|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
5.141.232.15
I was looking at doing this next month.
There are some refinements that need doing of course but the majority of it is there after lots of late nights checking all sorts of obscure details.
As you can see, the bottom pair operates very much as a straight fwd low potential push pull amplifier.
Cascode amps also exist in similar configurations but only single ended.
The main advantage of doing it this way, is the power transformer being driven current drive simultaneously in both halves of the primary instead of the customary "half on half off" see saw arrangement.
This obviously doubles the current, has NO quiescent current in the primary at idle (so you can get a lot more from the core with a lot less saturation problems), and doubles the voltage swing on the transformer owing to the differential (+190/-190V) power supply fed to the amplifier.
The G2 feed is complication in ultra linear mode so at present this is done with a 1:1 coupling transfo for the top valve pairs, but this can also be accomplished with a dual triode as the current is very low.
Moving from Pure pentode to hybrid triode behaviour has the advantage of keeping the high gain but much lower impedance drive to the OPT, as this is only 2K load, but total current flow in the primary is 120m/a at 300+ volts which is quite a bit....
The input/bias circuitry is insensitive to any variations in the supply rails because it's referenced to 0V and controlled by sub min gas discharge devices which also exist like all the other valves from old Russian military stocks...
See what you think!
Follow Ups:
OK.
After lots of enquiries, there's progress being made at last.The OPT needs a double TERTIARY wound core similar to above.
This cures the problems of feeding all the Grid 2s in one hit.Basically the bottom grid2s get fed in much the same way as per normal, but completely independently as per the image linked* and a second winding with the same characteristics again, fed from the +VE feed on the centre tap to the top grids.
As the transformer will in any case be the most expensive bit, then it makes sense to do it this way on a toroid basis as there's no quiescent current flowing in the primaries, (ONLY a small amount in the G2 circuit).
ahum!
Edits: 04/27/15
A post vanished here>
I wonder why?
The patent I was referring to was in the public domain since:-
2,791,646
ELECTRICAL SYSTEM FOR ULTRALINEAR
AMPLIFIER
Herbert I. Keroes, Philadelphia, Pa.
Application October 26, 1955, Serial No. 542,825
And is available from patentimages.storage.googleapis.com
So here is the corrected schematic again using the principles outlined.
You see the current path from the differential voltage rails?
The bottom circuit looks like classic push-pull, where ONE valve turns off as the other turns on, from a single rail 0> -190V.The main difference being as one valve turns OFF underneath it turns ON the top valve from the positive rail +190> 0.
That is a cascode SRPP amplifier, as explained by the excellent Mr Broskie in his tube cad journal public domain articles on cascoded stages dating from nearly 15yrs ago.
The load is driven through the 0V centre rail in both cases.
A cascode amplifier IS a push-pull amplifier. The load is developed at the centre.Classic SRPP output stages use a capacitor (valve/solid state alike) to drive the OPT because they don't usually have split supply rails (usually thru some form of Darlington arrangement).
That capacitor is invited to charge and discharge in both directions across the load.The interesting part is that the transformer in this case is wedged between the 2 halves and there is no need for an isolating capacitor, because it's a differential amp like an op-amp.
There is no quiescent current at idle, because both ends of the transformer are effectively at half VCC (already at 0V potential).
The only DC element is a small amount of grid current.The primary is being current driven in both phases, (as of 2 op amps in opposition).
It brings twice the primary load current, rather than the usual one half having no drive in a classic push pull arrangement while the other would swing current from only the supply rail to near ground and back.In effect, the in phase drivers are across the diagonals and the G2 drive is arranged such as to correspond with this reality to produce the classic ultralinear hybrid pentode/triode drive also top and bottom.
What the OPT sees is a conduction path which is phase inverted in each arm 1/ +190> 0 while arm 2/ 0 > -190 and vice versa, in what amounts to a classic bridged arrangement at the same time.
If you remove the connection to the primary CT you will revert to a normal P-P amplifier (with then double the primary impedance) so the power output drops back to non bridged mode.
Illustrated is an old (German circuit? Saba?) from years ago, where the signal for the top pentode G1 is derived from a resistor, and uses a non split supply.
It doesn't need a phase splitter, because it already is one.You can also derive the signal for the top grid from a voltage divider network, and that makes it act like a grounded grid amplifier.
In effect what we have here, is exactly what Broskie designed as an ultra-linear symmetrical cascode.
The top G1 of the top TRIODE has the SAME behaviour as the G2 of a pentode.
In fact a triode cascode stage behaves in almost the same way as a pentode.There are a number of options on the signal feed to the top grids via resistor voltage dividers/transformer coupling, or you can choose to drive the bottom valve, or on a pentode you can GROUND the G1 and drive the valve direct from G2.
In effect Broskie's stage illustrated below is in effect driving the stage ultra linear via a resistor network, while I choose to cross couple it with a smart transformer.
The circuit is to all intents and purposes near enough identical apart from bridging 2 symmetrical cascodes in antiphase.
I have no doubt it would be possible to drive headphones directly from such circuits without a transformer from HT as low as a +/- 24V line.
Edits: 04/28/15 04/28/15 04/28/15 04/28/15 04/28/15 04/28/15
Ah, OK, but in your schematic, the plates are driving the transformer, instead of the mu-follower's cathode.
I think if you look a little more carefully you will find BOTH the TOP cathodes and the bottom anodes are connected to the transformer.
?
Where's the cascode? It looks like the 'top' tubes just stiffen/load the output tubes' plates where the output is taken from. I'm missing something.. Have you run this through SPICE?
This belongs on the Tube DIY "board". I've asked the moderators to move it
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: