Nina Mu Stage Preamp - Circuit and Details

[Mike C]

The project started when I tried E80CC's in my Basie preamp i.l.o. 12AU7's and got a major improvement; this made me wonder what results one might get from a circuit designed specifically for the E80CC.
I was hooked by the idea that it could be seriously good if done well, so I had to have a go (even though I'm not that experienced).

As Basie uses active loading, I chose that again, but in a triode/pentode mu stage. Remember I was looking to better a tweaked Baise with attenuators, cap upgrades etc.

For better sound I reckoned to run the E80CC's at fairly high current of about 5.5mA.
I planned about 8mA in the load resistor below the pentode, giving about 13mA in the pentode.
Pentode choice was easy; I'd used the E83F before and know it to be an excellent active load, and it's a very high quality item at reasonable price.

The PSU will be covered in a followup post.

Results:
It has exceeded my expectations and is even better than I dared hope for.
It does outperform the Basie by a signficiant margin, with better dynamics, more detail and clarity, and above all a very natural, musical, sound. Oh, and female vocals that turn me to jelly.
These opinions are confirmed by several friends into high-end hifi, with comparisons against another high-end DIY valve pre (and against a Naim
pre but that was so one-sided it doesn't count )

When I started this project, my wife Janet said she didn't know why I was bothering as she didn't think it was possible to improve on what I had.
I wasn't too sure myself to be honest. Anyway, Janet is now convinced.

More later ...

[Bas Horneman]

Hi Mike,

How about trying the following tweak...the 475k grid leak resistor is "bootstrapped" just as is your 12k anode resistor.

Doede Douma suggested lowering that to 10k or so. Makes a minute change but apparantly it does sound slightly better that way.

Cheers,
Bas

[Mike C]

Re a lower grid leak;
Evidently it's something to try, but I'm struggling to understand one point.

If I do a normal time constant calculation using the coupling cap and grid leak, the circuit as shown gives 3Hz (probably a bit lower than necessary).
But the alternative of 10kohm looks to give a lf rolloff at 160Hz. I guess the 'bootstrapping' must overcome this in some way?

Anyway, it's on my list of things to try.
I also intend to try fixed bias; and I want to take the attenuators out of the Basie and try those, along with a pair of Panasonic pots.

However, the way it's singing I'm very reluctant to mess about with it. I guess I may try a Mk 2 one of these days.

Thanks for the interest.

[Mike C]

The mu stage cct has good power supply rejection ratio, so I chose not to regulate the PSU. I'm aware of the various opinions about whether to regulate or not (and I respect them all); but in this application, on balance unregulated seemed best to me.

With unregulated, I'd guess that the performance might be more affected by the basic quality of the PSU so I went for the scheme shown below.
Note the use of 6CJ3 damper diodes, as these have a very low voltage drop as well as being 'clean'.
The mains TX is about 3 times over rated for low DCR.

PSUD simulations and pole and Q calculations were used to refine the PSU design, to get one with very low ripple, high isolation between channels and good stability, with a pole about 17Hz and a Q about 4 (if I remember right).

Refinements:
Two mono volume controls are used. I need a balance function (due to the room), and I personally have heard how much a normal balance control affects the sound; so it's dual mono controls for me.
These are shunted pots. This little tweak is very similar to half of a stepped attenuator; for the cost of two good resistors it's a no brainer.
Remember, I didn't want to throw too much money at this project, as I didn't really expect much success; so the pots are just Bourns (from RS).

The selector switch and volume pots are mounted direct in the signal path to keep wiring very short, and are operated by extension rods. As a result, the circuit is very compact, in one small corner of the chassis.

The PSU cct:

[dhaen]

Mike,

It all looks good
I didn't realise that damper diodes were available without top-caps
The Bourns pots you mention deserve more credit than they're given, as do the Spectrol equivalents.

[Bas Horneman]

The dynamic value of a 10k in that circuit. If Doede is right that is is being bootstrapped would be 2M.

Your 12k resistor would have an Rp of 2,4M... assuming a pentode voltage gain (cathode follower) of 0,995.

The Kimmel article:
http://home.zonnet.nl/horneman/ml/akimmel.htm


The Doede tip:
http://www.dddac.de/at08.htm

Cheers,
Bas

[Mike C]

John esp, and all who are interested ...

Just to confirm the 6CJ3 damper diodes don't have top caps; this makes them easier to use.
However, they are single diodes so you need two in a normal full wave PSU.
Also, they use an uncommon base; large 9 pin. I think the right one is 'Novar', but very close is 'Magnoval' which are easier to find. The latter have the same base, but is made for valves with thicker pins than the 6CJ3. But it's possible to bend the lugs slightly using a jewellers screwdriver inside the socket (ouch!), to get a tight fit.
The heater is 6.3V 1.8A; don't forget you need a pair so the heater supply needs to be 3.6A.

There are some octal dual damper diodes which are said to be good for preamp PSUs but the voltage drop is considerably more than the 6CJ3. With 'Nina' I was aiming high so I went straight for what seems to be the ultimate.

The 6CJ3 is rated up to 2.1 amps!
Voltage drop in a preamp is about 5 volts; not that much more than a SS bridge.
When run conservatively, I can imagine them being good for many years.
I think they are available from Billington at moderate price, or cheaper from the USA. Even allowing for needing two, it's still cheaper than those metal base GZ34's and probably better sounding.

[fdegrove]

Hi,

I think the right one is 'Novar', but very close is 'Magnoval' which are easier to find.

As you found out those sockets aren't quite interchangeable.

The magnoval is used for valves like the EL509/6KG6 and won't fit in a novar socket.

So be careful and try to find the correct socket. If I'm not mistaken AES and Triode Electronics both carry them.

FWIW, voltage drop would be about 25V/700mA according to the datasheet.

Thanks for sharing your circuit, Mike.

Cheers,

[bequerel]

Will four damper diodes in a full wave bridge give me 1.4 times the transformer output voltage just as ordinary diodes?

[fdegrove]

Hi,

Will four damper diodes in a full wave bridge give me 1.4 times the transformer output voltage just as ordinary diodes?

No, voltage drop depends on current draw.
So it's not possible to calculate accurately unless you have a very complete datasheet with all possible measurements taken at various operating points.

Cheers,