Build therad: OLA-WG 2.5 - a 2.5way with SB midwoofers and a Bliesma tweeter

**ergo** · 08 February 2019, 19:41 Friday

Waveguide experiments

I've tried the Bliesma tweeter on a flat baffle and 4 different waveguide profiles/depths this far. Flat version was with a 160mm ring spacer to fit same hole as waveguides.

Lets see if the "big picture"link will work here.

https://npsxja.bl.files.1drv.com/y4mzjS_BW14GswDmFH5CpcWw67hWZ_pneEbY_4RRKe7eJYclvtqf-V-ZAwMs3OQuApM4qfa6dVqRZrsMOc1mC-TCkrBxpH38Q063nFZXQBBiZvmjFRiSf8PFqzoYZjX3BjV4jD98onVUGWvi4iYE0KbwuqUjnzD-ECK44D5Y0kMCY7_ewpJcKFxWMgUUG2unUQj1-SajWYr9p-IiTb7iAOE_Q?width=3349&height=1119&cropmode=none

As the test baffle has sharp edges it makes the edge diffraction evident especially on flat panel mounting as well as for the shallow waveguide. The deeper the WG gets the less it is influenced by the front panel edge diffraction.

Update: 02.March 2020 - the download link for STL files of all of the above waveguides + holder bracket HERE

**ergo** · 08 February 2019, 19:42 Friday

Placeholder for box details and build

One of the for interest trials done was to measure T/S
a) 'out of box' and after applying only 3min of high excursion sine at Fs
b) T/S after 12 hours high excursion burnin at FS

UniBox simulation for the 30L box to be built

Box is a relative simple MDF construction with 2 layer front panel

Front panel has rounded corners with R22mm roundover bit (biggest I found for DeWalt I got)
Paint is a "Granite simulaton" spray can paint (with 2 layers of undercoat paint - also spray can variant)

A 3D printed template was nice and easy way to sink the terminal plate

**ergo** · 08 February 2019, 19:42 Friday

Look of the speaker in final form

**ergo** · 08 February 2019, 19:43 Friday

Placeholder for Xover and final measurements

Final Xover after numerous small tweaks based on listening

And the Holm Impulse measurement of 1m L and R speaker measurements at their final position. About 10deg off axis horizontally

**ergo** · 09 February 2019, 12:32 Saturday

Adding an image with two of the above waveguides + the midwoofer off axis as well for reference

**sdl2112** · 09 February 2019, 14:31 Saturday

Very nice project! I appreciate your detailed testing. I am very interested in the Bliesma tweeter and your listening impressions once finished. It is unlike anything on the market specs wise. Huge Xmax and good dispersion. I almost want to try it in my SSA-WG, maybe one day. I look forward seeing your build progress. Good luck!

**Pythagoras** · 09 February 2019, 18:44 Saturday

I have the T34B-4 tweeters too for use without a waveguide.
Have listened to the tweeters with a quick test setup and they sound just perfect really. Unlike anything I personally heard before.

What I do wonder though. Did you order them without the protective grill or did you remove it yourself?

**wolf_teeth** · 10 February 2019, 02:35 Sunday

I think he stated on the diyaudio forum that he removed them himself.
Wolf

**ergo** · 10 February 2019, 13:59 Sunday

Hi,

Yes we removed the grilles together with my friend. The manufacturer was actually very helpful and gave guidelines how to do it. Stating of course that it will void the warranty.

Heating needed to looseen the glue around the grille is about 80deg celsius. We checked that measuring the air flow out of the heat gun at 10..15mm before tackling the tweeter. For my gun that mean setting the temp to 110deg as the gun controls it at exit pipe I assume.

The other aspect to worry about is that the magnet on Bliesma is super strong. So one has to be very careful with the tools used AND the grille itself is magnetic, so when loosened it is danger that magnet pulls it into the dome. We mitigated risks by using bamboo sticks and then with 4 hands we managed to heat and remove the grille with low risk of it jumping to a wrong place.

In the tweeter box you see also the temporary covers I made to protect the dome when the tweeter is not intesting or in WG.

**Pythagoras** · 10 February 2019, 16:23 Sunday

Oef that takes nerves!

Thanks for sharing the process. Glad it went ok.

Btw did you measure any differences with the grille removed? I was already wondering if it had any acoustic effect at all.

**cochinada** · 11 February 2019, 09:18 Monday

Hi Ergo,

Do you have pictures of the disassembled Tweeter? I can't understand how the plate is attached as I see no screws at all. :roll:

**ergo** · 11 February 2019, 11:48 Monday

There are no screws on front nor back. So probably a heat activated glue is also used for front panel attachment. Changing the dome/coil assembly seems not possible for end user.

