Many of you know the Musical Fidelity A1 and A100. The circuit (of the power amp section) is easily available now and pcb's can even be got from some chinese ebay sites. In my opinion this circuit can still compete with other and later serious class-A designs (thanks to Tim de Paracivini who designed the A1/A100 'on a late afternoon'). At that time MF also had the MA50 and MA100 monobloks. The MA50 was just the power amp section of the A1 (same pcb, same enclosure) but 2 channels bridged. The MA100 were 2 A100's. Does anybody know how the bridging was done internally?
- connect the inputs of both channels
- disconnect the minus output of 1 channel from the chassis/earth
- connect this minus output to the plus output of the other channel
- now the problem remains of the inversion in phase/polarity of one of the two channels? I can't remember there was any extra circuitry in the bridged amplifiers (they used the same pcb's!). Did someone have a look inside the MA50 or MA100? Please tell me how this had been done.
- connect the inputs of both channels
- disconnect the minus output of 1 channel from the chassis/earth
- connect this minus output to the plus output of the other channel
- now the problem remains of the inversion in phase/polarity of one of the two channels? I can't remember there was any extra circuitry in the bridged amplifiers (they used the same pcb's!). Did someone have a look inside the MA50 or MA100? Please tell me how this had been done.
Thanks for the link but this one can't give me all the answers except:
- MA50 does not use a phase inverter but just puts 2 channels in parallel. Is it as simple as that, solder the inputs and outputs together, not one component (except transformer with higher voltage and dito capacitors) has to be changed? Never seen that before.
- What about putting in extra parallel pairs, how does it work out in this circuit? Which components have to be added, changed or deleted?
- MA50 does not use a phase inverter but just puts 2 channels in parallel. Is it as simple as that, solder the inputs and outputs together, not one component (except transformer with higher voltage and dito capacitors) has to be changed? Never seen that before.
- What about putting in extra parallel pairs, how does it work out in this circuit? Which components have to be added, changed or deleted?
Personally I'm with Mark Hennessy on these amps, the A1 design is junk. There's certainly no way I'd pay all that money for one.
It's not even true Class A, it's just heavily biased Class AB, and doesn't seem to have any thermal compensation into the bargain. It's not surprising they blow up.
I wouldn't risk bridging one. Remember that it will see half the impedance. Maybe if it had more than one pair of outputs per channel.
Note that in the later amps, they actually went with the B200 circuit, a totally different animal altogether (opamp + MOSFET)
It's not even true Class A, it's just heavily biased Class AB, and doesn't seem to have any thermal compensation into the bargain. It's not surprising they blow up.
I wouldn't risk bridging one. Remember that it will see half the impedance. Maybe if it had more than one pair of outputs per channel.
Note that in the later amps, they actually went with the B200 circuit, a totally different animal altogether (opamp + MOSFET)
rmgvs said:
*** yup = simple as that - doubling current capability of entire amp . going from ~7W/8E/Aclass to double that
old , known from tube era ... also used extensively with Quad 405 , but with additional power resistors ( around 0R1 ? ) from each live channel output to LS (current sharing )
****** you need double outputs , with dedicated collector resistors;
value of these resistors ( both present ones and additional ones ) must be doubled - to preserve same Iq as now .
or decrease them for your liking , to increase overall Iq
regarding quality - original A1 ( same as many similar MF products ) is junk - from engineer point of view ;
entire line stage is junk ( phono almost decent )
entire PSU barely adequate for job
hot ,scorching hot ......
anyway - just power amp stage -when realized decently ( from the scratch ) is on par with JLH and little Hiraga .....
differences between them are matter of taste .....
I am not interested in the A1/100 as we know it from the firm MF and in the particular boxes the chose for it. If you build the thing yourself you can make it as reliable as you wish. Also, the amount of class-A can be chosen as you think is best, just a matter of changing a few resistors (assuming you have enough cooling capacity).
The only thing that remains then is the question whether the circuit itself is capable of delivering the goods. And here we have a very interesting and puzzling question which I from a theoretical point of view can not answer. In the last 20 years or so I have built all fancy things in diy amplifier land, like Pass Zen's, Hiraga's, Linsley Hoods, Sugden's and many more. I am not saying that the a1 betters these designs (because that is not the case per se), but I am saying that this design has some real magic and quality that is not so easy to find for a commercial design.
