and see what life looks like without the 25W resistors in the woofer circuit.

[Wolf]

Thing to also know is that panel mount resistors derate when not using a heatsink. Those 25W are closer to 10W without a heatsink. Granted, that would be with continuous sine waves, but it's of concern. As long as you're not going as loud as Pellegrene did during a Queen DVD, you'll likely be okay. I've used many of those style resistors, and have not had any problems.

Later,
Wolf

[KEtheredge87]

Hi Guys, Thanks again for all the suggestions. I did a bit more fiddling in XSim and it looks like I could probably replace the 4.5 mH iron core inductor with a 5.0 mH air core with higher DCR like Chuck suggested. That would let me get rid of the two resistors in the woofer path. Beyond that, I haven't been able to get rid of any other resistors like that paralleled set of 6.2 ohm zisters in the HF section. When I try getting those out, either the response sucks, or the response is pretty good but the system impedance dips to 3 ohm or worse with high electrical phase.

I did a quick poll in XSim with all the resistors I have (minus the woofer ones). I set the XSim amp to 100W at 4 ohms (not a level I expect to play at continuously) and the highest power dissipation I see is ~30 W on the coaxial midrange "R3". Assuming a fudge factor for music being dynamic rather than a constant impulse at all frequencies, I think I should be within thermal limits. All the same, I intend to measure the external temperature of these 25W "zisters" with an infrared thermometer after playing them for a while (thanks JavadS for the idea!) For what it's worth, these haven't even gotten warm when playing some music at 80ish dB for a while. The aluminum heat sinks built into the body of these resistors does a great job!