Split plate loads

[echuta13]

Another good thread on split-loads over at Watkins/sewatt (need to join to see it though): http://www.wattkins.com/node/18083

I've used split loads in a couple of amps, and have been pretty happy with the results, and I liked the idea of less components in the signal line, but Merlin was advising against their use for the same reasons that Steve Conner mentioned.

[Chuck H]

After my realization on the ghosting (which I hate) I'm probably going to steer away from using them. Tough about the drawbacks. Too good to be true. Like when you see a girl at the bar. From behind she looks pretty good. And she's all alone! So you start that direction. As you get close she turns around and you can see that she looks like Geddy Lee!?! So you walk past and head for the can like that's what you were doing in the first place. Oh well... Too good to be true.

[Gingertube]

Chuck,
No need to steer away from them entirely - its more a matter of just being aware of the limitations in their use.
We have a heap of circuit techniques in our "toolkit", its just a matter of choosing to achieve the end you want. I tend to use interstage attenuators in preference as they also offer the opportunity of a little tone shaping by adding a cap or a zobel (cap + series resistor) across either the top or bottom arms of the attenuator, but so long as the rail is clean then the split anode load can work well and I have certainly used it in the past.

My posts above and (probably Steve's as well) were intended to be in the vein of "sure use the split anode load but be aware off / pay attention to these possible limitations/problems".

A bit of my philosophy:
I'm pleased that you read such posts and take note - I certainly take note of yours, I can maybe claim to more technical knowledge (I've been a design eng in the day job for more than 20 years and was a tech for 15 years before that) but the sort of practical experience and knowledge you bring to this forum is invaluable. Thanks for that. That is why these forums are so good, we can all get the "blend" of what the theory says and when practice says that there was something wrong with our theory. This is why I'm not a huge fan of modelling, in the final instance you have to just have to build the thing and listen to it. Believe me, I long since ran out of fingers and toes to count the number of my theories that have "bit the dust" when subject to that final definitive test.

Over at DiYAudio I have a Tube HiFi Power Amp design I posted, thread runs to over 1200 responses and 245,000 views and it has now been built by more than 50 folk around the world. That design was done according to theory but I deliberately never used an oscilloscope or distortion analyser etc. in final set to work and tweaking, all that was done by ear. I'm told that it measures quite badly, none of us who have built it give a "rodents hind quarters", we are too busy enjoying the music.

Cheers,
Ian

[Steve Conner]

In my experience, if you had taken the time to really engineer your hi-fi amp and make it measure well, the extra complexity and cost would have put DIY builders off and nobody would have built it. The DIY community will choose a simple circuit that doesn't work over a scary complicated one that does work every time.

Not to say that your amp doesn't work. If you consider that the goal of building a tube amp is to get a tubey sound, then you want it to measure badly!

All this business of the split load is splitting hairs. For modest amounts of attenuation I would probably "just do it". And the Geddy Lee girl too. :-O

[jpfamps]

Using split anode load is my preferred method of attenuation, as it seems a more elegant solution to using a resistive attenuator after the decoupling cap.

As stated above the advantages are reduced Johnson noise, and consequently I'm very surprised not to see more split anode loads in high gain amps where noise is an issue, and interstage attenuation is de rigeur.

The obvious disadvantage (again stated above), the reduced immunity from power supply rail noise, can be assuaged by appropriate design of the power supply.