Musical Fidelity A1 - Mods:

This page details some suggested modifications. But before taking the amplifier apart, read the following disclaimer:

Disclaimer:

Introduction:

Because of the mechanical difficulties, the A1 is not a good amplifier to experiment with. There's limited space inside, and you can't run the amp without the top cover. With no output protection, a faulty output stage could do some serious harm to your loudspeakers. In other words, it's not for beginners!

Component changes:

I know of people who have gone to town with their A1, replacing components with exotic variations. Apart from using low-ESR 105°C replacement capacitors, I personally wouldn't bother - but do bear in mind that I'm an engineer.

I suppose you could try replacing the TL084's for something more fashionable... the OPA4134 would probably sound much better and result in a better signal to noise ratio. But my feeling is that if you go to the trouble of removing IC1 (the line pre-amp), leave it out! Build my replacement preamp instead.

So, if you want to try magic capacitors or silver-plated cable then be my guest, but I personally feel you'd be wasting your money.

Fan cooling:

This was something that I considered in the past, but never got around to doing. But it's recommended, especially for the summer months...

Noise is the potential problem with this; there's no point having a decent hi-fi that generates as much noise as a typical PC! So as you would want to run any fans as slowly as possible so that they are not audible from your listening position, some sort of speed control is essential. Given that the A1 was designed to work with no forced air cooling, even a relatively gentle draught will reduce the heat sink temperature appreciably.

My plan was to build a separate unit that could sit alongside the amplifier, containing one or two fans and a simple speed control. I preferred to not build anything inside the amplifier because of the possibility of the noise generated by DC fans getting onto the power supply rails. Also, there is limited space inside for anything other than the smallest fans, and these have to run at high speeds to shift air.

Based on my experiences with fans, you should choose the largest diameter you can get away with. I was considering 60mm or 80mm fans - the latter size is typically used in PC power supplies, so they're easily found if you don't mind a bit of "skip diving". These would have been taller than the amplifier so the air would have been directed over the top of the amplifier as well as through the internals.

DC fans are easily controlled by varying the applied voltage, so either build a variable power supply based on an LM317T or similar, or if you're not that handy with a soldering iron, buy a plug-top power supply with a variable output - these can be found with a slide switch that lets you choose 3, 6, 9 or 12V.

Preamp replacement:

You will have gathered by now that I don't like the preamp! It simply had to go, as living with the amplifier was becoming a bore. I know of a couple of people who have bypassed the preamp entirely, wiring a replacement potentiometer in the conventional manner and feeding it straight to the power amplifier. If you only ever listen to CD's, there's enough gain in the output stage to get away with that.

But as I use my tuner and other sources extensively, I had to have a gain stage of some sort as their nominal outputs are somewhat lower than the CD player. It had to be uninstrusive sonically, have none of the DC problems of the original design, and be much less noisy.

The last requirement needs explanation. Imagine that you've bought or made a 'pot-in-a-box' to get around the problem of a noisy volume control. This actually isn't a bad solution for the less-technically-minded... You'd set the volume on the A1 to a high-ish level and leave it there, making day-to-day adjustments with the external pot. You'd choose this level so that you'd get full output from your lower-level sources...

But, with the external volume at low settings for night-time listening, you'd be hearing a large amount of noise from the amplifier because the internal preamp is always running at a high gain. Not sure what I mean? Just listen to the amount of hiss the amplifier generates with the volume higher than 12 o'clock...

I wanted to have just this topology, because this avoids headroom problems and ensures the power amplifier is fed from a low-impedance source (compared with having a gain-stage followed by the volume control). But the amount of noise the existing preamp generates is basically the reason I didn't build this preamp much sooner...

To minimise the noise described above, the gain stage needed to have no more gain than necessary, as any noise generated by it will be present at all times... With an 'average' CD, the unmodified amplifier would start to clip at just after 12 o'clock. The resistance of the original potentiometer at this setting is 8K - working out the overall gain of the two op-amps at this setting results in +1.5dB from input to power-amp. At the maximum setting, the overall gain is a massive +17dB - far more than is needed... I decided that a gain of around +10dB would be sufficient. Of course, this means different volume settings - the control is not as fierce now and there is more range at lower settings...

I decided to use the venerable NE5532 for the preamp. I know that it is an old device, but you have to spend a lot of money to improve on it. The biggest advantage is the excellent drive capability, meaning that with a non-inverting configuration, you can have very low-value components in the feedback path, minimising Johnson noise. You can try different op-amps in this role, but if you do, I'd recommend that you increase the value of R3 and R4 to reduce the loading on the op-amp. You could simply increase them by a factor of 10 to 6K8 and 2K2, but you're running the risk of increased noise. To keep the same gain, R3 needs to be around three times the value of R4, so perhaps 3K and 1K would be a good bet. Good quality but affordable alternatives include the OPA2134 and the OPA2604. I haven't tested these, so you're on your own here!

The first part of the circuit is an ultrasonic filter. Note that this is simply to avoid noise in the MHz range upsetting the op-amp*, and is not meant to be a bandwidth-setting component. The cut-off frequency varies with source impedance: -3dB @ 340KHz with 0Ω source impedance, -3dB @ 190KHz with 1K source impedance. The amplifier bandwidth is set later by R4 and C5 on the main PCB.

