Wilmslow Classic Floorstanders Project

So, my last post was three years ago, and the Dynaudio TL project in my avatar is long gone. For the last couple of years I've been listening to a pair of Mirage M3Si's I rescued from a skip and rebuilt, and slowly putting together a front end to feed them. The Mirages are now gone as my good lady thought they were too big for the current house (which they absolutely were) and we're aiming to downsize.

My DIY skills are enthusiastic and knowledgeable but lack any real ability, so I’m keeping myself sensible as to what I take on. My mantra for the whole system has been “Grand Enough for Spotify”. Every time I think about spending money on “Audiophile” stuff, I go back to the Mantra… the “Grand” is the nominal budget for the whole system. Source material is Spotify Connect only. I stopped kidding myself I needed any other source when I realised I couldn't hear anything over 11kHz anyway. :rolleyes:

Front end consists of a Raspberry Pi (3B+) with 7”screen and a HifiBerry Digi2 Pro HAT, running Volumio. The optical output from the HAT goes via an old Musical Fidelity V-DAC and a passive volume control to a power amp using Icepower 300AS1 and 300A1 modules. Total cost of all that is about £500, leaving something similar for speakers. Now, if I’m honest, I should buy myself a pair of half decent standmounts and live with them, maybe with the BK sub I still need to fix, but that doesn’t scratch my itch to spend money unwisely on something unsaleable that I “built myself”.

I've been hunting for something smaller than the Mirages while remaining interesting, and up pops a pair of Wilmslow Classic for sale. I decided they were what I wanted and organised a long weekend with the Mrs at the wrong end of the country to collect them. What are they? A Volt BM220.8 mid/woofer and a Scanspeak D2905/950000 in a very solid, 40l floor-standing tower. Similar to the current SM108 kit from Wilmslow, but in a taller and larger box. I was surprised when I heard them (quietly) in the seller's flat. They sounded like a mini-monitor and had no bass worth mentioning. Of course, I bought them anyway and brought them home - I like a challenge!

Got them home and turned them up, hoping a bit of volume might help, but they only got worse… Playing quietly, they were detailed, clean but somewhat uninteresting due to lack of bass. Turned up they just got harsh, sibilant and generally unpleasant. Without measurements, they seemed to have far too much output in the crossover and lower tweeter range.

I pulled them apart to investigate…

IMG_0677.jpg
 
Last edited by a moderator:
So what do we have…

The boxes are very presentable, nicely veneered and absolutely solid. A quick calc of the port tuning suggests they are tuned to around 34Hz. The Volt literature claims it works in volumes from 20 to 50l, so should be no problems there. I'll point out at this stage that I paid about the same for the speakers as Wilmslow add to the SM108 kit price for an mdf flatpack…

First thing I noticed is that there’s a sheet of white acoustic stuffing wadded up behind the bass driver and the port. Not right intuitively, and not where the Wilmslow instructions tell you to put it. Yes, I got the original build instructions and invoice with it (£759 for the kit in 1999). Pulling that out added some bass but obviously does nothing for the crossover region.

Drivers are in visually good condition. All four are working. I can only assume the Volts are good as new. If weight is anything to go by, they are serious items at about 3.5kg each. The tweeters looked fine but a clean out revealed slightly damp and dirty coil gaps and no actual ferrofluid… Gave them a good clean out and popped in some new fluid I happened to have left over from the Mirage build.

Internal wiring is interesting. Van Den Hul (CS16?) on the LF circuit. Thinnest unmarked wire on the HF. I’m sure it was done for a reason, but no idea what it was!

Which leaves me with what should be correctly working drivers in each speaker and the crossovers…
 

Attachments

  • Opened.jpg
    Opened.jpg
    536.1 KB · Views: 93
  • Tweeters in bits.jpg
    Tweeters in bits.jpg
    326.7 KB · Views: 94
  • Invoice 2.jpg
    Invoice 2.jpg
    351.3 KB · Views: 98
Ah, the crossovers. These were, in theory at least, supplied from Wilmslow ready-built and with a “Hovland Musicap” upgrade. They contain these expensive film caps and air cored inductors on a circuit board. I’ve chased it through, measured values where possible and mapped out the circuit. It didn’t really make sense to me. I’ve been attempting to get Wilmslow to send me a circuit diagram, or at least confirm that what I had is what they intended, but I’m getting no joy… A) “they’ve been archived” (lost?) and B) “we don’t give out circuit diagrams willy-nilly”. I’ve sent my layout of the circuit to them for checking, but I’m not holding my breath.Their customer service technique appears to be ignore you and hope you go away.

I modelled the crossover in X-SIM (shown). The tweeter circuit had an inductor and resistor in parallel, both in series with the tweeter. This after a third order high-pass, so what LF the inductor had left to pass through, and why you’d want to, I don’t know.

In the absence of any info, and with dissatisfaction in the way they sounded, I assumed they were wrong and that the inductor should be across the tweeter to earth to complete a 4th order filter. Doing this in X-Sim drops the output in the 3-5kHz region without affecting much else, so I got the soldering iron out and moved the inductor output to the earth line to try it.

Bingo! Sibilance all but disappeared and they now sound nicely balanced; still detailed, but much smoother without the harshness that was there before. Still no idea if this is what Wilmslow intended, or how they measure, but they sound perfect to my ears, so they’re staying that way. Next step is to reassemble them with the crossovers back inside, clean up the wiring etc. and get on with enjoying the music.
 

