In the quest of refining the 6C4C push-pull design I started not long ago, it was suggested by 45 to try a different driver. The famous E180F russian equivalent with gold grid: 6J9P-E. So I went back at my design and replaced the 4P1L driver with the 6J9P-E. Instead of filament bias, the bias was provided by a pair of red LEDs. The low dynamic impedance of these avoid the use of a bypass capacitor. Setting the operating point with the existing supply is a bit tricky but I found that Va=210V / Ia=14mA provides good swing and minimum distortion. Biasing the valve at around -3.7V keeps the valve away from grid current which will increase the distortion unnecessarily. So here is the circuit:
The rest of the circuit remains unchanged. Of course if higher anode current is desired i.e. about 18-20mA, the anode voltage will have to be around 220V. So a different supply arrangement must be sought to achieve this.
Clearly this amplifier version will provide a strong sweet fingerprint given the higher H2 levels reflected in the simulation. H2 cancellation is not significant and linearity slightly worse than the venerated 4P1L brother. Interesting to see that the THD increased now to about 0.95% for 8W.
This driver will provide the character of this amp I suspect. I tried 6e5P with slightly better results (e.g. about 0.8% @ 8W) and will be worth listening to.
Ale, it was referred to the 4P1L as output valve following the comment of Raductar. It is useful anyway. I thought the 6J9P was a bit better at low level.
For the 6C4C PP, I would stick to the 4P1L. I would just add a 3rd stage in between (4P1L PP driver). So 3 x 4P1L + 2 x 6C4C. This way would also be a more universal power amp.
Hi 45,
Adding a third stage would be a good idea. Either with input transformer or with a IT between the 4P1L diff pair and the output stage. Alternatively , the diff pair can be choke loaded and DC coupled to the output stage. The latter will require a change in the bias arrangement i.e. a tail CCS for the 6C4C pair.
Have you tried any of these combinations in the past and do you have any comments/preference between them?
Thanks
Ale
Yes,
the push-pull driver works better for A2. It has more headroom. I have used the 6H30 but the 4P1L should just be as good if not better sounding in this application. In fact I have bought some 4P1L’s because I am thinking to try this.
The 6H30’s were running at about 175V/20mA coupled with the LL1635/5mA. The input stage I used was a simple concertina with another 6H30!
In this case I would start with minimum modification, so I would use the present 4P1L driver + 1660 in SE to PP as input stage and then add the 4P1L PP driver. Probably the LL1660/10mA in PP to PP is a better choice with the 6C4C because its grid current appears to be higher than the 45 from your simulation when you push into +15-20V grid region. One other thing that would be useful to know is what happens in terms of driving requirements if you drop the 6C4C primary load from 8.8K to about 5K.
Thanks for this. I think I have an LL1660/pp I could use with the 4P1L differential pair.
My OT is the limitation and unfortunately I don’t have a multi impedance PP OT, nor have a 5K one..
P.S.
With the 4P1L you have a lot more freedom in terms of operative conditions as they are not limited to 4W dissipation as the 6H30 and are also more efficient because of the filament consumption. To start, I would try something like 230V/25-30mA at about – 20-21V bias for the 4P1L.
Yes an significantly much cheaper! A (not that) hidden gem. That operating point is quite good for the 4P1L….
PS: I sent you an email with a differential pair version in A2, did you receive it?
What do you think: http://wp.me/p2r2tK-oq