**cochinada** · 11 February 2019, 15:21 Monday

Thanks Ergo. Do you glue your waveguide to the adaptar or what?

**ergo** · 11 February 2019, 16:00 Monday

As I knew that I want to try several waveguides I created all WG-s with outer diameter of 160mm and back side flat with 140mm. Then there is a 140mm mounting plate that holds the Bliesma without screws for now. I then use hot glue to just glue the rounded spots around the perimeter. That is enough strength to keep things in place for testing. This is not air tight, but I did not need that yet. For final design I'll consider creating a WG where the Bliesma will be held in place with screws also and we'll put some thin casket in between the tweeter and WG.

Hope the pictures explain the testing version.

**ergo** · 16 February 2019, 16:15 Saturday

I've been playing with the measured data on 4 alternatives
a) Bliesma on flat baffle (no waveguide)
b) Bliesma in shallow 10mm deep waveguide
c) Bliesma in MCM (Modula MT mkII) type waveguide
d) Bliesma in deepest (46mm deep) oblate spheroid profile WG

Upper woofer was measured and merged with a low frequency measurement at 30cm gated with 56ms - so has a bit of room modes in it.

Lower woofer data was generated from top woofer data by using different baffle step simulation profile and adding the 57mm (165us) extra delay.

Power response matching of midwoofer / tweeter improve with each deeper WG. The deepest one shows pretty good graphs... but the directivity index climbs quite high in end. My friends new room has quite a lively reverb, so it might help there. We'll have a think about which WG to go with.

Current proto front panel has sharp 90deg edges. Final design will have rounded over edges - so all plots will probably look better around 2.5k and 5k, but how much exactly is hard to guess.

Next step is building true proto speaker boxes. Then I can put both woofers in a box and we can get a listen. I also ordered the Hypex UcD36MP kit, so I can run these speakers with LspCAD crossover emulation too before we build the final crossovers. Maybe well even try two different waveguides still.

**cochinada** · 16 February 2019, 17:11 Saturday

I like the last one better because for one, the power index is more 'regular'. I would just try to get a more horizontal on-axis response but that's only my personal preference and I do understand the trend that Troels G. also appears to be following these days which is a tilted response.

**ergo** · 17 February 2019, 06:58 Sunday

The overall voicing is something thats hard to set at this point. Both Modula MT mkII and Minerva projects as well as one earlier project with waveguide have all indicated that they sound a bit 'shiny and sharp' with a flat on axis response. Tilting it down smoothly has helped with that. It might be to do with a power response that drops less in high and there is more energy in room as a result and to compensate one needs to take the on axis down a bit. The other aspect with waveguides is that they often work best with less tilting toward the listener. For my Minervas for example the 20deg off axis pointed towards listener seems to work the best for me as an example. In above images I decided to take a middle ground and show all of them at 10deg off axis as the main axis.

**cochinada** · 17 February 2019, 07:42 Sunday

That's a perfectly logic assumption. What matters the most in the end is how they sound in your specific room and if you like it or not.

I've been reading this book from Floyd Toole "Sound Reproduction - The Acoustics and Psychoacoustics of Loudspeakers and Rooms" which most of you guys have probably read as well, so I'm surely biased also by my beliefs. I'm probably another 'victim' of psychoacoustics but I can't let go of a flat response concept that easily but pursuing a tilted power response at the same time.

BTW, I've already an acoustic project ongoing for my living room where I hope the results will be highly noticeable. Do you have any specific treatment in your room Ergo or are you limited by that terrible WAF factor?

**ergo** · 17 February 2019, 10:08 Sunday

This specific project goes to my friends living room. He has some acoustic treatment plans, but options are somewhat limited as usual. The rooms is sort of part of an open space, so the kitchen is in one corner of the room and front entrance to the flat and access to balcony and other rooms are on various walls. So getting a good sound will be a bit of challenge but one that we are hoping to solve among other things with waveguides and good power response

**sdl2112** · 17 February 2019, 15:04 Sunday

I have found voicing one of the more difficult aspects of loudspeaker design. When I started my design I was like...what is my design target? I really didn't know. One of my goals was to correlate what I see in graphs to what I hear. An on-axis response really doesn't tell you much. So I proceeded to design for flat then adjust with external shelving circuits (for now passive between the preamp and amp). As a matter of fact I just made a change Friday to increase the downward slope. Anyway I found this review of the Dutch and Dutch 8c by Mitch Barnett interesting. He is a proponent of a downward tilt and has referenced work from Sean Olive, Floyd Toole and others. He even goes off on a soapbox on it. As I said it's an interesting read and it sure looks like the DD 8c use the Wavecor 8" woofers. Overall a pretty interesting speaker.