But I am not intending to discuss the quality of the design, I just want to know how putting in extra pairs of output transistors can be accomplished. Then after building, everybody can be his own judge.
The only thing that remains then is the question whether the circuit itself is capable of delivering the goods. And here we have a very interesting and puzzling question which I from a theoretical point of view can not answer. In the last 20 years or so I have built all fancy things in diy amplifier land, like Pass Zen's, Hiraga's, Linsley Hoods, Sugden's and many more. I am not saying that the a1 betters these designs (because that is not the case per se), but I am saying that this design has some real magic and quality that is not so easy to find for a commercial design.
But I am not intending to discuss the quality of the design, I just want to know how putting in extra pairs of output transistors can be accomplished. Then after building, everybody can be his own judge.
Hi,
the A1 desperately needs a reliable and appropriately sized heatsink.
But, then it's not an A1 any longer.
Do not bridge it.
If you want high bias ClassAB from this design then consider paralleling the output devices only, not the whole circuit.
Parallel the existing one pair, but add emitter resistors of similar value to the existing collector resistor. May be you could get away with Re=0r1 at these very low voltages.
Keep the single collector resistor at the same value. 0r22.
Consider using a more robust output device. The MJ15003/4 is superb.
A two pair output stage using paralleled 15003/4 would be even better.
Try reducing the bias current in each of the four output devices to 500mA resulting in 16W (2Apk) of ClassA into 8r0, for a total dissipation of 48W/channel.
the A1 desperately needs a reliable and appropriately sized heatsink.
But, then it's not an A1 any longer.
Do not bridge it.
If you want high bias ClassAB from this design then consider paralleling the output devices only, not the whole circuit.
Parallel the existing one pair, but add emitter resistors of similar value to the existing collector resistor. May be you could get away with Re=0r1 at these very low voltages.
Keep the single collector resistor at the same value. 0r22.
Consider using a more robust output device. The MJ15003/4 is superb.
A two pair output stage using paralleled 15003/4 would be even better.
Try reducing the bias current in each of the four output devices to 500mA resulting in 16W (2Apk) of ClassA into 8r0, for a total dissipation of 48W/channel.
Hi Andrew,
this seems to be the most elegant solution. No problem with the feedback then, this stays as it is. All you need is the extra pair of transistors and 4 power resistors of 0.1 Ohm per channel. The idle current remains the same (with respect to the 'old' situation) but is spread across the two pairs. I did a quick simulation and everythings seems to fit. Given the proper cooling, a power amp of 50 Watt class-A can be created:
- ac of 2 x 25 V (300 VA or more)
- dc of around plus and minus 32 V
- standing current 1,7 Amps per rail, 850 mA per device
- 27 Watt dissipation per transistor, 110 Watt per channel.
Right?
this seems to be the most elegant solution. No problem with the feedback then, this stays as it is. All you need is the extra pair of transistors and 4 power resistors of 0.1 Ohm per channel. The idle current remains the same (with respect to the 'old' situation) but is spread across the two pairs. I did a quick simulation and everythings seems to fit. Given the proper cooling, a power amp of 50 Watt class-A can be created:
- ac of 2 x 25 V (300 VA or more)
- dc of around plus and minus 32 V
- standing current 1,7 Amps per rail, 850 mA per device
- 27 Watt dissipation per transistor, 110 Watt per channel.
Right?
1.7A @ 8 ohm is around 50 Watts RMS class-A. The transformer is 300 VA per channel (2 x 25 Volts AC 6 A), making 600 VA for 2 channels. The DC of around 32 V seems fit to deliver around 50 Watts in 8 ohm too (it is the voltage used for the A100, but in that case with less idle current). I think the MJ15003/04 can stand the dissipation easily. We'll see.
This Krell topic is very informative. Well, it's not my aim to make a Krell-like A1. In the end, if you like the sound of the A1 and the power is enough for your needs, it is probably best to keep close to the original apart from using ample cooling and a bigger power supply (still keeping to the 20-25 Volts DC of the original A1), maybe getting 1 Amp of idle current using two pairs of output transistors type 15003/04.
Or in other words: do as Andrew says ...
Or in other words: do as Andrew says ...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.