Is there any reason why this ps topology can't be made to perform as well as any other (LTP, etc)? Obviously matching the anode and cathode resistors as close as possible is important, but are there any other tweaks/measurements necessary?
The split load/"concertina"/cathodyne phase splitter is just fine. Keep in mind that the "concertina" phase splitter is a "tricked out" voltage follower and, as such, has a small insertion loss. I say voltage follower given a MOSFET works at least as well as a triode.
FWIW, Mike Samra and I have been and continue to work on a radical rebuild of an EICO ST70. The plan includes conversion to Williamson style circuitry and there is 1 Noval and 2 Octal sockets available for small signal purposes, which is not enough for "textbook" Williamson. Hefty power MOSFETs serve in the "concertina" phase splitter role.
FWIW, Mike Samra and I have been and continue to work on a radical rebuild of an EICO ST70. The plan includes conversion to Williamson style circuitry and there is 1 Noval and 2 Octal sockets available for small signal purposes, which is not enough for "textbook" Williamson. Hefty power MOSFETs serve in the "concertina" phase splitter role.
Attachments
There's limited voltage swing at the outputs. Parasitics should also be similar at both outputs.
In theory, yes.
But are there exactly matched output tubes ?
Usually not, therefore plate- or cathode resistor should be adjustable to compensate
the difference of output tubes
Broskie has info about the split load phase splitter (concertina) and Zo.
Octal Aikido All-in-One & Split-Load Phase Splitter Zo
Octal Aikido All-in-One & Split-Load Phase Splitter Zo
Only a single 9 pin socket and a change is NOT happening. Experience working with Jeff Yourison on his 6Y6 project shows the FET split load to be excellent. Volts were in really short supply and a triode's inability to swing close to the B+ rail would have ended in a failed project. Tom Bavis stepped in and helped get things squared away. Even with a FET's ability to swing close to the B+ rail, Jeff was initially experiencing some clipping.
Attachments
As alluded, it has insignificant affect on frequency response or distortion if used appropriately. The Williamson amp followed it with a balanced driver stage before the output stage, which avoided issues related to large signals and non-ideal loading that often show up in other amps.
As with all use of MOSFET's, especially where there is a large difference in drain voltage, cascode them so the input capacitance does not change.
cheers,
Douglas
cheers,
Douglas
The concertina is the 'best' phase splitter with the major caveat of not having any gain.
LTPs, paraphrase's and differential with CCS don't sound as good.
If you are going to use a concertina followed with a cathode follower to drive higher power, you are then using 2 stages (and 2 + 1/2 in the other channel) which provide no gain!
So the obvious solution is to use another type of splitter which provides some gain if you need more power, like grid current, high voltage swing and unbalanced loads.
Remember that in an amplifier each stage should have the more gain, high input, low output impedance and low distortion.
LTPs, paraphrase's and differential with CCS don't sound as good.
If you are going to use a concertina followed with a cathode follower to drive higher power, you are then using 2 stages (and 2 + 1/2 in the other channel) which provide no gain!
So the obvious solution is to use another type of splitter which provides some gain if you need more power, like grid current, high voltage swing and unbalanced loads.
Remember that in an amplifier each stage should have the more gain, high input, low output impedance and low distortion.
I just converted my Stupendous amp to split load (concertina/cathodyne) from transformers. Much improved.
All true, and you rarely see a cathodyne without a preceding gain stage. For most fairness in comparison it's best to look at the volt-amp/concertina pair as a compound circuit, the same way you would look at the two triodes of a differential pair. When you look at them as a whole it starts to diminish the larger difference between the two schemes. The PSRR of the differential is a pretty big advantage, but it's not too tricky to make the lower PSRR of the cathodyne non issue.
FWIW, I think the cathodyne sounds better most of the time, probably due to the difference in cancellation of harmonics inherent to the way a differential operates... not that it's a huge deal or difference though.
I prefer the cathodyne for 95% of my builds, as they are a little easier to set up without the addition of a negative rail or a tail CCS. For most practical purposes I feel it's easier to implement a cathodyne compared to a LTP, but if you already have a negative rail for bigger or more difficult to drive output valves in fixed bias, a differential starts to make more sense as you will likely be using some sort of follower anyway too.
My favorite phase splitter for lower demand grids tends to be a cathodyne with preceding gain stage, using the 6SL7. For the big boys I've used cascaded differential stages with mosfet source followers, like in the now famous 6L6GC AB2 thread.
Last edited:
- Status
- Not open for further replies.