Attachments

  • Crossover in box.jpg
    Crossover in box.jpg
    343.3 KB · Views: 113
  • Crossover.jpg
    Crossover.jpg
    764.5 KB · Views: 105
Last edited:
  • Like
Reactions: 1 user
That Volt driver has a (very) low Qts of 0.27 and a Vas is 47L.

This means that in a 35-40L cab you will get a classic Extended Bass Shelf (EBS) alignment so low frequencies will drop by about 6dB below 100Hz or so, stay at that level until they drop at 24dB/0ct according to port tuning (should be driver Fs).
This can be relatively easily fixed either by placement closer to boundaries or the 100Hz tone control amps and pre-amps used to have.
 
  • Like
Reactions: 1 user
Thanks. Don't have any tone controls, but they definitely have more bass closer to the wall. Difficult at the moment as the crossovers are still on the floor behind them!

Tonight's job is to replace the power caps in my BK subwoofer, which should comprehensively sort out any remaining issues below 100Hz. :cool:

HifiBerry do a DSP card that adds to my Pi setup which will also sort the problem out. I'll probably do this next and lose the sub.
 
  • Like
Reactions: 1 user
Reading your old post again, sounds like the Kord had problems in the crossover as well...

The 2mH inductor rolls the Volt off from low... maybe below 1kHz... it seems the design uses the sharp mechanical roll off of the driver at 3kHz at the crossover frequency, which seems to be up near there. Unfortunately, I suspect the low roll off sucks the midrange out. At one point I was running the tweeters through the (original) crossover and running the Volts full range... it sounded better to me. I'll have a play at reducing the roll off in the LF circuit in X-Sim and see what happens.
 
Last edited:
First thing I noticed is that there’s a sheet of white acoustic stuffing wadded up behind the bass driver and the port. Not right intuitively, and not where the Wilmslow instructions tell you to put it. Yes, I got the original build instructions and invoice with it (£759 for the kit in 1999). Pulling that out added some bass but obviously does nothing for the crossover region.

If I understood your description it seemed like the correct place for wadding in a ported speaker. I suggest having a browse of the site to get a feel for how much faith to put in instructions from Wilmslow Audio (though this has varied over time as personnel changed). The wadding in a ported cabinet is seeking to do several things in a 2 way: damp internal cavity resonances at midrange frequencies (wadding wants to be where the particle velocity is high for the relevant modes which tends to be away from walls (zero particle velocity) and towards the centre of the cavity) and to prevent the midrange frequencies leaving via the port (wadding wants to be between the driver and the port but without getting too close to the port so that the slug of resonant air can enter and leave with little restriction). Ported cabinets don't require large amounts of wadding (lots of wadding like a sealed speaker is usually undesirable) but it does need to be in the right place to do it's job effectively.

Drivers are in visually good condition. All four are working. I can only assume the Volts are good as new.

This is not a wise assumption for old drivers whatever they may look like. For example, the mechanical properties of the suspension and surround usually change significantly with time.

Which leaves me with what should be correctly working drivers in each speaker and the crossovers…

Hmmm... A 2 way involves a lot of compromises particularly in the crossover region. It is going to be nigh on impossible to do a reasonable job particularly w.r.t. to identifying issues and sorting out the crossover without basing one's decisions on reliable information such as that provided by measurement.
 
I had a bit of a brainwave this morning. In the absence of the ability to measure the speakers, I can add some EQ until they sound good, and work out what's wrong in the frequency response.
Connected an android tablet as source, then played with the EQ in Spotify's internal equaliser. After much fiddling the problem is simple... reducing everything above 4k by 4dB (and 2k by 2dB) sorts them out. Found some old 22k resistors and fitted them in parallel with the tweeters (and removed the strange inductors) which (according to XSim) should drop the tweeters by about 2dB and immediately they are much more listenable if still a bit bright.
I've ordered a selection of resistors to play with the L pad values, but I now have an end in sight!
 
20231015_111609.jpg


After playing a bit, I deconstructed the old Wilmslow crossovers and used the bits with one new cap in the LF circuit and some different resistors to build my own. Point to point wiring, inductors all at right angles, completely separate boards for LF and HF, and fitted them in the bottom of the cabinets away from the drivers.
I've settled on a circuit with the odd series inductors removed from the HF (still there in the pic, no longer connected), tweeters reversed polarity and padded down about 3dB which sounded right to me.
Had a go at measuring the response this pm. Played white noise through them from an app on my tablet and picked it up on my phone with a spectrum analyser app... Not exactly accurate, but they do look reasonably flat.
More importantly, they're starting to sound good... I've now stopped cringing and started enjoying the music.
 
Yeah, I know my soldering isn't pretty - I'm learning!

As an aside I seem, entirely unintentionally but with absolutely no regret, to have removed the Wilmslow name from these speakers... it was on the "Wilmslow Supersound" caps that I replaced and in the "WA" markings on the circuit boards. Seems fitting given how much help they've given me (still none) and the fact they stole/copied/licenced the design from Kord, Volt or Solen (not sure which came first) and I've modified it anyway.

I'll have to come up with a new name for them now :unsure: