Paraphase Inverter driving modulator grids

[Ed/KB1HYS]

I want to remove the Driver transformer from my Valiant and drive the two modulator grids directly with a Paraphase phase splitter. The advantage of the paraphase is that you can get some gain out of it, whereas a plain dual triode splitter has a gain of unity or less.  I've already modded up the audio chain (Timtron Mods mostly) and this would seem to eliminate the limitations of the driver tranny.  Besides, working on radios is fun.

 I've ginned up a schematic of the proposed change. 
Take a look, whatcha think?

<<<---EDITED WITH CORRECTED SCHEMA--->>>

[w4bfs]

this PI circuit is similar to the 'long tailed' phase inverter used in many amp designs ... it also has gain and does not require a tweak ....let me know ifn you need mo info ...73...John

[Steve - WB3HUZ]

http://www.amwindow.org/tech/htm/valiant/valiantmods.htm

[W4EWH]

Take a look, whatcha think?

I'm not clear on why the input to V16 is taken from the plate of the 12AU7 (through the blocking cap), but the input of V17 is taken from the high side of the plate resistor.

Bill, W1AC

[steve_qix]

Take a look, whatcha think?

I'm not clear on why the input to V16 is taken from the plate of the 12AU7 (through the blocking cap), but the input of V17 is taken from the high side of the plate resistor.

Bill, W1AC

That is definitely a bug in the schematic!  Of course, the connection to coupling cap should go to the plate of the tube, and not to the B+.  While we're at it, I would use larger coupling caps.  If you're using resistance coupling anyway, might as well extend the frequency response of the driver.  Maybe a .1 into 200k and not .01.  Where the .01 works into the 1 meg - maybe .05 into the 1 meg resistor.  Just some suggestions 

Regards,

Steve

[WBear2GCR]

Personally I would not go AB1 (resistance/capacitor blocking) coupling in a modulator.
I'd go AB2, since that gives you a bit more output snot, and the benefit of a "smoother" clipping characteristic on peaks...

The problem with the "long tail pair" in general is that the output of the second triode is not automatically equal in amplitude to the input triode. So there are a number of schemes that attempt to rectify this issue. The easiest one is to simply trim the plate resistors. The other way to go is to "balance" the sections at the cathode end, with a "balance pot" - that would have the wiper going toward the resistor to ground and each end of the pot going to the cathodes.

In that schematic, the 220k pot would appear to be for addressing the "imbalance" issue, although I think I have seen that used in a Fender guitar amp schematic with a different value as a presence control, I may be imagining that though...

But I prefer a driver that can sink some current and drive the grids of the tubes positive, running in AB2.
Add one more dual triode and you can do that without a transformer.

                 _-_-bear

[w4bfs]

yes indeed ...cathode followers ...be careful of cathode to fil bkdown V spec ...are you using 6146 as mod tubes? ...there are better choices esp 6550 ...73 ....John

[Ed/KB1HYS]

Ok,  this is why I like to post here!

Well I corrected the Schematic,  I didn't realize I'd drawn it wrong and would propably have built it that way too! 

Anyway I have corrected the error and changed the coupling caps as recomended. 

The 220K pot is there to offset the imbalance induced by the phase splitter.  It probably will effect the audio, but I am going to leave it internal and once set to balance the output not touch it.

When I was researching this idea, a lot of audio amps were mentioned, so I adopted the circuit from one of those.   

I hadn't thought about weather to operate the mod tubes in AB1 or 2.  Interesting point I will have to consider the option. 

finally, I am running the stock 6146s as mod tubes.  I had planned on getting the phase splitter built and working and then changing the 6146s to a 6DQ5 or similar for a little added punch.  I'll look at the 6550 too though.

[k4kyv]

Personally I would not go AB1 (resistance/capacitor blocking) coupling in a modulator.
I'd go AB2, since that gives you a bit more output snot, and the benefit of a "smoother" clipping characteristic on peaks...

Resistance/capacitor blocking would not work with AB2.  It would clip as soon as you start to drive the tube into the grid current region.  AB2 would require either a cathode follower driver, or else transformer coupling (what you are trying to get away from).

[WBear2GCR]

One tube everyone can consider today is the "KT90" supplied via ruskie importers, seems to give more snot than the 6550 will, perhaps even will handle more B+ (not sure have to check that).

I'd reconsider the 220k pot method for "balance" and go for the pot in the cathode first.
The other way to "enforce" balance is with a current mirror supplying the plates, not resistors.
This can be done <gasp> solid state.

Also consider a "White Cathode Follower" arrangement - again an extra pair of tubes, but that's no big deal imho.

          _-_-bear

[Steve - WB3HUZ]

I wouldn't worry much about the esoteric circuits. Any improvement the make will be swamped by the distortion from the mod transformer. Spend your time on the modified-Heising setup as shown in WA1HZK's article.

[Ralph W3GL]

A quick check shows you did not change all the .01 caps...
One still in there from the right plate to a 100 ohm resistor!

Why bother?  Go with what  Dr Huz said!

[WA1HZK]

I use that in every rig I own.
http://www.criticalradio.com/833%20Rig%20Project/Pre%20Amp%20-%20Phase%20Splitter.jpg
Works Pissa - No Limits.
Keith
WA1HZK

[WBear2GCR]

sorry, not a White, VanScoyc (sp?) - it's in Radiotron... "cross coupled phase splitter".

don't have my copy here or I'd get the reference right.

Note how the HZK splitter uses different value of plate resistors to try to balance the two sections...

         _-_-bear

[w4bfs]

hey Keith ... interesting 'universal' line amp to phase inverter ... I am curious about your 600 ohm input stage...unusual to see grounded grid stage here ... I am also using 600 ohm line level (0 dbV Bell standard) but using balanced inputs in a transformerless tube ckt with moderate common mode rejection ... by the way, I'm not trying to stir up anything, just want to see if I missed something ...thanks..73 ...John

[K3ZS]

The NC-183 uses a single tube.    Drives come off of the plate and cathode to drive the push-pull 6V6's.

[Ed/KB1HYS]

Single tube and most double tube phase splitters only provide a gain of unity or less.  I looked at the paraphase circuit because it would provide gain, up to 70volts to the grids of the modulator tubes depending.  Since Didn't want to add a stage to the audio, and just wanted to remove the iron from the circuit, this approach seemed to be a good one.

I was also thinking about changing the finals from 3 6146 tubes to a single 4x150 since I had found a source on line for pulled 4x150s that would cost what a single 6146 is going for now. Ratings were close enough to be workable...  but that was just crazy talk.

Why?? cause it seemed like it would be fun/neat...

[w4bfs]

Ed ... one of the things I looked at was replacing 6146's with 2 4x150 in my JVII. ..about same load Z ... The stock plate xfmr was good Volt'g ( cap input) but not enuff current for higher power level ... be SURE to look at EIMAC application note ... screens must also be modulated from low Z source due to tube internal construction ....good stuff ...John

[Steve - WB3HUZ]

You'll need a blower.

I was also thinking about changing the finals from 3 6146 tubes to a single 4x150 since I had found a source on line for pulled 4x150s that would cost what a single 6146 is going for now. Ratings were close enough to be workable...  but that was just crazy talk.

[Ed/KB1HYS]

Yea Man,  a Valiant wid a Blower.... TURBO Valiant, Cawmon!

[WD5JKO]

Guys,

    Look at the work of Patrick Turner. He has some very innovative tube amplifier circuitry. His phase inverters are mostly hybrid and use a center tapped choke. I find his web site from the land of Oz very interesting.

Jim
WD5JKO

http://www.turneraudio.com.au/poweramps.html

stereo 6550 amplifier using four 6550's per channel:
http://www.turneraudio.com.au/8585-amp-october-2006.html

[w4bfs]

Hi Jim ... thanks for passing along Turners website ... this is one very serious dude !  73   John

[WA1HZK]

Running the input stage grounded grid lets me feed my radios Lo Z and still make use of the original Hi Z gain pots. Works perfect and eliminates input iron or mismatches jamming 600 into 1 meg.
Bread Board it up and take a look at it with yer scope. I have never had any problems with this part of the rigs.
Keith

[WBear2GCR]

Jim,

I know Patrick from several online groups... he's a semi pro audio builder afaik... did not see the circuit you refered to, which amps use it?

                _-_-bear

[WD5JKO]

   Bear,

   Patrick uses a very similar phase inverter on all his push-pull amplifiers. The distinction with his circuits is he uses  BJT CCS along with a center tapped choke. The result is near perfectly balanced outputs without having to trim load resistor values to balance the outputs. Just scroll through the schematics.

  I wonder how well his triode-connected EL-84 (6BQ5) phase inverters would drive 6146 output tubes when driven into AB2?

Jim
WD5JKO