Something stirs in the Undergrowth

[IslandPink]

Bloody Travis Perkins isn't open today - on a Friday five days after Christmas ! - even I had to go into work yesterday.
I'll have to go to B&Q later.

[IslandPink]

Here's one solution to the headache of having to allow for the possibility of running a compression driver on a longer version of this horn. I'd really like to avoid doing this experiment before building the Mk.2 , so if there's a way of trying it later, after I've got the system working with Fostex drivers , then that would be good.
The concept is to split the horn somewhere around the rear lower bend, and make it modular. It's probably going to be an advantage to have the thing break down into two pieces anyway, for transport and for ease of working in the build.
I notice the 12a and 13a WE horns tend to have a removable mouth section, so there's nothing really new in this.
The sketch shows a flange, but probably you could overlap the side pieces to double thickness and have something a bit more elegant.
I'll look into this more when we get into the CAD modelling.

[IslandPink]

Here's the next mod of the Mk.1 prototype, that should hopefully give me confidence in the Mk.2 design. The idea is to open out the 2nd bend significantly and then come round in as smooth a curve as possible to join up with the inside of the 'old' third bend. I've already had the big mallet out this morning and got rid of the Mk.1B mouth structure and the outside of the 2nd bend. I like the fact that wood glue is not quite a strong as wood, and parts can ( not always !) therefore come apart quite neatly with a big hammer.

[IslandPink]

Did an interesting extra graph today.
I took the averaged data from the Radioshack (listening position ) meter readings for the Mk.1a and Mk.1b variants, and combined and averaged these two sets of data.
The result, at least from 100Hz to 400Hz, is remarkably flat. Also, the extension to 100Hz, then falling from there in the bass is more or less exactly what I expected based on the simulations.
I'm hoping this represents the 'potential' of this horn design ( length & area expansion ) in the Mk.2 horn if the folds can be made gradual and non-detrimental.
There are dips at around 600Hz and 900Hz remaining. There is definitely some of this characteristic in the basic driver data if you look at the response curves. However there may be some contribution from the first and second bends in these two horns - which is common to the data - and maybe deepens the two dips. Time will tell. The Mk.1C could be different in this respect as I am opening-out the second bend.
Basically though as I have said to SimonC , there is no obvious need to tinker with the length or area-expansion of the design.

( Bear in mind these plots include a 700Hz 1st-order low-pass filtering from the amp.)

[IslandPink]

Crikey things are going well at the moment.
Sold the LL1620 outputs, unexpectedly, on AoS . Got nearly £200 in, Paypal today. Had a look for prices on FF225WK's on Ebay , they've gone up from £240/pr to £280/pr due to Brexit, I was a bit disappointed, then I spotted a nearly new pair s/h for £165 inc postage . Now I'm still £40 quids-in and can get the OB's working better

[IslandPink]

Latest addition to the family :

[Scottmoose]

Nice. Interesting unit, the 225wk. If you can kill the minor amount of ringing (not too difficult; mainly a dustcap / cone interface issue) it's quite flexible. IIRC Steve sorted his out with some black-tak around the periphery of the dustcap. Dave fixed his via strategic placement of his favoured EnABL variation.

[Cressy Snr]

Yep, the FF225WK is a great widebander. Kill the ringing with the black-tak and the performance of the unit on vocals becomes superb. Add in a high crossed over tweeter and you have a very potent combination.

[IslandPink]

Yes, good potential , just listening to some Miles Davis ( Get up with it ), drivers ( both ) tested on the floor.
Not sure what to do just yet, if I run stereo then it's just the Supravox and FF225's on the two Baffles and I will have to see how the treble sounds and tweak as suggested.
If I stay mono & bi-amp as per current experiments, then the FF225Wk goes to 700Hz then low-pass 6db/oct from the amp and the Yuichi and ribbon come in from there to give nice upper mids & treble.

Meanwhile the work on the folded horn must continue, and Simon giving great assistance on the beginnings of the CAD model.

[IslandPink]

Just put the FF225 on one of the Quasar OB's , alone, no crossover. It's got quite a lot of bass already, sounds great, though the upper mids and the 3kHz lump are not ideal. Anyway, some fun listening to be had, there's loads of tone in the upper bass and lower mids.

[IslandPink]

OK , been doing some work.
OB matters first . Here's the FF225WK and Supravox 285GMF on one of the baffles. It's missing a top panel which would slightly affect bass extension, but that's a minor issue.