**ergo** · 18 February 2019, 15:16 Monday

Thanks for the "Dutch and Dutch 8c" link. Interesting speaker and worthwhile to read review. The off axis and power response preference studies are also covered deep in the before mentioned Floyd Toole "Sound Reproduction - The Acoustics and Psychoacoustics of Loudspeakers and Rooms" which I own and read also.

***
My friend drew up the box and calculated the pieces and will order the material and cutting to dimension from his friend who is in furniture business. So it'll take some time to get those and then we can progress with speaker.

***
Meanwhile I got the Hypex 6ch amp and enclosure from Audiophonics.fr. I had thought about getting an amp for LspCAD speaker emulations so I could run project with that before ordering the passive parts. I did that a lot years back, but then the big analog amps got stashed away and seem too hard to dig up when needed. A small class-D seems a perfect fit for the task. The nice small enclosure was also on sale, so the amp + enclosure + terminals set me back 200EUR which seems not bad. I'd just wish that Hypex would also include a wiring kit. Had to hunt those specific connectors from RS components and it'll take extra time to arrive.

I'll also add a 6 channel pre+volume control I built back in 2007. That too has been sitting on shelf and will get a new life. Connecting amp direct to soundcard always makes me a bit nervous when tweeter is connected there without protection too.... software can never be fully trusted. So having a master volume before the amp makes things safe + it's easier to adjust than the volume from LspCAD

**ergo** · 17 March 2019, 11:37 Sunday

A bit of progress on both fronts

MDF for proto boxes for the OLA-WG arrived. Now waiting for the Festool DF500 domino tool that I ordered with inspiration from this forum too. Should make it times easier to make a precise assembly of the box.

Also mechanical placement for the 6channel Hypex is done with help from my friend. Now some wiring ahead. I have all connectors and cables needed, just have to find the time.

**JonMarsh** · 26 March 2019, 19:55 Tuesday

Originally posted by ergo

Thanks for the "Dutch and Dutch 8c" link. Interesting speaker and worthwhile to read review. The off axis and power response preference studies are also covered deep in the before mentioned Floyd Toole "Sound Reproduction - The Acoustics and Psychoacoustics of Loudspeakers and Rooms" which I own and read also.

***
My friend drew up the box and calculated the pieces and will order the material and cutting to dimension from his friend who is in furniture business. So it'll take some time to get those and then we can progress with speaker.

***
Meanwhile I got the Hypex 6ch amp and enclosure from Audiophonics.fr. I had thought about getting an amp for LspCAD speaker emulations so I could run project with that before ordering the passive parts. I did that a lot years back, but then the big analog amps got stashed away and seem too hard to dig up when needed. A small class-D seems a perfect fit for the task. The nice small enclosure was also on sale, so the amp + enclosure + terminals set me back 200EUR which seems not bad. I'd just wish that Hypex would also include a wiring kit. Had to hunt those specific connectors from RS components and it'll take extra time to arrive.

I'll also add a 6 channel pre+volume control I built back in 2007. That too has been sitting on shelf and will get a new life. Connecting amp direct to soundcard always makes me a bit nervous when tweeter is connected there without protection too.... software can never be fully trusted. So having a master volume before the amp makes things safe + it's easier to adjust than the volume from LspCAD

[ATTACH=CONFIG]28957[/ATTACH][ATTACH=CONFIG]28958[/ATTACH]

Isn't it interesting how sometimes things come together with pieces of knowledge and/or hardware accumulated at another time? Interesting project, Ergo!

:T

I will confess that the most interesting part was finding out about the BlieSMA organization and their new tweeters- especially the T34B. So interesting, that now Solen in Canada has two less in stock than earlier today...

There are several possible homes for it... if it pans out as testing of their two models has done for HiFi Compass, Troels, and AudioExpress, there are at least two projects it will go into. Possibly three under consideration, considering the sensitivity.

Remarkable.

**JonMarsh** · 27 March 2019, 08:59 Wednesday

Originally posted by ergo

Hi,

Yes we removed the grilles together with my friend. The manufacturer was actually very helpful and gave guidelines how to do it. Stating of course that it will void the warranty.

Heating needed to looseen the glue around the grille is about 80deg celsius. We checked that measuring the air flow out of the heat gun at 10..15mm before tackling the tweeter. For my gun that mean setting the temp to 110deg as the gun controls it at exit pipe I assume.

The other aspect to worry about is that the magnet on Bliesma is super strong. So one has to be very careful with the tools used AND the grille itself is magnetic, so when loosened it is danger that magnet pulls it into the dome. We mitigated risks by using bamboo sticks and then with 4 hands we managed to heat and remove the grille with low risk of it jumping to a wrong place.

In the tweeter box you see also the temporary covers I made to protect the dome when the tweeter is not intesting or in WG.

[ATTACH=CONFIG]28920[/ATTACH]

Also appreciated that you shared the process. :T

Gives food for thought and planning...

Now, where did I put those bamboo sticks? :W:

**ergo** · 12 April 2019, 04:40 Friday

I had a question in PM about the graphs and curves, so answering here too. The new VituixCAD version allows exporting individual graph with legend, so that might help.

The off axis measurements I do are geberally 10deg steps in horizontal direction from 0deg ot 180deg (sometimes 0 to 90 or 100 to make the graph less 'busy')

The line with 'bold' in VituixCAD means the "reference angle". That is user defined and in my case it is 10deg. With waveguide speakers I usually do not use the 0deg on axis curves as a reference axis as experience shows that pointing then a bit off axis from direct path to ear sounds better.

***

This weekend should bring some progress to this build finally. I have meanwhile got all the walls for the boxes and the speaker holes and recesses for the front panels have been milled already. So the glue up is next using the new and shiny Festool DF 500 domino system for alignment.

**ergo** · 22 April 2019, 04:02 Monday

With the slogan learned from Jon "Slow work takes time"... but there is some progress.

The 6ch amp is ready. With a kit from Alibaba it has a remote volume and power on/off functionality as well. Tried a 3D printing with transparent PETG to make a 'window' for the power LED and IR. The IR penetrates that well enough. I might consider redoing this by milling some semi transparent acrylic, but for now this works too and is ready for "xover simulation" with LaspCAD as soon as I get the boxes together.

The OLA boxes are glued up. Outer front panel glue up to come still. Those are cut already including the driver recesses and holes. Festool Domino method was used and I must say that after some initial learning it is easy to see how efficient fast and precise this can be. Usually it is hard to get all 3, but the DF 500 sure helps.
My friend wants to try with acoustic panels as internal wall lining, so we will give that a go. We'll see what it does to apparent box volume and if the box stuffing can be 'too dead'.

**cochinada** · 22 April 2019, 09:13 Monday

Hi Ergo,

I've bought this for my amplifier in the making.

**Zvu** · 22 April 2019, 18:50 Monday

Check the tracking between L/R channels in that motorized potentiometer. It can be way off.

**ergo** · 23 April 2019, 02:01 Tuesday

Joaquim, thats a nice looking pre indeed. My main pre-map uses the Bruno Putzeys Balanced Preamp + a relay switcher for a volume. This one is just for occasions where more than 2 channels are needed.

Zvu - that a good point. I have not done that on this unit. Years ago I had an amp with the same series Alps pot and that one happened to be relative ok, but yeah I'll try to measure this one as for xover comparison thats quite crucial.

**Zvu** · 23 April 2019, 03:18 Tuesday

I like to use THESE in front of an amplifier for setting the levels of individual drivers and i use software for higher resolution attenuation. That way i am sure that the levels of both channels are exactly the same.

**ergo** · 23 April 2019, 06:29 Tuesday

Thanks for that reference Zvu. Affordable and nice option indeed it seems.

I'll check the Alps performance and if not good enough I can either go similar route or max the vol out from the pot and use the Master volume in the RME UCX mixer that I'd use for this.

**ergo** · 11 May 2019, 11:03 Saturday

It's been a while since I've updated this thread, but we have made progress with the build. Yesterday the double front panels got attached with a 2mm bitumen in between inner and outer front panel layer.
Paint is a "sandstone effect" can paint for both the box and front panel. Plan is to paint the WG too once I print it.... Just prepared the new STL files and will try out the progressively thinner layers to make the outer part of WG more smooth

**wilbur-x** · 11 May 2019, 13:29 Saturday

Nice ! Skills, nerve and talent are a formidable combination.

**ergo** · 12 May 2019, 01:39 Sunday

The variable layer thickness works quite well. Most of the WG is with 0.2 mm layer and upper part is then gradually finer. Took 10.5h of printing, but thats ok when it goes on over night.
Gray one is for comparison where all layers are 0.2mm

**Scottg** · 12 May 2019, 02:09 Sunday

Originally posted by sdl2112

..Anyway I found this review of the Dutch and Dutch 8c by Mitch Barnett interesting. ..it sure looks like the DD 8c use the Wavecor 8" woofers. Overall a pretty interesting speaker.

