TG "The Loudspeaker" Build (Sort of)

[Ant]

Love the three bobbin CNC winder, but it's unsuited for 1.6mm wire - you need a hand winder for that, or even just a stick and count the turns mentally. Agreed, whilst people talk about winding coils on a lathe and using its autofeed (but how many have toggle reverse on autofeed?) even a little lathe has enough power to make it dangerous.

I bashed together a hand crank for my md65, mainly for thread cutting on thin stock, i wanted that functionality available. Just shoves in the back of the bore, tighten the nut and it expands, a reverse collet, i cant for the life of me think of the correct term.
The bit that goes in the bore is a tube with 4 cuts in the end, a tapered rod with a thread at the other end goes through the tube, nut on the other end bears on the end of the tube so tightening it pulls the taper in and opens out the cut end of the tube to grab the inside of the bore. Then you can crank it around under hand power, engage the carriage and run it. Cant reverse the carriage direction while turning the head the same direction with the md65 though i dont think. Might be wrong i cant remember

[Morgan Jones]

... so tightening it pulls the taper in and opens out the cut end of the tube to grab the inside of the bore.

Expanding collet? Just like bicycle handlebar stems into steering head in the 80s and 90s. That's something I haven't made for my MD65. But it would be useful. And you're right; MD65 can't do LH threads. Not unless you add a reversing wodgit to the banjo. Now that would be a useful thing...

Mistyblue: If you can measure current as a function of frequency with constant voltage, you can calculate inductance. How powerful is the motor on the TOS lathe? 2HP, 3HP? Far too much for safe coil winding; it would be perfectly capable of messily severing a finger. Fifty years ago, I bunked off school one day to go to a mechanical engineering exhibition. One stand had brochures showing what their safety equipment could prevent. I suppose with WWII in living memory of many people, they were less squeamish then than now, but even I was surprised by a picture of blood-splattered machine tool with severed finger on it. Made the point, though. Hand wound on a stick will do just fine for 1.6mm.

[MistyBlue]

Sorry the largest I have are 5.6mH, 1.4mm dia. wire , 0.15 ohms. They are transformer coils for sub-woofers and are these: https://www.monacor.com/products/compon ... -/lsi-56t/

That’s OK. Thank you very much for taking the time to look though, I really appreciate it.

The TOS SN40 is 5.5KW (!) motor. (7.33HP)

To be brutally honest, that machine always scares me, ever since we first got it. I remain in what I consider a very healthy state of controlled fear every time I use it. I don’t use it unless I feel my best and have my wits completely about me, and I make sure that the other member of the household is aware of what I’m up to. It’s bigger than what we need, but beats the Chinese offerings we used to rely on.

I did actually have a go at the toggling the feed earlier to simulate what would be necessary for it to work, and despite having some lag in the gear train on the apron, with appropriate gearing on the feed and some fine tuning of the speed via the VFD, there would be a sweet spot where it would be possible.

Whether I would want to is another story. There is just too much risk for my liking. That sort of coil winding job is better suited to a Myford or similar size machine. In fact, I wouldn't want to consider it unless there was a real possibility of being able to physically stop the spindle in an emergency.

The lathe also features a hefty braking system, mechanically coupled to the control rod that stops the spindle dead every time the lever is pulled into the stop position. However, there is no footbrake, and the spindle stop does not stop quite so instantly at higher speeds. I fitted an extra emergency lever near to the chuck as a safety precaution, but if something goes wrong with a machine this size, I accept that I would be f*****.

Don’t worry – what I had in mind was hands free operation with properly designed and constructed accessories with loads of safety features.
Unless I was going to make a habit of winding coils, it’s easier just to buy them. The tooling necessary to do it properly and safely is just too much effort for the sake of a couple of weeks and the cost of setting that up. Although it’s good to consider what is possible “on paper” as a purely academic exercise.

I have heard about some of the horror stories, and had one or two close shaves myself before. Enough to drum it in to me that you don’t take chances with machine tools. I consider it in the same way I was trained with safety in chemistry, and having had just one or two very minor mishaps there was enough for me to reconsider my life choices.

[Morgan Jones]

I've just looked up and discovered just how big a lathe that is. Controlled fear is definitely appropriate. It's a serious monster. Unquestionably too big for winding coils. Nice, but I'll stick with my Bantam and Hobbymat. I occassionally fantasise about a Hardinge or Schaublin but realistically I don't have the skill to benefit.

However... The nice thing about winding your own coils is that you can quickly modify them or make another. A few months ago, I wanted 1500VDC at <1mA from a 12VDC battery, and a custom wound ferrite transformer plus BD131 transistor and a few Rs and Cs did the trick. The whole thing fits in the bottom of a yeast tin. A better transformer worked even better. All (fairly) quickly wound on manual coil winder. Controlled fear was engaged when testing supply. Oddly, the best probe proved to be a 2.5kV rated oscilloscope probe rather than the (rather clumsy) 40kV EHT probe for DMMs.

[MistyBlue]

Re: the -14.8db problem.

I had a look through my simulations made months ago before I commenced the speaker build and was reminded that the passive crossover for the LF/MF sections was the most sub optimal. I had a go with some of the measurements I made, but it looks unsatisfactory, so I’ll stick with the active 24db filter. The whole thing works, and we are loving the sound, but we’ve lost the headroom, so I’m going to have some boards fabricated that try out a number of options.

The simplest (and probably most sensible) option is to fit an opamp based gain stage after the crossover in it’s enclosure as part of that device. I’m spoilt for options here, and if I do it through hole with a DIP8 socket, I can swap out opamp devices until I can no longer hear a difference and go blue in the face.

If I go with a dual one based on a common device like an OPA2134, that makes a nice and compact PCB that I can make modular (one PCB for 2 LF channels, and another for the 2 HF), I can swap around and optimise for frequency as well. If I’m feeling like splashing some cash, I can try some of the rather expensive discrete drop in replacements from various manufacturers and frustrate myself further.

The other option is to have some more FE2022 boards made up and see how they compare with the opamp stage. The 2 FE2022 boards I have were purchased from DIYaudio for a very reasonable amount that included the matched JFETs. DHL, however, wanted to indulge in extortion, so I will begrudgingly reverse engineer the boards in KiCAD, have JLCPCB make them, and match the JFETs myself.

I will probably do both and see.

[Paul Barker]

Sorry the largest I have are 5.6mH, 1.4mm dia. wire , 0.15 ohms. They are transformer coils for sub-woofers and are these: https://www.monacor.com/products/compon ... -/lsi-56t/

That’s OK. Thank you very much for taking the time to look though, I really appreciate it.

The TOS SN40 is 5.5KW (!) motor. (7.33HP)

To be brutally honest, that machine always scares me, ever since we first got it. I remain in what I consider a very healthy state of controlled fear every time I use it. I don’t use it unless I feel my best and have my wits completely about me, and I make sure that the other member of the household is aware of what I’m up to. It’s bigger than what we need, but beats the Chinese offerings we used to rely on.

I did actually have a go at the toggling the feed earlier to simulate what would be necessary for it to work, and despite having some lag in the gear train on the apron, with appropriate gearing on the feed and some fine tuning of the speed via the VFD, there would be a sweet spot where it would be possible.

Whether I would want to is another story. There is just too much risk for my liking. That sort of coil winding job is better suited to a Myford or similar size machine. In fact, I wouldn't want to consider it unless there was a real possibility of being able to physically stop the spindle in an emergency.

The lathe also features a hefty braking system, mechanically coupled to the control rod that stops the spindle dead every time the lever is pulled into the stop position. However, there is no footbrake, and the spindle stop does not stop quite so instantly at higher speeds. I fitted an extra emergency lever near to the chuck as a safety precaution, but if something goes wrong with a machine this size, I accept that I would be f*****.

Don’t worry – what I had in mind was hands free operation with properly designed and constructed accessories with loads of safety features.
Unless I was going to make a habit of winding coils, it’s easier just to buy them. The tooling necessary to do it properly and safely is just too much effort for the sake of a couple of weeks and the cost of setting that up. Although it’s good to consider what is possible “on paper” as a purely academic exercise.

I have heard about some of the horror stories, and had one or two close shaves myself before. Enough to drum it in to me that you don’t take chances with machine tools. I consider it in the same way I was trained with safety in chemistry, and having had just one or two very minor mishaps there was enough for me to reconsider my life choices.

you are more mad than me by a factor of 7.5 horse power to half a hordepower.

Isatrted 28 years ago winding coil on my Harrison L5. In those days all the retired engineers who set up sizeable organisations met at Billy Sweetings cellar to play.

They all said you cant do it Paul. You have to buy a winding machine . They were right. The laithes are too jerky and fine winding wire snaps.

I bought a winding machine and in comparison there was hardly ever a break, and it was much faster I think that one did 10,000 rpm.

I sold it stupidly . I think it was down to hard work life at the time, and thought Id never go back to winding. But now Ive retired andhave the time. But anyway I have a chinese made avo coil winder .

Something like this direction is what you should consider.

I admire you, But youre laithe is not the tool for winding coils.

[MistyBlue]

I’ve been busy……

I took some advice and decided to rip out the profiled acoustic foam from the cabs and replace it with an underlying layer of heavy carpet wool underlay. The profiled foam was then put back on top of the felt. It was a nightmare but was worth it for the increased damping and improved midrange clarity. Despite using some tooling including a wide chisel and right-angled short pry bar, the exercise took 2 days and did a good job of shredding my hands….

WoolFelt001.jpg (100.41 KiB) Viewed 55267 times

Also, the active crossover has now been fitted with amplification stages on the outputs to bring the level back up to eliminate the -14.8db problem with the headroom. I was worried about maintaining transparency, and was expecting to be disappointed, but I suppose the resulting sound is just different. Gains in some areas, but not so good in others. I would have preferred the sound to be a little more clinical. The vocals have improved, and the way the whole system now handles heavy rock. Not so sure about the simpler stuff though, which seems a little more closed in.

