Mini Karlsonator (0.53X) with Dual TC9FDs

Founder of XSA-Labs
Joined 2012
Paid Member
Hi xrk971,

Thanks, the additional information looks good, on a quick read I think I can work from that. A lot of good ideas, I haven't yet tried to build math functions into AkAbak scripts to automatically recalculate the cross-sections and duct length from external cabinet dimension changes, but I'll try as soon as I find some "quiet time". This should be great for other boxes too, e.g.: MLTL, and TH.

Thanks again, Regards,

PS.: Nice build logs too. :)

Thanks and you are welcome. Someone who works in an oil field and can crank out AkAbak sims?! You must be a petroleum chem E :). Applying Holmesian rules of deduction. :)
 
The AN Super8 works quite nicely in the full size 12 in K'nator. It has a high rising response that you may want to tame with some BSC or EQ'ing, otherwise it might be too bright or shouty.

I also tried constricting the top channel to 0.5X and that really helped flatten the response out (second plot). Third plot is impedance, fourth is cone displacement. All for 0.5X choked case with 1 watt in at 1 m.

Hi X,

Thank you very much for that. Most appreciated!

With regards to constricting the top channel, being a noob, just want to clarify that: So we are looking to extend the internal panels so that the gap between the top panel of the box goes down from 1 5/8" to 13/16"?
I've attached a pic to illustrate.

How would you compare the mini Karlsonator with the 20" Cornu?

(Which BTW I'm currently building out of foam core with 5mm MDF faces and the TC9FD's so all very similar to your first Cornu. Thanks for that design too! Bl**dy VIFAs cost USD $40.- to ship to NZ.. grrr)

Again, thank you!

Kiwi Pete.
 

Attachments

  • Karlsonator_mod.jpg
    Karlsonator_mod.jpg
    93.5 KB · Views: 2,033
Founder of XSA-Labs
Joined 2012
Paid Member
Pete,
You are correct, but it is easier and more flexible to build to plan and then add strips of removable foam core to reduce the cross sectional area. That way you can tune it to taste and nothing is permanent. Glad to hear you enjoyed the Cornu with Vifa's. Too bad they cost so much to ship. The Visaton FRS8 is similar. Is that easier to get?
 
Visaton FRS8 available in OZ from...

FRS 8 - 8 Ohm

Haven't bought from them so I can't comment on their service, but the price is pretty good

If this is the Vifa in question

Vifa TC9FD-18-08 3.5 Full Range Loudspeaker

then you can get that in OZ also.

Appreciate the comment blakkmarketeer , but by the time you take the higher price into account plus airfreight from AUS, it 's pretty much the same as ordering from the US. So, I'll just have to grin and bear it ...:D

Nevertheless, I hadn't come across them before, so will check them out next time I want some speakers

Cheers

Peter
 
Hi again xrk971. I have noticed that in the sims used for this tread, both for the two by and nine by, that above about 1kHz there are a lot of bumps in the response. I cant help but wonder if this is because the K slot itself is too wide at the top and is perhaps also opening out too quickly. Can you address this easily in your sims and see if that is indeed the case. Looking forward to any assessment you may be able to do.
jamikl
 
Hi again xrk971. I have noticed that in the sims used for this tread, both for the two by and nine by, that above about 1kHz there are a lot of bumps in the response. I cant help but wonder if this is because the K slot itself is too wide at the top and is perhaps also opening out too quickly. Can you address this easily in your sims and see if that is indeed the case. Looking forward to any assessment you may be able to do.
jamikl

Typical Karlsons will exhibit two somewhat large notches, the first one at the top-end of its gain-BW (200-400Hz depending on size), the second usually ~1 octave above this. The response above this is usually smoother than either Akabak or Hornresp simulations, which are ruthless in showing that frequency hash.

IG
 
Founder of XSA-Labs
Joined 2012
Paid Member
Effect of Smaller K-aperture and Distance from Wall

Hi again xrk971. I have noticed that in the sims used for this tread, both for the two by and nine by, that above about 1kHz there are a lot of bumps in the response. I cant help but wonder if this is because the K slot itself is too wide at the top and is perhaps also opening out too quickly. Can you address this easily in your sims and see if that is indeed the case. Looking forward to any assessment you may be able to do.
jamikl

Jamikl,
I wanted to see what the effect of a narrower K-aperture starting gap and smaller end aperture size was on the frequency response after reading your question. I was quite sure it would not affect those smaller bumps - which really are a function of room placement. However, I went ahead an generated a modified K-aperture function that starts at zero width and is narrower at the wide end just to see the effect for myself - something to play with which up to this point, I have not yet started messing with. Below is a plot of the original K-aperture function (blue) fitted over the data points from the K-plans. The dashed red line is my attempt at removing the initial gap and making it close faster - it never reaches the full opening of 1.0. Second plot is the freq response of the dual Vifa setup (0.53X) with the original K-aperture, the third plot is the freq response of the modified K-aperture. As you can see, it affects the lower bass response and has no effect on the higher frequency bumps. This also shows that making the slot smaller detrimentally affects the response. Finally,the fourth plot shows the original K-aperture and the effect of the placement in the room and distance away from the back wall. In this case, it is set 60 in away from the wall, whereas the above sims assumed sitting 1 in away from the wall. As you can see, the peaks get much smaller and you have more of them - so it really is a room boundary interaction that we see there. What is neat about this plot is that it shows that this speaker still performs very well bass-wise when not sitting next to a wall for bass reinforcement - continuing to show that the Karlson approach produces a speaker that kicks butt in larger spaces. :) Due to the narrowing of the freq bumps, I would say that it is preferable to use this speaker farther away from a back wall if you can.
 

