Also check these MG15 pages:
http://www.troelsgravesen.dk/MG15XO-up-grade.htm
http://www.troelsgravesen.dk/MG15-problem.htm

A pair of dual concentric Tannoy drivers have been on my wish list for a long time. Try Google "tannoy monitor gold" and you will be flooded with web pages describing aural nirvana from these drivers. Read here: http://www.hilberink.nl/speaker.htm
I rush to tell that I did not become a Tannoy owner this time due to price. Something in the range of 2,000 USD is serious money and possibly the price is right, but this was not the time to invest in speakers that wouldn't be my main speakers anyway.

Fortunately we have a "Mister Tannoy" here in town and a call provided me with two 100 litres closed cabinets and 5 x 15" Monitor Gold drivers plus a box of crossovers. "Take your pick! And be sure that you will never hear what the Golds can do from these 100 litre cabs, they rather need 150-175 litres - or even better, 250 litres!" Initially I asked for some 12" drivers, HPD 315, but the face of "Mister Tannoy" went pale and his quick advise was to forget all about anything other than the Golds - and preferably the 15" drivers being superior to the 12" Golds!

250 litres is definitely into red WAF zone, but who knows. So, I took the whole lot home and started measuring the SPL response of the bass and compression driver without any crossover attached, trying to pair two drivers - because these 35-40 years old drivers do not necessarily measure the same!

Left: These 15" MGs are huge! Here on top of ScanSpeak SP38 speakers. Right: No, this is not how my test cabs looked, but surely some of these vintage cabs are beautiful.

Vintage drivers and what can be found:

Left: Apparently the HPDs are from a period where everything at Tannoy went wrong. The factory moved to Scotland with major start-up problems. The membranes got thinner in order to maintain efficiency from new and poorer magnets, the outer surround was made from rubber rather than corrugated, coated paper; the compression driver phase plugs were sometimes badly machined, etc., etc. All adding to poorer performance compared to the golden age of Tannoy, the Sixties and early Seventies. I found this picture on the web from a HPD driver, and yes, this phase plug is a really bad.

Middle: Actually one of the units had the compression driver voice coil stuck at one side in the magnet gab - and this driver has probably been delivering the "Tannoy sound" for decades with a major 10 dB dip around 2 kHz as can be seen below. This was how the pole piece (= phase plug) was centered:

Right: The aluminium dome of the MG15.

Above is seen the impact on frequency response from various settings of the "level" knob on the crossover and yes, if I put the "level" on maximum and "roll-off" on maximum, I could get a profile that started looking reasonably linear, but this was not how the MG was meant to be! This driver was over and out. So, even if you're buying MGs, be prepared for a bad day in the Sixties at the assembly line of Tannoy. I have to tell I once owned a British car build early Seventies, and it wasn't until I bought a Honda Civic that I found out that cars could start every morning for ten consecutive years! Hmm...
Hard to believe that this driver was not rejected in some quality check procedure. These drivers were big money even at that time. Eventually I found two drivers and two crossovers that seemed to work properly for initial listening tests. Crossovers? Yes, these were anything but consistent in performance. Here's a picture of the interior:

MG15 Crossover

Most of these components are of good quality and the reason for serious deviation from target is the electrolytic capacitor in the bass lowpass section. I measured from 13.3 to 24 uF (!) for the intended 16 uF. The lowpass section is simplicity itself: a choke and a capacitor (1.2 mH and 16 uF). 12 dB/octave and a point of crossover around 1200 Hz.

Late MG 15 crossover:

Left: Rear panel with terminals and rotary switches for adjusting tweeter level and tweeter roll-off.

In this late - and very well kept - version of the MG15 crossover, the electrolytic cap is replaced by a new capacitor, most likely made the same way as film caps, only here from etched foil based on physical dimension. Also for the tweeter section new - and better - caps have been used.
I can only admire the nice wiring job done here. All in best of order with twisted wires and secured leads. It takes time to make a crossover like this.

MG15 Crossover Up-Grade

Left: Click image to view large. Right: Up-grade kit finished. Click image to view large.