ModifiedActiveCrossover001.jpg (109.09 KiB) Viewed 55267 times

Whether this is to do with any voicing that has been done on the FE2022’s (they are meant to be front ends for MoFo’s etc), I don’t know. I have designed some through hole/DIP8 opamp boards that are being made by JLCPCB at the moment for experimenting, and I have also purchased some adapter boards for SMD devices so I can test different ones at leisure.

I imagine that I will drive us both nuts with swapping around!

I am also going to design a new linear supply for the crossover unit, so I can have split rails on the amplification stages using some super regulator boards I have got, and a separate single rail 24V supply for the crossover board.

[MistyBlue]

It's been a while since I posted anything on this, but while waiting for bits for my Mofo adventure, I got diverted to solving that -14.8db problem and get the crossover back up to line level. There was a simple and obvious answer, but I was a little anxious about the result, given past experiments with the FE2022. I needn’t have been. I swapped out the two FE2022 front ends on the HF outputs of the 6-24 active crossover board for a standard op amp board I had made up and added another one for the LF section. IC used was LM4652 at x5.7 gain.

I had to come up with a linear power supply to replace the 24V SMPS I had feeding the crossover PCB, and add another +/-15V supply for the op amp stages. I used one of those single channel “Studer 900” PCBs that are littered over on Ebay/Ali etc for the 24V supply and a Jan Didden Super Reg board to provide the dual supplies for the op amps. I recycled an old Lacie remote hard drive case for the PSU enclosure that I have had in the loft for years waiting for a project.

ActiveCrossoverNewOutputStages001.jpg (112.27 KiB) Viewed 47721 times

ActiveCrossoverNewPSU001.jpg (120.13 KiB) Viewed 47721 times

ActiveCrossoverNewPSU002.jpg (74.52 KiB) Viewed 47721 times

Sounds excellent. Everything has tightened up and is more detailed, but without being fatiguing. It’s almost as much of a positive change as I had when using the MoFo in the HF section. The bass has reverted to the tight, clean sound it originally had, which was my main concern, but I have been happily surprised by just how much of a difference it has made across the board.

[MistyBlue]

IC used was LM4652 at x5.7 gain.

Should read LM4562

[drivas]

I posted some of the following on DIYaudio as a homage to Lynn Olson’s “Beyond the Ariel” thread for the inspiration that it gave me, and I’ll repeat some of it here so you get the picture and where I’m coming from.

My current speakers were built in 2018 (30mm B/BB birch ply heavily braced and damped 230 litre cabs with Faital 15PR400 and HF146 HF drivers, crossover 3rd Order electrical at ~800Hz) which originally had JMLC425 horns. They are similar to the design that was discussed on that thread some time ago (15” Bass/Mid PA driver and 1.4” exit CD) with the GPA/Altec and Radian drivers, but with European drivers that are more easily obtainable in the UK and not to mention considerably cheaper. I played it safe with the choice of drivers, took some tips from others by trawling the various forums, got traces of measurements from multiple sources and simulated extensively before I took the plunge. When the cabinets were built, I took my own acoustic measurements and used VituixCAD to model the crossover. All the usual stuff.

They were a revelation to my ears, and I found them preferable to the Tannoy Windsors they replaced which sounded too overly refined and polite (being polite here - that old Linn advert comes to mind LOL). Both of us also prefer them to my previous Lowther based speakers and conventional cone/fabric dome hi-fi floorstanders.

speakers pic 99sml.jpg

In 2021 I changed the original JMLC fibreglass horns because they were just too forward for extended listening sessions and 3D printed some freestanding OSSE waveguides with the rounded edges to reduce diffraction. This improved things greatly with the HF, reducing colouration and they became much less forward in presentation and simplified the crossover. Much better.

However, since then both of us in the household have found that we still get listening fatigue, and whilst the sound is brilliant, it is just a bit too much excitement. They do all of the music we like excellently (Mark Knopfler, Norah Jones, Tracy Chapman, Pentangle, Roger Waters, First Aid Kit, Carcass (!) etc, they are detailed, they have that big sound that we like, they work well with films/games/DAW/TV etc as well as music, and they are absolutely mesmerising with Nightwish (Floor Jansen is a Goddess!). So there is much that is good about them, and we will never go back to the conventional HiFi stuff. But it’s just this one main issue. TBC

Actually, I'm not clear on which horns you're currently using. Are they the JMLC AH425, from Martin Seddon's Azurahorn.com, or are they your own design and build?

[MistyBlue]

The waveguides on the current speakers were designed using ATH - https://www.diyaudio.com/community/thre ... h4.338806/. I posted some of the construction details in this thread, when I was making them, here - https://audio-talk.co.uk/phpBB3/viewtop ... &start=195.

On the previous 2 way speakers, the original JMLC horns were indeed very close to the Azurahorn AH425. I made the mould on a lathe with the profile calculated from Hornresp software. The horns were layed up in fibreglass.

The replacement waveguides for the 2 way speakers were designed in ATH, had a similar mouth size to the JMLC, but were 3D printed.

Hope this helps.