Attachments

  • K-Slot-Fit-vs-Smaller-Slot.png
    K-Slot-Fit-vs-Smaller-Slot.png
    10.5 KB · Views: 1,958
  • Karlsonator-0.53X-Dual-TC9FD-Orig-K-Slot-Freq-1m.png
    Karlsonator-0.53X-Dual-TC9FD-Orig-K-Slot-Freq-1m.png
    33.9 KB · Views: 1,908
  • Karlsonator-0.53X-Dual-TC9FD-Smaller-K-Slot-Freq-1m.png
    Karlsonator-0.53X-Dual-TC9FD-Smaller-K-Slot-Freq-1m.png
    34.1 KB · Views: 1,873
  • Karlsonator-0.53X-Dual-TC9FD-Orig-K-Slot-Freq-1m-60-in-wall.png
    Karlsonator-0.53X-Dual-TC9FD-Orig-K-Slot-Freq-1m-60-in-wall.png
    34.4 KB · Views: 1,878
Founder of XSA-Labs
Joined 2012
Paid Member
Build Log - No. 2 - Stuffing & Glue Final Side Panel

Continuing, on... the next step is to add some polyfill stuffing (I use filling from an old pillow that we would just throw away) into the section from the closed end down to where the 180 deg turn is at the bottom. Start with more of a dense packing of the stuffing at the top and gradually reduce the density as you get to the bottom. Add the stuffing in smaller clumps that are loosely pulled apart at the bottom. This will let you selectively pull it out through the driver cutout when you are trying to tune the bass response to your liking. The more stuffing you have, the lower the bass SPL but the smoother and tighter the response will be. The best way that I have found to do the final stuffing adjustment is to use a microphone and software like Holmimpulse - that is the only way to quantitatively balance the speaker outputs so they are symmetric. I use about the equivalent of 20% of the filling from a standard size pillow for the stuffing just to give you and idea of the amount used.
;
The next step is to cap the speaker with the final side wall. This is a potentially tricky operation and there are several ways to do this to make sure that the side wall is sealed properly so that there are no leaks. First of all, if you cut all your internal pieces using a ruler and straight edge, they should all be withing +/- 1 mm in width. This will help ensure that the side wall can easily seal with glue. If you have larger gaps, use some hot melt glue to build-up that gap to be even with the rest of the edges prior to final gluing. You can use white PVA glue to glue the final side panel on, but it sometimes results in an internal leak where the divider is - and that will produce a speaker with poor bass response (basically ruin the whole speaker build). Use of an expanding foaming acrylic adhesive will help during this operation (Gorilla Glue).

You cannot use hot melt glue for the entire last panel on because there is too much linear distance for you to apply hot melt that it will harden before you can set the side panel on. If you are not sure, make sure your edges are flat, even and level - if you lay the side on, there should not be any gaps larger than 0.5 mm. Then apply a bead of PVA glue to all edges and lay the final side panel on top and align it with the edges of the box. Then place a stiff flat piece of plywood or other object on top of the panel and use weights to clamp it down (books work well). The observe the edges where the PVA should be oozing out due to clamp pressure, make sure there are no gaps. Clean up the excess squeezed out glue with your finger or tissue paper. Let it dry at least 6 hours before taking the clamp weights off. After you remove the clamps, double check the PVA seal for any gaps and use more PVA or hot melt glue to seal the gaps. You can add a fillet of hot melt glue at the 90 deg transition between the front baffle pieces and the side now for extra reinforcement. I use the back of my fingernail as a fillet smoothing tool to produce a nice radiused fillet joint in hot melt glue. Be careful not to burn your self! I don't have any photos of this part of the process because I was working fast to use a hybrid PVA glue and hot melt glue for the capping operation. Here, I use PVA on the external wall and edges, except the front panel where the driver is, and I use hot melt on the internal edges and the front driver panel and divider panel. This way, I have enough time to apply hot melt and I know the hot melt forms a better seal for non-uniform edges since it has good gap filling properties. It is tricky though because you still have to work very fast to apply the glue quickly before it dries. Then I continue as before with clamping as if I used only PVA but now with knowledge that the insides do not have pinhole leaks.

You are almost done now, the next step is to connect the drivers and plop them in place. I use self-tapping drywall or MDF screws to hold the drivers to the double layer foam core. Do not over tighten or else you will strip the foam/paper "threads". A neat thing to observe is to actually have music playing as you put the last driver in - you will see that the speaker come to life with the bass suddenly pouring out as you screw the last driver in.

The last step is to attach fittings on the front for screwing the K-aperture panel into place. It could be just glued but then you would not be able to remove the top driver easily. That will be for the next build log entry.

This took about 15 minutes to glue, 6 hours to dry, and another 15 minutes to do final sealing and fillets and install drivers. You can listen to it now, and it actually sounds pretty good, but the dispersion and bass extension will get much better with the K-aperture in place.

Have fun. :)
 

Attachments

  • Build-stuffing-14.jpg
    Build-stuffing-14.jpg
    200.7 KB · Views: 2,010
  • Build-fit-drivers-17.jpg
    Build-fit-drivers-17.jpg
    172.4 KB · Views: 1,034
Construction adhesive might be the best thing for that final cap-off operation. It's thick enough not to warrant surgical precision on the cut of the panels and will give you all the time you want to apply and align things. It won't run either, so you can re-orient the project as you need while doing it. Possibly longer to dry than some other stuff though. I'd also suggest laying a plywood with some weight on it once it's all in place, just watch for shifting when doing so.

IG
 
Thanks for doing that sim xrk 971. Although basically agreeing with what you said about it there does seem to me to be a little more output between 300Hz and 480Hz although still rough. The spikes on either side of that dropout seem to be a little lower also. When I am able to build this I have some 4"dia. pvc pipe and will try and make some k tubes that hopefully will come down to about 300Hz. I also wondered about a small klam with a 4"driver sitting on top of the kalson instead of the pipe tube, might need to be tilted, to get from 300Hz up to where a small dia. tube could take over. What do you think about that idea. Freddi, ig81 and Morayjames might like to comment on that idea as well. Which of the Karlson threads would be best for these queries?

jamikl
 
Founder of XSA-Labs
Joined 2012
Paid Member
Build Log - No. 4 - Mounting Wing Panel & Completion

The next step is to cut a long strip of 0.75 in wide foam core, then section those off into 1 in lengths so that you have 3 pieces for each support that will be used to screw the wing panel into place (27 pieces). Then hot melt glue the 3 pieces together to form a block, make sure one side is squared flat before the glue sets as this will be the side that is glued to the cabinet. Then glue them to inside edge of the front cabinet as shown. I measured distances and put marks to know where they go. Draw marks on the front of the wing panel where the screws go (use a ruler and make them even so they look nice). Lay the panel on top and use the self tapping drywall or MDF screws to fix the panel in place. On the very bottom, I use a dab of hot melt glue to keep it from flapping - this can be cut with a razor if you need to remove the panel later. That's all there is to it! The speaker is completed. Here it is next to its twin, a 12 in ruler is shown for scale. Note that I did not add an internal brace to the wing on this speaker. I want to see how it sounds and if it really needs it.

Ok, so now I wire it up to my homebuilt Texas Instruments TPA3116D2 class D amp powered by a 19 volt laptop brick (good for 45 watts/ch into 4 ohms). I am driving directly with the headphone outputs of a Phillips carousel CD player. Here is the first listening impressions in STEREO: I put on Dave Brubeck, Dire Straits, and Fiona Apple. The sound is simply AMAZING! Very detailed, rich, and most of all super DYNAMIC. There is punch when it happens, and the micro-dynamics, tiny sounds, fingers scraping on the wound steel string of the stand up bass, vocal clicks from the singer starting a new phrase or taking a breath, the shimmer of a high hat, the punch of a kick drum, it's all there. And it can be LOUD if it needs to be. I can stand back 10 ft to 15 ft and it still sounds very good. The best part was that the spatial imaging was still there, and the sound stage was very wide - your head did not have to be locked into a 2 x 2 foot sweet spot to get the imaging correct. I could actually move around the room over a 10 ft x 10 ft area and the imaging was still great. These are one of the best speakers I have heard and you will definitely not be disappointed if you try this build.
 

Attachments

  • wing-supports.jpg
    wing-supports.jpg
    231.1 KB · Views: 902
  • wing-supports-glued.jpg
    wing-supports-glued.jpg
    172.5 KB · Views: 757
  • finished.jpg
    finished.jpg
    167 KB · Views: 758
  • completed-pair.jpg
    completed-pair.jpg
    157.3 KB · Views: 3,599
Founder of XSA-Labs
Joined 2012
Paid Member
Update on the second speaker: the K panel needs the brace. I found that the second speaker had a lower spl when I recorded a sound clip to see one channel down. I think the brace reduces the flex losses on the K aperture. It might justify double layering the front panel. The brace also may have some effect like deflecting the initial wave into the upper slot? Anyhow, it sounds better with the brace now added. I did not have time to do a measurement yet to balance the two speakers by adjusting stuffing.