This sounded pretty good with the 'first stab' crossover or 8.2mH & 100uF in place. Subsequent listening and experimentation showed 130uF was quite a bit better in terms of dynamic coherence and upper bass tone, 140uF was going a bit 'off' , so back to 130uF - that's where I've left it. Given the low-ish sensitivity of the 225, I put a 1R in the feed to the Supravox. This was still too bass-heavy, now 2.2R in there and that sounds about right , poss 2.7 or 3R in the fullness of time with a stereo set, remains to be seen.
Anyway, it's much easier to integrate than the FE208Ez which has too big a magnet ( sorry Steve ! ) and too low a Qts.
Here's the crossover.

Now the 225 has an annoying peak at 3.2kHz and another at 5kHz, so I wouldn't want to listen to this long-term without a tweeter. it makes vocals too harsh.
The intention was to integrate this now with the Yuichi horn with a low-pass crossover at 700Hz.
Here's the bad news :
As suspected before with the FE208 struggles, I've now found for sure that adding the 700Hz low-pass ( left channel of current amp ) 'on top of' the OB and its series crossover screws the coherence and tone of the OB. I think the additional low-pass at 700Hz is already putting in enough phase-shift on the mids at 200Hz & upwards to destroy the delicate balance achieved on the stand-alone OB. Damn I don't at present know how to solve this.
So, the easiest solution ( & not easy ) is to put the crossover higher-up, say 1.5k or 2k and just use something like a ribbon to smooth and extend the FF225's response. This would require some 'treatment' of the 225 though, as the 3.2k peak will be too high up in the mix .

I have some ideas about how to do that without adding mass to the cone, as an alternative to the mods Steve C did, but will post that somewhere down the line if I can do some experiments.

[Paul Barker]

I'm urrgh ing on the side of going back to OB. Looking with green eyes at your perspex. But the doc mods concept recently introduced to my life (inspight of enormous obstructions thrown at it from friends and foes) makes me think there may be mileage in lining the back of my wooden baffles with steel. On the other hand the seminal Briggs dimension Wharfdale open baffles has a sandwitch construction with sand poured in the gap.

Might try sandfilled.

[IslandPink]

Yes, various ways of doing it. Incidentally, checking this OB for vibration a couple of days ago, the main mode is the 'flapping' of the sides in and out, there's not much fore/aft motion of the front panel to be felt at all. Suggests there needs to be a pole of some sort joining across the rear of the side-wings just above the level of the bass driver.

As always the problem of these OB designs is what do you use for the mid-driver ?
AER the obvious answer, but where to get a good pair ?
The FF225 can't match the AER, full range.
One or two of the Lowthers ( PM7C, PM8C , PM6A perhaps ) might do a good job but would need rolling higher up to bring in a proper tweeter.
Voxativ - great but where do you get £5000+ ?
One option I've not bothered to try but might work - a couple of Saba greencones, the bigger ones, Maybe even the oval 9" x 7" ones. Might be OK up to upper mids, then roll & blend in a tweeter. They are alnico which is a big plus pount - and cheap.
James speculated a ring of the smaller Fostexes - FE108Ez ..... perhaps 4-off would be enough ?

One more little thing built just now , see pic below. This is a stereo>mono summing gadget, Nick suggested last time he was here. Only works on output of the Longdog DAC because the transformer secondaries are floating, not grounded. Hence you can ground the '-' side of one transformer and stack the other one on top of it. The grounding is done at the power amp.

[IslandPink]

Meanwhile some horn progress, Mk.1C .
Inside of the new final curve being built up. The second bend is now not so abrupt - about a 70 deg bend, not a 160 deg bend. The design is still being worked out, I haven't got it clear how I do the section from half way around this rear part to the mouth, as it gets wider than the original construction. Luckily there's lots of spare wood hanging around and it doesn't have to look neat or last very long.

[IslandPink]

More horn work and more surprises in the 3D world of curves and flats interacting.
The new bend is progressing and a side piece for the next 40cm or so has been added on one side.
One of the thorny problems I was building up to, is the side-piece that has to go around the last curve to form the basis for the thickness and guide to the outer curve build-up. I knew this would be tricky , and decided a cardboard template would be the best approach - trim with scissors and then mark out the wood, in this case 12mm ply.
Pic below shows this before cutting the first piece of wood. Note the way it looks as it interfaces to the end of the existing horn ( bend 2 ) . Lovely slant of about 10 degrees there