ampage
Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. New: view Recent Searches.
New: visit Schematic Hell!
The sunn still shines online!

 
Listen to great tunes streaming live right now!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  

Building a 'clean' PI


 
5/31/2000 10:05 PM
Chris Harden
email
Building a 'clean' PI
On my amp I'm currently working on the PI, trying to get the most headroom I can. I've removed one of the gain stages so that the preamp is clean, but I only need to turn the master up slightly before I get distortion, almost certainly on the PI. I'm using a modified 12AT7 see-saw arrangement with 56k anode loads, a common 470R cathode resistor with a 220uF bypass, and a 300V B+. The stage seems to be running very cold at only 5mA for both triodes together, but I can't seem to get it to run any hotter, I think I need to have it at around 15-20mA combined anode current. Any ideas on increasing the headroom availiable, as I want to be able to play super-clean as well as OD.  
 
Chris
 
6/1/2000 4:23 PM
Doc
email

Chris:  
 
You didn't mention anything about the rest of the amplifier, its tube lineup, or its intended use except that you want the best of both worlds, to play super clean and overdrive. Is there a feedback loop? I'm not certain what you mean by the term "modified see-saw inverter". With 5ma through a common 470R cathode resistor, grid bias voltage is only -2.35v. That will saturate easily, so your preamp output signal needs to stay below that level to have the PI remain clean. You may need to eploy some kind of master volume control to keep the incoming signal level down. MV's will alter the sonic character a little.  
 
For decent headroom and better linearity, you might try the long tail inverter circuit. For maximum headroom, but lower stage gain, you could employ a lower-mu tube such as a 12AU7, 6CG7, or 6SN7 rather than the 12AT7 or 6SL7. Run the stage at a high B+. A 6SN7GTB can handle up to 450v plate to cathode.  
 
There are some excellent tech articles on the Aiken Amplification website. In the advanced section, you will find a how-to design article on a long tail phase inverter.  
 
The split load inverter can give good results in balance, linearity and headroom, but requires some sort of gain stage ahead of it. Two triode sections will do it. You could direct couple the gain stage's output to the PI's input grid for a circuit exhibiting lower phase shift than with R-C coupling. You might borrow a circuit from a successful hifi amp, such as the Eico HF87 or Dynaco. The Dyna uses the split load inverter (the triode section of the 7199 is like a 12AU7) and the Eico uses the 6SN7 long tail. Or, you could borrow the proven Fender SF Twin Reverb circuit which uses the 12AT7, but can be adjusted to run with lower mu tubes.  
 
These are just my suggestions, and by no means the last word in phase inverter selection.  
 
Doc
 
6/4/2000 5:54 PM
Ray Ivers
email

Chris,  
 
I just finished running scenarios thru TubeCAD, using both the 12AT7 and 12AU7 tubes, and, well... it seems that at 300 volts B+, you need to use such low values of plate load to bring the current up that you lose gain, headroom, and signal symmetry, which granted may not be that much of an issue for a push-pull driver tube - you might be better off with a 12AX7 driver. A negative supply for the common cathode resistor might help, but I admit is a little bit of a weird fix. I'm assuming you're using what TubeCAD and I call a 'differential amp' configuration - two common-cathode gain stages sharing a cathode resistor. With see-saws, cathodynes, split-loads, concertinas, paraphases, long-tails, bootstraps, and other such floating around out there, I get a little confused sometimes. I did notice that all the output peak AC numbers are far above any negative bias values that could possibly be in use, so I don't think that driver-tube headroom will be an issue in this configuration at all.  
 
Ray
 
6/4/2000 10:44 PM
Chris Harden
email

I've only just found this post, because the nice people at AMPAGE have had a reorganisation and spit the guitar amp section up. Nevermind, I'm here now.  
 
Firstly, the circuit I'm using. I don't know the name for it so I'll explain as best I can.  
 
Take both sections of a 12AT7 and connect the cathodes to a common cathode resistor. The outputs to each O/P tube come from the two anodes via caps. One gate is connected to the output of the previous stage (via master volume), the other is connected (via cap) to a potential divider across the outputs of the two sides (the inputs to the O/P tubes). It's similar to a long-tailed pair except for the spare grid being connected across the outputs rather than to ground, which should give better symmetry, because as soon as the sides are un-equal, a signal appears at the spare grid to correct it.  
 
I have a blank sheet to work with, you mention the B+ being a little low, what would be better? I'm going over to EL34's soon, which will be run at 550V, but as I think I mentioned, I'm waiting for parts.  
 
The rest of the preamp at the moment is as follows, 1/2 12AT7 I/P stage, tone stack, 1/2 12AX7 common cathode, master volume, PI. I have more stages to add, but I'm working on the clean tone first.  
 
I've also been playing with a current source instead of the cathode resistor in the PI, I may have mentioned earlier. I changed it for a resistor when the circuit wasn't working right, but it didn't seem to make much difference.  
 
Thanks for your thoughts  
 
Chris
 
6/5/2000 2:59 AM
Ray Ivers
email

Chris,  
 
I would call that circuit a 'self-balancing paraphase' inverter, but it may well have the name 'see-saw', as well. I've always thought that circuit would be more popular in guitar amps, for the very reason it isn't in hi-fi amps (potential for more distortion). It's a hi-gain, hi-output-voltage design, and allows easy adjustment of AC output stage balance. At 550 volts B+, a 12AT7 should be pulling almost 380 volts peak-to-peak AC with 56K plate loads, 220K output tube grid resistors, and a 470 ohm common cathode resistance - headroom city! Current draw is still a little low at 8.6 mA, but it's better than it was.  
 
Ray
 
6/5/2000 7:50 AM
Bruce
email

Ray, have you tried running that sim on the 5965 tube?  
Bruce
 
6/5/2000 12:24 PM
Ray Ivers
email

Bruce,  
 
That's a cool tube! I ran it thru TubeCAD and it seemed to like the lower B+ voltage (300 V) more than the others. I ended up with 33K plate resistors, 470 ohm common cathode resistor, 7.5 mA total current, for -118/+129 peak-to-peak AC out. I was able to go as high as 500 volts B+ before I started hitting problems with this tube; the voltage across the tube became excessive, so I had it pull more current, which caused high plate dissipation, etc. (I didn't want to just keep increasing the plate resistor). The 500 volt numbers were 39K plate resistors, 560 ohm cathode resistor, 10.5 mA, and 1.52 watts plate dissipation; signal swing was -200/+227 volts peak AC!  
 
Chris - this is an excellent tube to consider for your driver. Antique Electronics lists them for $2.75 each. The British military designation is CV5843.  
 
Ray
 
   Page 1 of 2 Next> Last Page>>