(*) It's worth saying that this ultrasonic hash can result in audible differences between CD players (and interconnects) that don't exist with better (read: properly filtered) amplifiers. Conclusion? Poorer ampler is actually more "revealing" than better ampler that isn't being confused by a "dirty" CD player, and gets a better review as a result! That's just a little pet-theory of mine, and probably has no basis more reliable than extreme cynicism!

The volume control is a good-quality 10K log pot - deliberately low in value to minimise Johnson noise(*). Modern kit should be quite happy to drive this sort of load, although some exotic valve gear might not like it... This pot is AC-coupled to the op-amp by a non-electrolytic capacitor. As above, the exact -3dB point is variable - this time the position of the volume control will influence it. However, it is low enough not to be a problem and the LF cut-off is properly defined by C6 and C7.

(*) When I was a kid, I remember building an amplifier with a 1M pot as a volume control. At minimum volume, the amp was quite quiet. At mid-volume, it was rather hissy. But at full volume, it was quiet again! As a child, I didn't know about Johnson noise...

This circuit is not critical and can easily be built on Veroboard. The deliberately low impedances ensure that inter-track capacitance is not an problem. The small Veroboard assembly is held in place by the potentiometer and is connected to the PCB via short lengths of wire. These are soldered at both ends and made long enough to enable easy access - I decided not to use any extra connectors on reliability grounds.

Replacement preamp is small enough to be supported by the new volume potentiometer.	Shows the low-value resistors used in the feedback network...	Ultrasonic input filter, showing the high-quality capacitors.

This picture shows the modifications required to the original PCB - fortunately it's very simple. Remove IC1 and the two wire links that carry the signal from the tape-monitor switch to R33. Easily reversible, if required.

Next, take the power connections from pins 4 and 11 (red/black pair). Input comes from the removed links (red/yellow pair). Finally, power-amp inputs are connected where the potentiometer was, feeding R4. There is also a connection labelled RV1 GND which is conveniently near to this latter pair (see the red/yellow/brown bundle)

Before installation, I took the time to ensure the preamp was working by testing it with a bench power supply, signal generator and oscilloscope. As this unit connects directly to the power amp there is the chance of real fun if there is something wrong! One thing is worth mentioning - I originally replaced C6 and C7 with tantalum capacitors. The preamp has a slight dc offset, due to different dc resistances seen by the inputs - I measured around -200mV. This meant that the tants saw a small reverse-voltage across them, and I was worried that this would affect long-term reliability. A leaking capacitor could have serious consequences on the output stage, so I replaced them with 1µF polyester capacitors (the red ones as visible on this picture).

Finally, this shows the preamp in position before the top cover goes on. There's plenty of clearance between the lid and the back of the Veroboard, but a different potentiometer might reduce this margin... The slack cable is dressed away from the power amp, and an attempt is made to keep the inputs and outputs separate, but I doubt there's any real reason to worry...

This has worked reliably since installation, and has transformed my day to day life with the amplifier. The pots and switches have remained noise-free, and the stereo image benefits from the improved tracking of the pot. The preamp is extremely quiet - with your ear held against the speakers you can just about hear some low-level hum coming from the power amp - there's no hint of hiss from the preamp. Obviously, this applies at all volume settings, unlike the original preamp. It's also well-behaved - there's no sign of instability, and the amplifier retains it's civilised pop-free power up/down. The strange effect of fading up, down and up again (described above) at power-on has gone. It's hard to drawn any reliable conclusion on the effect on sound quality until I'm able to borrow an unmodified A1 to do a direct comparison. However, experience shows that the 5532 is a much better opamp than the TL084, so it's hard to imagine how this can be anything other than an improvement.

Comments:

Several people have modified their A1's as detailed above, and have reported positive results. Here's a typical comment, with thanks to Marcus Ackel:

So to your changes and my experience. Before the new preamp, the a1 had a touch of gain character in increasing volume like an guitar amp. Since built in your preamp it works very reliably and it sounds more like a highend machine, no more desturbing sounds or gain and both channel are at the same level (well balanced if you decrease volume, which is also professionalized by an alps poti).

So you made it really becoming an perfect professional grown-up amp.

And another, from Pedro Vasconcelos:

For the preamp I used an OPA2134 (with which I had great results in an headphone amp) and the components list your sugested, except for the pot, where I re-used the Alps 50K from the A1. The preamp is working fine, is much quieter than the original, better tracking a lower volumes. The sound quality has definitely improved: wider soundstage, very smooth highs, the sense of "grain" in the treble is gone. It also sounds more powerful, thought I attribute that to the replacement PS caps.

Preamp parts list:

To assist constructors, the components required for the replacement preamp are listed here. Components starting from 100 are for the right channel. Note that the feedback components are different to mine, but as mentioned above, these are recommended if you don't want to use an NE5532.

Notes:

• Part numbers and prices are from Farnell, in UK pounds. Note that these may vary, and will be affected by having to buy in quantity - for example, you need to buy resistors in packs of 50 (!) - for this reason you might consider using other suppliers.
• MF means Metal Film - these are recommend for best stability and accuracy, but feel free to experiment.
• The input filter capacitor isn't critical, and values either side of this are suitable.

Finally, more information to help constructors will be added at some point soon, including a layout diagram and more detailed instructions.

Onto the next section - other models...