Third Generation CMF Amplifiers
Third Generation is a brand name for Adam Hall, well known suppliers of studio and stage hardware, from rack-mount strips to plugboards. Sherlock can't tell you exactly what 'CMF' stands for, but the last two letters in the name undoubtedly refer to the MOSFET devices found inside these amplifiers.
Taking 8 ohms as the norm impedance for monitors, the three amps we're looking at this month (CMF 200, 400 and 1000) are broadly rated at around 60, 150 and 350 watts respectively, per channel. There's also a power-bridging facility, so high powers are available, in mono only, of course. Table 1 gives the quoted output powers into other impedances, while Table 2 illustrates actual test results, and the sensitivity into various impedances of the middle model, the CMF 400; indeed, with all three amps being similar, we've concentrated on this, the 150 watt model, because it's the one most likely to find its way into studios. Differences in the CMF 400's brethren (where they exist) will be briefly assessed later on.
Judging by the full complement of XLR connectors on the front panel, and the literature, the CMF series amps have been designed with the PA user in mind: for each input, there's a 'link out' female, and a pair of paralleled output XLRs. "This is all very nice" you are thinking, "but I don't want messy patch cords all over the front of my rack". But don't jump to conclusions: there are also terminations at the rear, and these do suit studio requirements, as we'll see later.
Homing in on specifics, the inputs are unbalanced, and the XLRs are wired to the American Standard, that's with pin 3 hot, and pin 2 disconnected (or grounded). If you drive this amp from a balanced feed, all will be well, except that the absolute phase (or signal polarity) will be inverted. This is usually inaudible. Most home studio feeds will be unbalanced, and 99% of unbalanced XLR interconnections in the UK operate with pin 2 hot - the European Standard - so assuming Third Generation don't modify their pinouts to meet the needs of their UK users - and we hope they will - then there are two options.
One, if you're wiring up fresh input cables for this amp, simply connect the male line XLR's 'hot' wire (inner) to pin 3 instead of pin 2 as usual. Then wrap red tape around the XLR and write "3" on the tape to remind you that the 'hot' on this cable is hooked up to pin 3, ie. non-standard. Alternatively, unscrew the front panel, fold down and move the input XLR hot wire from pin 3 over to pin 2. By the way, gaining internal access to the XLRs on this amp is considerably easier than on most.
The impression that "there's a lot going on" across the CMF 400's front panel is added to by the presence of two power switches, no less; each channel can be activated independently. We'll go into this later, but for now we'll content ourselves with the observation that a pair of independent switches need handle only half the turn-on current surge in contrast to a single switch generally used to power-up the entire amp.
It's therefore in order to predict a longer than average lifespan for the CMF 400's power switches, an important factor in view of the predilection of toroidal transformers to pull 100 amps or more through the mains at switch on.
The upper half of the front panel is recessed, and the panel artwork and VU meter bodies are overlaid with back-printed perspex which prompts a quick aesthetic judgement. Finished in 'Belgrano' grey with cream legend, the recessed panel is stylish and yet practical too, in that it's easy to keep clean and looks tasty. The surrounding metalwork, with smooth, radiused edges and metallic bronze paintwork is also worthy of praise.
The moral here - and one which many manufacturers are slow to grasp - is that amps no longer have to be rigged out in black with square edged metalwork to look sexy, and people as a whole, moreover, may no longer want these sort of aesthetics. Black is after all strongly associated with death, gloom and ill-doings; fine for Heavy Metal musos, but hardly an encouraging colour for the rest of us. The colour scheme adopted by Third Generation may, therefore, be seen as innovative and (literally) symbolic of the third generation of transistor power amps. Only the rivets serve to spoil the good impression. Used to mount the XLRs, they not only look ugly, but make replacement very awkward if the connectors are damaged or fall apart.
At one inch diameter the gain control knobs are larger than average, so whilst the scale is of the usual 1 to 10 sort, it's easy to interpolate any settings between the numerals. This mitigates for the gain controls being ordinary rotary types, as opposed to stepped attenuators and though the tactile qualities of these pots are good, for PA use I would prefer more rugged controls.
Though the output metering lights up, it's sadly in the form of VU swing needles. Just to remind you, VU (Volume Unit) metering is useful solely for assessing the average level and perceived loudness of a music signal, and these meters are next to useless should we want to see the onset of clipping at all accurately. To wit, these meters will read 8 to 15dB or more in error on music, with short-term transients exceeding the mean (VU) level by this amount. Of course, the metering is handy for basic PA work, because it confirms that "watts are coming out the back", but for studio and professional PA usage, we feel that the Third Generation amps would benefit from accurate, peak-sensing overload metering, namely a red LED!
As I implied earlier, the input and output terminations are repeated on the rear panel. Normally, the shocking expense of XLRs rules out this sort of generosity, but the Third Generation amps are fitted with Key Technology's new low-cost, glass-fibre types. Of course, there's a snag: these XLRs aren't as rugged as the original Switchcraft or Cannon types, but still, an XLR with a reduced specification can more than hold its own against other, competing audio connectors.
The XLR outputs are different. First to put your mind at rest, they're wired pin 2 hot, in the conventional fashion. Second they're not repeated on the rear panel; instead, there are 4mm sockets-cum-binding posts. Though less convenient for PA work, connections made with binding posts are on a par with XLRs, at least from the viewpoint of long-term reliability, so no worries here. Besides, because power-bridging depends on us making the connection between the hot side of each channel (the 'negative' or 'cold' outputs on each channel are left unused), binding posts are more convenient when the bridge-mode is in use. Why? Well because if you power-bridge with XLRs, you have to make up a special lead, which splits out the two speaker wire to separate XLR plugs. Messy!
Power-bridging on these amps is by means of a slide switch; a locking plate prevents it being moved around by little blue gremlins, an important feature for road work. When bridged, the left hand gain control is inoperative: mono only, remember, in this mode. In line with the independent power switches, each channel sports independent mains fusing. Short of the main breaker tripping out, this means a catastrophic failure in one channel can't disrupt the other, a useful feature, at least for PA.
Mains comes in via the usual IEC ('Euroconnector') socket, but beware of the adjacent mains voltage selector panel - it isn't connected up, presumably because the transformers used have no tapping facilities. However, Third Generation may incorporate this feature on full production models: and going by the selector panel, taps for 220V (handy for Continental Europe) and 100/120V would be available. The 200V tap would be of particular interest to musicians having connections with mainland Europe.
An excellent feature on this amp is the grounding strap (see 'Studio Earthing Techniques' Sept 84 HSR) which allows us to isolate the chassis from the signal ground. You do this by unscrewing a looped wire from a barrier strip. If you find your set-up hums least with this link removed, best gaffa-tape it inside the lid, so it's not lost. With the 0 volts and 'earth' sides being clearly marked, the barrier strip acts as a handy test point, and also a strapping post for any stray green wires inside the rack! The rear panel is generously labelled: key parameters (ie. normal input sensitivity for full output, minimum load impedances, and XLR pinouts) are all given; hands up if you've never mislaid the instruction manual, or vital data on a piece of gear?
The CMF 400's enclosure consists of 10 gauge aluminium (that's just under ⅛" thick), bent to form the back and sides, while the front panel is formed into a convoluted tray shape (hence the recessed perspex) to provide strength, and the rack-mount flanges are welded to the front panel. All in all, the metalwork is above average, and also innovative and intelligent.
In particular, the amp is no heavier than it need be, and the side members project just over an inch (30mm) beyond the back panel. Both are handy for road work, and the projection shields the connectors from 'Ye Bigge Boot' at no extra expense, and doesn't make this amplifier take up any unnecessary space; for when hooked up, the plugs will need a couple of inches in any event.
Less favourably, the raw mechanics would need more attention in order to cope with professional PA work - in particular, the torque available on the screws fixing the panel is below par, as is the overall rigidity. Both of these shortcomings are easily put right, though - are you reading this at Southend?
Next on the list are the handles, and indeed virtually all of the fixings: no attempt has been made to lock these and some of the screws had slackened after a (presumably) unluxurious journey to my home via Securicor's 'A' Service! Competition: how many screws will there be left after 3 rehearsals, 6 gigs, and 300 miles in the back of a Transit? However, if you're solely thinking of making these amps enjoy a comfortable, control room existence, then none of this need worry you.
The heatsinking is more perilous. Heat conducts from the MOSFETs to the external heatsinks and the chassis, via an 'L'-shaped bracket. But the bracket isn't flat enough, and only half the surface area on the amp we examined contacts the chassis. In effect, we have a thermal bottleneck, therefore the external heatsinking is mostly wasted. The remedy - a dead flat, machined surface - should be very easy to implement though.
Taking a peek inside, each amplifier channel employs a wholly separate power supply as inferred earlier. Another unusual feature is the incorporation of two thirds of the power supply components on the respective amplifier PCB, leaving just the two toroidal transformers attached to the chassis base.
Dealing with the PCB first, the concept of incorporating the rectifier diodes on the amp's card was first seen about 12 years ago, when Jim Marshall made power amps, and has resurfaced from time to time, but has recently become more prevalent as designers have come to appreciate (i) the cost savings and (ii) the role of the power supply vis-a-vis further improvements to the amplifier sound.
Broadly speaking, the closer you can position the power supply to the output devices (the MOSFETs), the better the performance, especially at high frequencies. In particular, 'on board' supplies make the routing of the 0 volts wires less critical. The difference may be lost on stage, but it's certainly relevant to serious monitoring applications. The dual power supplies also contribute to a better stereo image. The actual circuitry is wholly conventional, being based on the famous Hitachi MOSFET application note; and the even earlier LH 0001 hybrid driver-amp from National Semiconductor! Its keynote is achieving excellent results with only 50 or so components, where other designs would use 100 parts or more.
There are no points for the reservoir capacitor mounting: these are soldered to the PCB, but the terminations aren't actually designed for this, and the capacitors flop about as a result. A skimpy, tie-wrap has been added to alleviate the capacitor's suffering, but this was insufficient and was found to fall off easily. And the bad news isn't restricted to PA work alone: the capacitors are loose enough for regular, high-level bass in the control room to cause metal-fatigue in the connections. I suggest that the cost-conscious makers invest in some double-sided adhesive foam, which, when placed between and underneath the capacitors would dramatically improve the support - at no extra cost (tie-wrap costs 2p, needs £30 tool: tape costs 2½p, no tool needed).
Whilst on the topic of capacitors, the DC blocking capacitor on the NFB (negative feedback) loop, a large value electrolytic as usual, is bypassed with a 100nF plastic film capacitor. This plays a subtle role in today's amplifier designs, improving high frequency distortion performance. Sadly, this (subtle) amendment has perhaps been unwittingly counteracted by a large ceramic capacitor elsewhere in the circuit. These exhibit high dissipation. If this latter component were replaced with a plastic film capacitor say, then all would be well.
Gaining access to the PCBs is dead easy, provided you don't jump to conclusions! I'll leave purchasers to evaluate the puzzle - a severe case of looking for non-existent difficulties... Servicing is also aided by the cards being hooked up with Lucar ("snap-on") connectors. Again, the bridge-mode function creates a snag, being hardwired, presumably until a new PCB is produced which incorporates extra Lucar terminals. But given that this is on the cards, then the Third Generation amps score top marks for ease of servicing - alongside the Quad 405, taking just under two minutes to remove and replace a card. This is fortunate, because the DC rails are fused, and the fuses are placed where? Well, there on the PCB, which is mounted upside down, presumably to make them even harder to get at than usual!
In passing, can I remind you that supply fuses on MOSFET amps using the Hitachi design (ie. as used here) rarely present the full DC rail voltage to the speaker, when they fail. This makes it less likely that a catastrophic amp failure will take out a pair of expensive monitors (on ordinary Bi-polar amplifiers this is a frequent event when DC rail fuses blow).
The toroidal transformers in this amp are unusual in that they feature an interwinding (or electrostatic) screen. As mentioned in our studio earthing series two months back, this has a lot to offer when trying to get together a monitoring system that's wholly isolated from the RF garbage that's floating about in the average mains supply. At present, amps with this simple, civilised feature are rare - let's hope Third Generation have set a precedent for amps in the latter half of the 80s.
Last word on the mechanics: the wiring for the meters' lamps is terminated on a bare tag strip, shared also by exactly adjacent mains connections. Nor are these mains terminals sleeved. This is fine for 50s wireless sets, but won't meet modern day domestic safety regulations: for instance, any undue flexing of the chassis could cause tags to jump free and touch, creating a highly dangerous situation.
Only the CMF 400 was audibly tested, primarily because we felt the time available (always a problem on reviews!) was best spent on the amp most relevant to our readership. On listening tests, its performance was up to the high standards expected of a MOSFET amplifier, apart from a dull top-end, compared with the Bose F1 amp, for example. In fact, the frequency response is down 1dB at 15kHz. Whilst this might be hunky dory for the PA man who doesn't want any hassles with RF burnout, a more flat top-end response is (a) desirable for monitoring duties and (b) easily implemented, by changing a few small parts.
Looking briefly now at the CMF 1000, this is broadly similar to the CMF 400, but features fan cooling based on a wind tunnel, like the Turbosound SSE "Fan Amps" of a few years back. Because the fan is necessary if the amp's quoted power is to be achieved, and because the standing dissipation of the amp is higher in any event, and because there's no means of slowing the fan down to make it run quieter, this particular model is less than ideal for control room use: amplifiers of similar power ratings exist (such as Yamaha's PC2002) which can operate safely without a fan.
In passing, the CMF 1000 also had two worrying mechanical discrepancies. First, the meter lamps are wired to the mains. There's nothing wrong with this: they work at low voltage off a primary tap on the toroidal transformer, but it would be nice if they were sleeved (ouch!!), and also if there was a warning notice inside to this effect. After all, the lamps will blow at some point, and if someone in a hurry forgets to isolate the mains supply before changing the bulb, they may be electrocuted.
Second, the reservoir capacitors are mounted on the PCB again, which is good, but in the absence of PCB supports, their weight imperils the MOSFET devices, causing the leadouts to snap. The CMF 1000 is, therefore, in need of a few refinements before it can score its certificate of safety and roadworthiness, particularly as it's befitted almost exclusively to PA applications.
The CMF 200's output into normal 8 ohm loads is very modest, at 60 watts. Unless you are using very efficient monitors, it's therefore best kept for powering micro-monitors or foldback. The amp is perhaps unique in being a 19" rack-mount, yet with only one MOSFET device per side, per channel. Also, there's no metering on this model, and the push-pull drive needed for bridge-mode operation comes from a small sub-PCB bearing an op-amp masquerading as a polarity inverter. This isn't bad, but the CMF 400 and 1000 derive their phase-reversed drive signal with no additional circuitry, using the Rimmer technique: the output of channel one feeds the inverting (feedback) input of channel two.
These Third Generation amps score no marks for disguising the source of their circuitry, but 99% of electronics is plagiarism, and the circuit topology is in the top class. The styling and aesthetics are also well above average for a budget UK power amplifier, and doubtless some Japanese gentlemen will be losing sleep over it. These amps also score high points for their generous provision of connectors when it comes to plugging things in.
The failings noted, however, will prevent these amps achieving the reputation they deserve, but all concern basic safety aspects and simple mechanical details rather than any major design faults. They're all easily put right, and pending actual production amendments, a dealer should be able to help you out.
Lastly, each amp comes with a handy booklet for beginners. Called "Audio Mixing", it shows and describes how to wire up Third Generation amps, mixers, graphic EQs and active crossovers for your first gig. In other words, it's aimed mostly at PA newcomers, but there's a short section on home studio mixing, and a handy glossary. The CMF 400 costs £1.20 per stereo watt into 8 ohms, and is worth serious evaluation and auditioning if you're seeking the benefits of MOSFET amplification.
CMF 200 £315.10, CMF 400 £392.15, CMF 1000 £673.90 inclusive of VAT.
Details from Third Generation Audio Products Ltd, (Contact Details).
In reply to the points we noted, Ron Bailey of Third Generation has informed us that peak metering and other indicators are projected on future, revised products. In addition, the PCB mechanical failings will be attended to post haste. Replacement PCBs can also be shipped within 24 hours in the event of a burnout.
Gear in this article:
Review by Ben Duncan
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!