While sourcing the LS5/8 loudspeakers that formed the basis of Paul's system, we learned that some of the last LS5/8 systems made used Chord amplification. But until a system appeared on eBay recently, I'd never seen one. And by coincidence, while drafting this page a colleague found an old copy of ENG INF (an in-house publication for BBC engineering, technical, and operational staff) which contained details of this system. See below for scans of the relevant pages.
Normally, I wouldn't devote a page to a speaker that I've not had the chance to listen to, but these are so rare that they deserve to be documented. And having heard the "standard" LS5/8s so frequently, I feel that I'm almost qualified to write about these. But, please don't email me to ask what I think of the sound of the Chord LS5/8 because, again, I haven't heard them!
I'm indebted to Mike at the Audio Toy Shop, whose eBay name is "atoyboy", for allowing me to use the following pictures. And should the new owner of these want to contribute to this page, perhaps by sending me their impressions of them or maybe some more pictures, then I would be very grateful.
Overview
I'm assuming you're familiar with the standard BBC LS5/8, which shoehorned an active crossover PCB into a Quad 405. The Chord versions were produced because of the non-availability of the 405, and were chosen after listening tests detailed in the ENG INF article.
Here is a view of the Chord amps and the LS5/8s.
Apart from the addition of a Neutrik speakON connector on the rear panel, the loudspeakers themselves are entirely standard.
Here we see the amplifiers and their speakON interconnects.
As you can see, the modified Chord SPM800s have been designated "AM8/20".
Inside the Chords
These pictures seen so far were included in the eBay listing, and were tantalising enough. But the amps were what really interested me, and lots of questions needed answering, including what crossover PCBs were being used, and was the circuit redesigned in any way? Fortunately, Mike was happy to supply some internal shots.
Chord amplifiers are famous for using switched-mode power supplies, and this is no exception. The front half of the amplifier is given over to the power supply, and as you can see, the internal partitions appear to isolate this PSU from the rest of the amplifier. The outputs from the supply are further supported by the 6 electrolytic capacitors you can see in the centre-rear of the chassis.
The audio circuitry is confined to the lower section of the case:
And here you can see that the BBC crossover is a standard late-version PCB. For reference, the two different versions are shown below - the early crossover (left picture) is what we cloned for Paul's LS5/8s.
)
)
While later versions have some mechanical differences, as far as I can determine they contain the same circuitry with minor component value changes. The main change is the bass-boost switch; on early versions, an unusual 3-position toggle switch is used (blue square, bottom right) - this switch is in fact a 2-pole, 3 position switch (normal 3-position toggle switches have a centre-off position). On later boards, a 3 position slide switch is used, mounted onto the metal bracket that supports the crossover PCB in the Quad 405.
As you can see, the Chord has more internal space than the Quad 405, which makes installation of the crossover PCB somewhat more straightforward. Here we can see the PCB and wiring:
The volume control has been mounted off the PCB on the rear panel, and it appears that the bass-boost switch has been abandoned - it is hard-wired to the "flat" position. Very sensible - the last thing these loudspeakers need is bass-boost!
And this close-up of the crossover PCB shows a few small changes:
The resistor in series with the voltage regulator transistor is a different value, and a 2.5W vitreous enamel wire-wound device. Standard crossovers use a 1/4 watt metal film here, although I normally upgrade this to 1W as a matter of course. Of course, the Chord uses higher supply rails than the Quad, so this is no surprise. Note the red wire bringing in the +HT from the adjacent PCB that carries the output transistors.
A much larger input transformer is used, mounted on the rear of the PCB. The ENG INF report mentions that the original transformer caused some HF rolloff, which is exactly what I found when cloning the amps for Paul. We used different transformers to overcome this, and I'm glad to see that this is also the case in these models. Indeed, the type chosen by Chord is the same as I picked.
Finally, a couple of views of the switched-mode power supply:
Note the large input filters - the two torroidal cores, top of frame, which look as if they might be common-mode chokes - and the black thermistor (next to the 5A DC fuse) used to limit the inrush current. The bridge rectifier is mounted on the side of the chassis for cooling. There is a small conventional (50Hz) transformer nearby to power the control circuitry. This is good, as all attempts to design-out this component usually reduces reliability (just ask any service engineer!).
It's impossible to guess without access to schematic diagrams, but it could be a half-bridge topology judging by the two switching transistors and noting that the collector/emitter junction of these switches are AC-coupled to the transformer via the large orange capacitor. I would imagine that a forward or flyback converter would be inappropriate here because of the large power requirements of the output stage (which is rated at 160 watts per channel into 8 ohms).
The small vertically-mounted PCB contains the control circuitry - note that there are two pulse transformers (beige rectangles, to the left of the main switching transformer) to carry the base drive waveforms across the isolation barrier to the switching transistors. Given the age of the unit, I guess these are bipolar, but it's possible that they could be MOSFETs.
If anyone can shed any more light on these fascinating amplifiers, I'd be very interested.
BBC ENG INF (Winter 1992/3, No 51) report
As mentioned above, discovering this article at the same time as I was writing this article was total serendipity! I've scanned the relevant pages - click on them to enlarge, or use the the OCR'd version of the article.
)
)
)
Please note: The copyright of this article belongs to the BBC, and while I believe that the inclusion here constitutes fair usage given the historical value, please be aware that it may be removed at any time.
Reading this article, it seems that someone got a real bargain. These fetched £1120 on eBay, and I'm sure that there were worth every penny! As I said above, I would be delighted to hear from the purchaser of these. I would also love to learn more about the Chord SPM800.
Update: The AM8/21
Amazingly, there was another iteration of this concept - Chord managed to fit a stereo version of the LS5/8 amplifier into one chassis! This was designated AM8/21.
This was announced in the quarterly in-house magazine BBC ENG INF in spring 1993 - I don't have a paper copy to scan, but thanks to the efforts of BBCeng.info, you can view these publications on-line. Look for "STOP PRESS..." on page 2 of Issue 52. Yes, that's just after issue 51, cited above, so Chord weren't exactly resting on their laurels! I have seen so few of these that I can't comment about which version was more widespread; these were only ever made in small numbers anyway. Any data would be very welcome.
Of course, this requires the original amplifier to have 4 channels of amplification, so what was it? Two obvious answers spring to mind: perhaps the original amplifier was 4 channel originally or, perhaps more likely, it was a stereo unit that was bridged to reach the required output power - a not uncommon technique when very high powers are needed. If anyone knows anything about these, please let me know.
I haven't yet found the emails in my archive, which were kindly submitted for publication a few years back - but as soon as I do, I will include an appropriate credit for the photographs. In the meantime, if you recognise these, please do get in touch.
I'm no expert on Chord amplifiers, but this looks like one of their professional "SPA" series units.
This rear view shows the XLR inputs and the Neutrik loudspeaker outputs. There are blanking plugs covering holes left by phono sockets and conventional 4mm binding posts.
This internal picture shows relatively little of the "host" amplifier, but look at the crossover PCB. It's a new design - both channels on the same PCB. Let's take a closer look...
And what do we see? Op-amps! This is a completely new design. I wonder how this came about, and I wonder what involvement the BBC had with this, if any? Absolutely fascinating...
Sadly, like the AM8/20, I have never seen one of these in the flesh, let alone heard one, but if anyone can offer any more information, I'd be delighted to hear from you.
Update - December 2019
With thanks to Richard Jeffrey-Gray at Hoxa, I have learnt that the the op-amp version of the crossover was also used in the AM8/20.
This is something that occurred to me when I originally learnt about the AM8/21, but it's really good to finally have evidence of this.
Because there are so few of these units out there, it is difficult to establish the exact evolution of the Chord amplifiers. However, we know that the original crossover PCB - albeit with slight modifications - was used in the earliest of the Chord AM8/20 amplifiers. And, we know that a "stereo" version of the amplifier was released that used an op-amp based circuit, as seen above.
At this time, I still don't know who designed the op-amp version. If pushed, I'd guess that Chord did the design work, and would assume that they got BBC approval for the circuit, but I've no evidence to support this yet. However, when looking at the AM8/21, you can see why the design came about: the original BBC design based around discrete transistors would have cost more to produce than the op-amp version, and it was physically larger. Hence, for the dual AM8/21, it makes complete sense - trying to shoe-horn a pair of the original transistor-based PCBs in there would have been challenging, to say the least.
Now, having done the design work for the AM8/21, you have to question the logic of sticking with the original transistor design in the AM8/20. Remember; the mono AM8/20 would have been offered alongside the stereo AM8/21, as in many instances the "monoblock under the cabinet" arrangement simply made more sense than a common stereo amplifier.
So this explains why there is a version of the AM8/20 with the op-amp based crossover. And here's what it looks like:
It's not exactly the same layout as the dual version in the AM8/21 - it's not as if someone just took a hacksaw to that PCB - but there are obvious similarities. Of course, the transformer is mounted on the component side on this assembly, instead of on the solder side on the AM8/21 version, but note that the 4 mounting screws are in the same place. Note also the missing resistors at the top - in theory there should be 3 resistors in parallel to drop the 70-to-80 volt rails down to 15V, but the only one fitted is 56Ω - which suggests that the main resistance is elsewhere. There are 7 op-amp sections (three NE5534s and two NE5532s), and each of those take 4mA - so 28mA in total. Add a bit for the 15V zeners and call it 35mA in total. A suitable resistor for dropping from 70V down to 15V while passing 35mA might be 1.5k, and that'll dissipate 2 watts. That's a lot of heat to put so close to those big electrolytics.
While not apparent in this photo, there are quite a few differences apparent in the "host" amplifier. I suspect that Chord made changes to the design of the SPM 800, but I know almost nothing about the internals of Chord products, and nor does the rest of the internet - they appear to have done a very good job of keeping their products well wrapped up!