Musical Fidelity MVT mk3 preamp line-stage

[rmgvs]

I have a copy of the circuit of the musical fidelity MVT preamp (mk 3) of around 1988. The copy is very bad and therefore I puzzled somewhat and drew the version that I attached with this thread. The MVT did get a lot of positive reviews at that time and I think it must have a lot of potention. Especially the fact that it is a DC-coupled discrete and simple line-stage from input to output seems of interest. Furthermore the output stage is heavily biased in class-A.

I tried to understand and simulate the circuit but have a number of questions where we can help each other out maybe.

The circuit consists of 3 stages, and this is what I make of it:

1. a differential input stage with innovative use of the LM318 IC in order to get around 0 mV at the DC output

2. a midstage with a bc550 (and associated components inclusing constant current source) in the negative rail and a bc560 (and associated components) in the positive rail.

3. an output stage in push-pull using complementary bd139/bd140, class-A.

A few problems and questions:

1. How does the LM318 work in this circuit? I do not have this component in the library of my simulator and therefore did not succeed in checking the workings of the circuit. It seems to me that the use of this piece of electronics is getting a zero offset at the output since no other adjustment is possible in this amp.

2. The midstage seems to be symmetrical but if you look carefully the positive half has an extra bc550. I do not understand this set-up and I am inclined to think it is a deliberate error in the schematic or something special whick indeed I do not understand.

3. In the original circuit bc414/416 are used instead of bc550/560 and bd135/136 instead of bd139/140, but this is of no importance to the questions posed.

4. I did simulate the circuit assuming the input is a normal differential input pair with 47k resistors to the positve power rail and making the midstage symmetrical by leaving the odd bc550 out. In this case the circuit seems to work and has around 20 dB of gain, but 2 transistors in the midstage are idling at around 13 mA of current while the output stage idles at around 4 mA. Well, this is another circuit and my goal is understandig and building the original MVT.

[Ultima Thule]

Almost breaked my neck...

Cheers Michael

[rmgvs]

You're right. I shifted the picture for your convenience.

[jez]

OK......
Your first three points,
1/ (and also 1 in your second questions) The LM318 is not used merely as a servo, it is an important part of the design. The unusual way it , at first, appears to be used (your diagram looks correct) is due to the internal input stage being switched off/disconnected by the connection of the inputs to ground. The long tail pair coming from the op-amp pins merely replaces the internal LTP with a much lower noise version (see National Semiconductors "linear handbook" and application notes for the LM318. MF used this lots in various designs)
2/ Yes.
3/ Yes.

Your second list of Q's:
2/ The extra BC550 (diode connected) IS NOT A MISTAKE.
3/Bc416/414 were interchanged with 550/560 all the time at MF, there is no differance to performance. BC214/184 were used in early MVT. Bd139/140 will indeed work fine.

Finally (and I speak from experiance here), the MVT mk2 sounds MUCH better than the mk3. I've just converted my DIY mk3 to mk2 spec after hearing the differance between factory made mk2 and mk3 amps. The mk2 uses the same circuit that your query is about (with some minor variations) BUT it does not use the conventional volume control of the mk3...it uses TL071 series op-amps (I know more silicon in signal path should not be good but it works!) as a buffer and then another as a feedback volume control as used in the A1 etc, this then feeds the stage discussed above. The mk2 has four of the class A stages arranged to provide both inverting and non inverting outputs for each channel. I have built only the inverting ones as I think it sounds a bit better.
Jez.

[jez]

Oh yeah....there are some compensation caps 1nF missing from your diagram. it will oscillate without these!
Good luck.

[rmgvs]

Wow, this is some information. But as usual, new info calls for new questions:

1. How is zero-offset guaranteed for? In your diy version: how does it behave in this respect (is indeed no output cap needed for safety reason when using a DC power amp)?

2. What are the idling current in the output stage and in the mid-stage? Is cooling of some devices necessary?

3. I don't get it that an extra opamp stage better the sound quality of the preamp. Maybe just using the opamps and leaving this circuit out completely is yet the best option, just as in the original A1? Have you tried that? Do you have an explanation for this odd state of affairs?

4. I do not succeed in simulating the circuit with the BC550 as diode. Could it be that this has something to do with not being able to get the LM318 in the circuit or can the midstage be seen as a separate entity that must work anyhow

5. I would greatly appreciate the schematics of your best working version, if needed by personal email. Thanks. Do you have some pictures of the boards, would be nice too!

[jez]

Originally posted by rmgvs
Wow, this is some information. But as usual, new info calls for new questions:

1. How is zero-offset guaranteed for? In your diy version: how does it behave in this respect (is indeed no output cap needed for safety reason when using a DC power amp)?
There is no zero offset guarantee. Use well matched semi's and there will be no problem. (2mV or so.no output cap needed)
2. What are the idling current in the output stage and in the mid-stage? Is cooling of some devices necessary?
No idea. heatsinks are needed for output devices. (did you not have access to one to trace the schematic?)
3. I don't get it that an extra opamp stage better the sound quality of the preamp. Maybe just using the opamps and leaving this circuit out completely is yet the best option, just as in the original A1? Have you tried that? Do you have an explanation for this odd state of affairs?
I don't get it either!, even stranger is that using "better" op-amps than TL0 series gave worse sound!? Yes it occured to me that maybe just the op-amps could sound better still....have'nt tried it yet.
4. I do not succeed in simulating the circuit with the BC550 as diode. Could it be that this has something to do with not being able to get the LM318 in the circuit or can the midstage be seen as a separate entity that must work anyhow
The bc550 (strange isn't it) is probably there to help obtain correct bias for the output stage and/or to compensate for an offset in the output of the LM318 and thereby allow equal +- swing at output.
5. I would greatly appreciate the schematics of your best working version, if needed by personal email. Thanks. Do you have some pictures of the boards, would be nice too!

I was waiting for this!!!! I dont have diagrams in electronic form, I would have to scan it and convert to PDF etc.....will try to when I have the time. I did'nt use PCB's for this.

[jez]

Hmmm the "quote" function didn't work out as I intended

[rmgvs]

I did not see ever a real MVT in life. I via-via got hold of a bad copy of the circuit.

I succeeded in simulating the circuit including the diode-transistor. But only by making a LTP by myself (using 220k en 47k as resistors, else I couldn't get reasonable working) and not by using the intended LM318.

If anybody has a spice-model for the LM318 ....

Then I get the following results:

- total gain is around 21 dB.
- idle current of output stage is 34 mA, dissipation of BD's is 0,6 Watt per device. Not cooling them seems a viable option.
- idle current of BC's (4 of them) in midstage is 14 mA, dissipation 0,23 Watt per device.

[jez]

Small heatsinks are used for output devices but not any others in MVT.They run pretty hot so are probably needed.
P.S. MVT= (I am fairly confident) Micheal Vaughn Thomas, also designer of the Dr. Thomas power amp and original "The Preamp" from MF....just for info