Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › A pair of variations on the RKV theme
New Posts  All Forums:Forum Nav:

A pair of variations on the RKV theme

post #1 of 5
Thread Starter 

When I first started visiting head-fi, the Audiovalve RKV was one of the "ultimate" amplifiers you could get. Later on, I was lucky to find an early kit from Elrad, based on the same schematic (which can be found here , along with many explanations about the rkv). I've since sold this one but I still kind of regret it. So I tought that it might be interesting to consider building something similar but not a complete clone (that wouldn't be fun).

 

The RKV uses a highly heretic approach: a complete tube amplifier (gain stage, inverter, output stage) inside the loop of a pedestrian opamp. It has advantages: high psrr, stabilization of the tubes operating points even as they age, linearization of the amp leading to very low thd, reduction of the output impedance. It used 4 pcl805, which have a weak triode section (used for gain stage and inverter) and a strong pentode section (used for the output stage), strong enough to drive the AKG K1000.  The RKV has also some points which aren't that great. The main one is that the amplifier has very high gain (almost 50x). Why ? The amp can only be made stable if the opamp is set to have more gain than the input tube. The PCL805 has very high µ (60). Another problem is that they used a quite complex power supply to get bipolar supplies and various bias voltage. It can be made simpler (somehow).

 

Here's a preliminary schematic:

 

 

 

 

 

 

Let's get rid of the pcl805, with its high µ triode. Most triode-pentode tubes have high µ triodes, so they're all out. For the input stage, a common lowish µ triode is the ecc82/12au7, with a µ at about 17-20 (at best). Still a bit high but it can be reduced with the use of an unbypassed srpp. It will provide a voltage gain of about x5. The opamp gain can now be reduced from 50 to 8. This adds a triode. But to stay with two tubes per channel, I can remove the inverter and drive the lower output tube directly from the opamp. While I'm at it, let's dc couple the srpp to the output stage. What about the output tubes ? The pcl805 was more powerful than needed for almost all headphones. An ECC99 or 6N6P is probably a sensible replacement. At 20ma/150V, they can run at 3W each (well, one runs at 4w since it has to drive the feedback network). Still under the 8W total power of the tube. It could probably run at 15ma/130v.

 

To simplify the power supply, all the negative supplies have been shifted to gnd. The lower gain imposes a higher voltage reference at the input, which means that the opamp can do with a 0-24vdc supply. The output is now biased from the cathode (and can be adjusted easily). The ecc82  heaters can be supplied from the same 24V supply through a cheap switching reg (plenty of lm2596 boards on ebay). The 6n6p will require an elevated heater (the vhk max is 100v).

 

 

 

Of course, this could be even simpler and better, Torpedo inspired :p With a ecc99 instead of the ecl84, you'd loose some gain but you'd need only one tube.

 

 

So what do you guys think? Waste of time ? Worth trying one or the other ?

 

post #2 of 5
Quote:
Originally Posted by 00940 View Post
 

So what do you guys think? Waste of time ? Worth trying one or the other ?

 


Hey, anything is worth a try...thats the best way to learn what works and what doesn't.

post #3 of 5

Interesting.

 

The work I've been doing on direct-drive electrostatic headphone amplifiers caused me to think about differential circuits, which, coupled with the tubes I bought to mess around with push-pull, led to my developing an amp which coincidentally I called 'The Heretic'.

 

I'll post it up in a thread of it's own.

 

I'd go with the second of your diagrams there, with a few changes. I'd replace the current source with one based on a depletion-mode MOSFET, DN2540 will do fine in this case, it's good for 400V. If you're determined to get the absolute best performance rather than just go for simplicity, you can cascode it or yoke it to an IC CCS as here:- http://www.diytube.com/phpBB2/viewtopic.php?t=5615.

 

The other thing is that IMO it would be better to take the feedback from the output side of the transformer as this would result in overall higher fidelity.

 

w

post #4 of 5

Have you looked at the Bijou (or the futerman circuits it tries to emulate)? 

 

You may have success by replacing the tube *gain* stage with a simple phase splitter, and using the op amp for feedback & gain setting. 

 

Sounds like fun, whatever you do. 

post #5 of 5
Thread Starter 
Quote:
Originally Posted by nikongod View Post
 

Have you looked at the Bijou (or the futerman circuits it tries to emulate)? 

 

You may have success by replacing the tube *gain* stage with a simple phase splitter, and using the op amp for feedback & gain setting. 

 

you read my mind  ;) I don't think it is as "cool" as the previous circuits, as the opamp doesn't set the operation point of the tubes. But it sims good. 

 

 

This one is really nothing but a torpedo with both current and voltage operating points servo-ed. I think I'll start with that one when I get the transformers I ordered.

 

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › A pair of variations on the RKV theme