Measurements

Left: MG step response. As can be seen, the tweeter is connected with inverted polarity. Actually it's connected with positive polarity, but being mounted with the inverted dome facing the horn, the output comes out inverted.
Right: Above is seen a nearfield measurement of bass/mid and tweeter. Point of crossover appear to be around 1200-1300 Hz. Very steep slopes despite the 2nd order crossover. Based on medelling the sloped are close to LR4. A nearfield bass driver reading producing a flat response down to 50 Hz is not an everyday sight, I can tell. And a nearfield reading is highly dependent on where you actually place the microphone, so don't pay too much attention to dips and bumps. The Tannoys do not measure nice, but as seen from a lot of other big speakers, this does not seem to compromise sound. Another example of not putting too much into FR readings of loudspeakers. It tells very little of how they actually sound. So, with this fine-tuning I could start listening - and what a sound!

It's hard not to be completely swept off your feet when you first time connect these monster drivers. Every time I hook up a new pair of speakers I run a test CD to check left-right, polarity, etc., at it's not often you actually hear "something" from the 31 Hz warble tone on the test CD, but here it was.
I had never thought my Copland could provide the speed and transient attack as it appeared connected to these drivers. 90-92 dB sensitivity and 6-8 ohms impedance is yummy to most 2 x 50 W valve power amps - and the Copland loved it.
I have never been able to hear female vocalists this loud without serious distress to the ear (= distortion) and I've never heard applause from an audience from live recordings this clear. Acoustic bass starts getting the right proportions and a full drum kit from these speakers is aural nirvana. No doubt about it.

That's not to say the MGs like to be played loud, they are very good too at low volumes and don't need a whole lot of power to be "energized".
In my listening room I have the speakers on the long side. Usually 1 metre from the rear wall and with 1.8 metres to the sidewalls. Use to work fine and so did the Tannoys - except that they simply appear to be too large for listening at 2-2½ metres distance. The very first thing that came to mind when I played music on the Tannoys was my earphones. Listening nearfield to these drivers is like putting on a giant pair of earphones, except that you now can actually feel the bass. I'm certain the earphone association came from the coherence of the sound coming from these drivers. This dual-concentric principle is a little wonder but I have always been suspicious about the integration of a 90 grams heavy paper cone and a "no-weight", horn-loaded compression driver. How can this go well? But it mostly does here. I guess that the large radiating area of the horn driver - actually of exactly the same size as the paper cone, as the paper cone is the horn - is part of the explanation why it performs so well.

This does not imply that the Tannoys do not have a "sound" - they very much do so. When it comes to midrange transparency and neutrality they are somewhat behind the Acapella SE set-up. The Acapella SE has a much more neutral midrange presentation and much better pin-point imaging compared to the MGs.
Pin-point imaging: Well, we would think that the dual-concentric principle would make a further contribution to the overall coherence and seamless integration of midrange and treble, enhancing pin-point imaging, but this was not my experience. Sitting at 2½ metres distance I sometimes found it hard to render the same ease of localisation of vocals and instruments in the acoustic scenario. These membranes are so huge that vocals seemed to come from all over the place. I've experienced the same thing with panel speakers and I guess that's part of the reason Peter Walker added the delay lines to the famous QUAD ESL63. To get better dispersion and to target a point source radiator. So, we're back to the discussion from the "Infinite Baffle Speaker" file on pinpoint imaging, and again: Paul Messenger:
"Narrow speakers with a very wide radiation include more of the listening room and help create an illusion of bringing the musicians into the room; while more directional designs like horns and dipole panels give a precise view onto the recording itself.
The Tannoys are to my experience quite directional speakers. I have measured the horizontal dispersion - no need to do a vertical dispersion here as it will be the same due to the dual-concentric principle - and it is very good. You measure the same FR all over the place, but when listening in stereo, they are quite directional and you move the speakers around and you move your head around to find the best position for stereo integration. These speakers need large rooms and a listening distance of 4-5 meters to perform the best. Not my cup of tea for near-field monitoring. I guess the 12" or 10" would be better suited for this situation but with sacrifice of deep bass.

Monitor Gold Type LSU/HF/15/8

New crossover

From the frd and zrd files of bass and tweeter I modelled a new crossover trying to get as close as possible to the 1.2 mH and 16 uF used in the original low-pass section. This persistently came closer to a point of crossover around 2 kHz against the claimed 1 kHz point of crossover. Measuring all the drivers - with original crossover - there appear to be a major overlap in the crossover region and no matter what polarity was given to the tweeter, the nulling effect wasn't visible at all. The phase-tracking between these two drivers is indeed very poor.
The weird thing about the MG is that the coil in the LP section can be varied a lot without affecting the midrange much, this partly due to the impedance profile. And this driver is so big that measuring the FR at 1 metre distance is really too short. We're getting into near-field conditions and these drivers should possibly be measured at two metres distance, but my workshop won't allow this and I didn't have time to take the drivers outside for proper measurements.
However, with 1.2 mH and 16.4 uF+2R2 a reasonably flat response could be achieved with a point of crossover around 2 kHz. In fact the same I measured from some of the drivers fitted with the original crossover. The compression driver however, seems to target a point of crossover around 1.2 kHz but due to the drivers being out of phase the overall result is a reasonably flat response when the "level" and "roll-off" setting is used properly.

Left: The measured FR is close to the predicted. Disregard actual dB level. This reading was not generated with 2.8 V input.
Right: MG/original crossover (red) compared to Acapella SE (blue, only midrange+ribbon).

The frequency response (FR) of these Tannoys looks awful! Why is it this doesn't leave this driver useless? I've seen FR from electrostatic speakers that look even worse and yet these speakers sometimes deliver the best of sounds. After delivering back the Tannoys I connected the SP38s again and yes, these speakers are much more well-balanced, the dispersion is much better - it's simply a more truthful presenter - but it can never deliver the bass punch and hard beat of the snare drum that hits you like a bullit from the MGs.

So, what's to be learned from these Monitor Golds? Well, sensitivity is an important parameter and they tell you that most of today's speakers really do not have the transient response capability of these ancient drivers. I read that a lot of enthusiasts using single ended amplifiers will use high sensitivity speakers in order to cope with the sometimes very few watts available from these amps. We see 15" coax Fane drivers used with 5 watts SE amps, but 15" Fanes are meant for PA application and not for hifi. Aren't we getting back to some JBL sound - for better or for worse - in order to provide the necessary sensitivity? Where is the smooth, well-balanced sound from high sensitivity speakers? Where are the high sensitivity speakers that will cope will all sorts of music and not just trio jazz?

I tried playing some of the old Beatles LPs with the Tannoys - the very same speakers that were used for mixing at Abbey Road - and it really didn't sound too good. I wonder if this was what George Martin was hearing at the mixing desk.
These medium-sensitivity Tannoys come close in being able to handle almost anything. I don't find them particularly transparent but they will for sure tell you that there are a lot of things your current speakers probably are short of.

I cannot thank "Mister Tannoy" enough for lending me his drivers!

Comments from UK reader, who wants to stay anonymous:

Hello Troels, I enjoyed your story about the Monitor Golds very much. You ask yourself at one time after playing Beatles recordings: Is this what George Martin heard when he was doing the recording? Taken into account the many variables like room acoustics, amplifiers etc. It is save to say he probably did.
I used to work with Tannoy 15" units placed in large Lockwood cabinets at Polygram. These units were used in many studios around the world during the sixties and seventies. The sound was not spectacular HiFi nor very detailed. That is not what a recording engineer is looking for in the first place when selecting a professional monitor system. These Tannoys deliver a reliable sound pattern that allows engineers and producers to balance the sound and use the large array of filters and outboard equipment to do a mixdown of a multitack recording. In studios the HF horn was most likely to die long before problems with the LF unit would arise.
After we moved to the newly built Wisseloord studios in 1977 the Tannoys were abandoned. At Wisseloord a completely different system with JBL units was used that gave a far more detailed sound and better defined spacing of the instruments. This system was part of the acoustically designed control rooms.
To double check the voice/orchestra balance we often used  the Philips 9710M units in a relatively small enclosure. Out of large quantities  9710M frames selected units were used by Dutch radio to serve as monitor speakers in broadcast studios. I had the pleasure of meeting  Mr. Koppinga the designer of the 9710M during the seventies. In Holland there is still a strong demand for these wideband units after nearly 50 years since their introduction.
With kindest regards.