Though a bit off-topic:

SoundStage! Hi-Fi | SoundStageHiFi.com - Dutch & Dutch 8c Active Loudspeakers

https://www.soundstagehifi.com/index.php/equipment-reviews/1270-dutch-dutch-8c-active-loudspeakers

-of particular note, look at the non-linear distortion on the lower-end:

SoundStageNetwork.com | SoundStage.com - SoundStageNetwork.com | SoundStage.com

https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=2073:nrc-measurements-dutch-dutch-8c-loudspeakers&catid=77:loudspeaker-measurements&Itemid=153

IMO this is why you don't design small(ish) bass-boosted speakers.

**Scottg** · 12 May 2019, 02:10 Sunday

Originally posted by ergo

It's been a while since I've updated this thread, but we have made progress with the build.

It's looking really good!

**ergo** · 21 May 2019, 05:25 Tuesday

A short update.

Last Saturday we made some progress with adding the terminal plates. My initial plan was to make a milling template for the terminal plate with 3D printer. That had sort of a 'spectacular fail'

So all in all starting a print and going to bed too early is not a good idea until these things get 'eyes' in form of a camrea or such, so the machine would know to stop when something like this happens.

So on to making it out of MDF with a CNC.... and it resulted in pretty good result as expected (my CNC is still more precise than the Prusa 3D one and when it does fail it is usually less spectacular.)

Drilling all the screw holes for drivers. The dowel on the drill bit is to make sure the chuck would not damage the paint when a hand-drill makes a sudden plunge while reaching the end of the material.

WG-s are also painted and ready....

So getting very close to a stage where we can wire it all up and close the holes with drivers so I can do the measurements for the Xover...

**ergo** · 02 June 2019, 11:30 Sunday

Progress done with the build

All drivers are in and 3 pars of cables going out of speaker terminal holes for now for measurements and initial xover.

Got to test the new manual turntable I built some months ago with a big Lazy Susan bearing from Alibaba.
I CNC-ed the tick marks and then printed a 'spring assisted stop' for it

Raw Bliesma T34B-4 + waveguide directivity (mic at 1m on tweeter height)

Upper midwoofer MW-16P (mic at 1m at tweeter height still)

PS. Design wise - the tweeter waveguide screws will probably be painted with the gray as well. Black as they are they become a bit too "eye catching" which is not the goal.

**ergo** · 02 June 2019, 11:35 Sunday

Initial simulation of Xover that I'll be simulating first with LspCAD to voice it a bit further before ordering the parts

and the reverse null with tweeter inverted

**Zvu** · 02 June 2019, 12:39 Sunday

That looks ridiculously good. I'd probably just do something about that hump from 120 to 400 hertz. Other than that, i'd say your work is done.

**cochinada** · 02 June 2019, 12:45 Sunday

Very nice Ergo! Perhaps you already explained but I don't remember it. Do you have one of the woofers connected in series with a resistor or coil and then in parallel with the other?
Will you try to lower that small bump around 300Hz?

**ergo** · 02 June 2019, 14:25 Sunday

Woofers are in parallel, but each has it's own LP xover.

I'll try to measure also via simulated xover and confirm if I see this hump or not. Below is the raw measurement without xover and with two woofer in parallel. Speaker is mid room and the floor reflection (and other room modes) are visible.

Also my friend will mate these with two 10'' closed box woofer and he likes stronger lows + will use these for home cinema purposes also. So a bit of heavy in low end might be just to his liking. We'll see.

**Zvu** · 02 June 2019, 14:48 Sunday

I've seen couple of talks about trends in hifi audio these days. They say that 2 channel sound is slowly dying, sales go down, people listening more on their earbuds etc. While i do believe that people more often listen to music on their earbuds these days i don't think that 2 channel is dying. I also believe their sales go down and when i look at projects like this, i absolutely understand why. Sales go down because of numerous projects you can find online and make better speakers for less money than any factory can offer. I think we nailed passive crossovers and making our own stereo.

**draki** · 02 June 2019, 15:09 Sunday

Ergo, nice and elegant!

A question: are you not setting the vertical (Y) coordinates for the individual drivers? The simulation shows all Y's at zero. Obviously the X and Z (looking at the pics) are supposed to be zero.

Could you post the impedance graph set to 20 Ohms range?

Build therad: OLA-WG 2.5 - a 2.5way with SB midwoofers and a Bliesma tweeter

Build therad: OLA-WG 2.5 - a 2.5way with SB midwoofers and a Bliesma tweeter

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment