Thanks. It's a good question - looking at an old thread on MA50s they are closely related.
A poster called DonkeyTrousers reported major improvements when replacing the caps.
I recall (rustle, rustle) that Mark Hennessey had a site with much info on these amps
I wonder if the preamp in the A1 is the difference? Or the power supply - special editions of the A1 had external powersupplies (according to Mark)
Ha - reading through Mark's site pretty he much pans the pre-amp. Noisy and grainy. TL-084 with DC current through a pot.
Last edited:
Well, the improvement over the A1 makes sense, because the bias current places the MA50s further into class A. The A1 is scarcely class A, considering the lower power means it will be switching in and out of Gm doubling more frequently than MF's "99% class A" suggests, if you have average speakers, programme and room to listen in.
A fan would have to blow along the top cover plate, since that's the heatsink. Fitting it internally would mean heating some components that may not have been so previously and fans couldn't be bigger than perhaps 58mm overall height. The ventilation is at the sides which are not very deep at perhaps 250 mm. internally.
An external fan box parked alongside might be just the answer for cooling if you feel it necessary as you might in warmer climates than the UK. It seems not too many of these MA50s were made and few escaped the UK. Even the A1 was a an over-expensive slow mover here in the early '80s. The A100 was probably better all round- just haven't seen many.
Here's a link to Mark Hennesey's interesting pages Musical Fidelity A1
A fan would have to blow along the top cover plate, since that's the heatsink. Fitting it internally would mean heating some components that may not have been so previously and fans couldn't be bigger than perhaps 58mm overall height. The ventilation is at the sides which are not very deep at perhaps 250 mm. internally.
An external fan box parked alongside might be just the answer for cooling if you feel it necessary as you might in warmer climates than the UK. It seems not too many of these MA50s were made and few escaped the UK. Even the A1 was a an over-expensive slow mover here in the early '80s. The A100 was probably better all round- just haven't seen many.
Here's a link to Mark Hennesey's interesting pages Musical Fidelity A1
I had actually bypassed the preamp and was using the A1 as a power amp. The circuit is EXACTLY the same as the MA50s, the only differences being the transformer voltage and bias current. Transformer voltages are 15-0-15V AC for the A1, 25-0-25V AC for the MA50.
Ian, the MA50 bias current is actually half that of the A1, making it ultimately the same, the two channels being parallelled. So I dispute that the MA50 is operating "further into class A".
The main reasons for improvement that I can see are
i) increased voltage headroom
ii) halved signal current in the driver transistors
iii) easier load impedance - effectively doubled
I should think ii is the main reason. iii is just ii looked at in another way.
While we're looking at this, can anyone see the need for C6 & C7? (http://audiobbm.files.wordpress.com/2012/01/mf-a1.gif)
I can't see why there would be any DC across either of them.
Each MA50 has 4 of these caps. If I replaced them with wire links it would be DC coupled, or if I wanted DC blocking, I could put one big polypropylene cap on the input. My only concern is that it might have some effect on the bias current. Can an electronics expert please advise? It's quite a few years since my peak as a designer.
I can't see why there would be any DC across either of them.
Each MA50 has 4 of these caps. If I replaced them with wire links it would be DC coupled, or if I wanted DC blocking, I could put one big polypropylene cap on the input. My only concern is that it might have some effect on the bias current. Can an electronics expert please advise? It's quite a few years since my peak as a designer.
Last edited:
Sorry, my bad for taking Tim de Paravicini's description at face value and not being more careful reading what M. Hennessey's comments actually meant - prognosticating. Both misleading.
If both channels in parallel now share the same current, then there is effectively a lower AB crossover threshold and the figure of 0.36 A is the appropriate bias level to consider as defining the AB crossover. That means less, not more or same bias and hence less time in class A, more AB crossover artefacts evident.
If you want to maintain the same effective AB crossover point, you obviously need to double the current when doubling the number of output junctions of an amplifier which is what happens when you parallel 2 amplifiers running with the same supply. Otherwise your 2 amps simply run at half bias each which isn't the same as a single pair of transistors with the same bias current since Vbias (applied to the driver bases) will be higher, effectively as with the A1.
I follow your reasoning in (ii) and (iii), not that loading should be of great consequence, considering full class A currents would be around 3A per push-pull output pair (constant) and a few pairs would be needed for 50W. As you began, this is a dissipation problem exacerbated by 35V rails. The A100 is two high power channels with the same current again but fan cooled, so I imagine this is all quite doable in the same case and as Sajti suggests, reduced voltage supply fans can run virtually silent. Test the type first - some are just noisy by poor design.
Last edited:
I don't get this AndrewT, could you clarify please?
One input transistor will draw a positive bias current, the other will draw a negative current, similar in value if the hFEs are similar. By joining them together, they will to an extent cancel out leaving a net bias current which is smaller than either of the two individual bias currents. This net bias current is sourced from the zero volt rail via R9//R8//(R4+R3), leaving a tiny dc voltage offset at the join, just as there were tiny, though different, dc voltage offsets before. I just can't see a problem.
Come to think of it, I've built several amplifiers in the past, of a similar configuration where the bases of the top and bottom input transistors were joined directly together. And they worked fine. I've also joined together inputs of op-amps with no detrimental effects so I can't see how they could blow up. Surely it would only go wrong if you forgot to provide a current path for the net input bias current?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.