EL84 Amp - Baby Huey

[boywonder]

Originally Posted by Hylle

If this will work, why was a source follower used in the 6V6 Baby Huey in earlier posts?

There is also an EL84 version of the BH with mosfet followers; the mosfet followers provide the needed current to the output tubes (to drive the miller capacitance), relieving the input tube of the need to provide lots of current when they are high mu, low gm tubes like the 12AX7.

Do your output transformers have 4 ohm speaker taps?

[Hylle]

Yes, the transfomers have 4 ohm output.

By the what is the input sensitivity for max output on the BH from the first post?

[Audio_idiot]

Ian,
What do you think of instead of 12AX7/ECC83/ECC803, a cascode differential with a low capacitance JFET at the bottom & a 6DJ8/6922/ E88CC/ECC88 at the top. Should this give a better drive to get over the miller's capacitance of the output tube? This also opens up choices of FET since the voltage rating is lower, AFAIK, this will be easier to implement then a MOSFET source follower...
I don't think it will complicate your streamlined circuit any more than adding a source follower MOSFET.

[hidnplayr]

Originally Posted by Audio_idiot

Ian,
What do you think of instead of 12AX7/ECC83/ECC803, a cascode differential with a low capacitance JFET at the bottom & a 6DJ8/6922/ E88CC/ECC88 at the top. Should this give a better drive to get over the miller's capacitance of the output tube? This also opens up choices of FET since the voltage rating is lower, AFAIK, this will be easier to implement then a MOSFET source follower...
I don't think it will complicate your streamlined circuit any more than adding a source follower MOSFET.

No, a cascode has very high output impedance

[Audio_idiot]

So does 12AX7, ECC83 & ECC803. If I understand correctly, it is one of the pre-requisites for this plate to grid feedback to work.

[hidnplayr]

That is true, but I believe the cascodes output impedance will be too high.
Feel free to prove me wrong

[malcolmfraser]

Hello all,

So I get mounting the power transformer on the chassis and only have one grounding point on the chassis to reduce hum. My transformer also have an earthing lug on one of the bolts that clamps the laminations (top of transformer).

So, question..

Go with the standard option of just mounting the Xformer on the chassis and making sure there is good conductively between the base and the chassis (free of resin etc), or isolating the xformer from the chassis and connecting the earth lug back to the single earth point on the chassis via a substantial cable?

Thanks in advance ;-)

[gingertube]

Hylle,
All of the 6V6 versions I built and have posted here were fixed bias. That meant a maximum value of 100K for Rg1 on the output tubes. To accommodate that I used the source follower.
If you are going to use push pull parallel 6V6, then the Rg1 are in parallel and have a maximum value of 50K for fixed bias and 220K for cathode bias. Using the source followers becomes essential to buffer the front end.
Cheers,
Ian

[gingertube]

Audio_Idiot and others with questions above,
The requirement for the BH front end is low current (approx 0,6mA per side of the diffamp) and high rp. The high rp is required so that the feedback is not divided down too much by the internal impedance (rp) of the tube.
The low current is required to keep the anode load resistors high whilst maintaining resonable voltages at the anodes of the diffamp. Be aware that 12AX7 can run into grid current problems at low anode voltages.

Cascodes have high rp and would therefore be suitable. Choosing something which will run at low currents is the next trick. A JFET triode cascode diff amp would work well and in fact that would be a good solution for those who wanted to build a 2 tube per channel 6BM8 Triode/Pentode based BH.
6DJ8 have a reputation of not being great at low currents but I have seen them used at low current. If doing a JFET triode cascode diff amp I would probably choose BF245 plus 12AT7 but try the JFET/6DJ8 cascode by all means. There is also another option - The gain of a cascode is proportional to the gm of the lower device x the upper device load resistor. Bipolar transistors are the kings of gm. Perhaps try a matched pair of bipolar transistors rather than JFETs. I would need to look up the part number but 2N5551 and 2N5401 (can never remeber which is the npn and which is the pnp) are both available as matched pairs in a surface mount chip. For the bottom of the cascode you need npn.

I've been using a 6V6 / 6SL7 BH for many months now.
MY EL84 BH monoblocks are STILL sitting on the workshop bench to have their front ends re-done using 6BR7 (EF86-ish) pentodes for the diffamp.
Malcolmfraser,
Just ground the transformer to the case. To keep noise low you should have a single connection for the 0V (signal ground) to chassis. Keep this connection away (opposite side of chassis if you can manage it) to the safety ground chassis connection. If the power tranny has a shield then connect that to the safety ground as well, rather than the signal ground.

Hope I've caught up with all the questions.

Cheers,
Ian

[Audio_idiot]

Thanks Ian for the heads up.

2N551 is NPN.

While we are at it, would it be beneficial to implement an active load for the cascode or the raw 12AX7?

Regards

Ken