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!|
|2/9/2005 8:45 PM|
||Re: Does anyone use Paraphase inverters?|
I believe that the LTP has better balance and more gain than either of the paraphase ones you mention. It sounds like they are talking about the two different ones Fender used in the past in that article you are talking about. The self-balancing paraphase will balance better than the other one, but the LTP is better than both and gives more gain too.
|2/10/2005 1:33 PM|
||Benefits of better balance.|
What are the benefits of better balance???
|2/10/2005 2:16 PM|
Less distortion - I guess.
Of course, that's not necessarily the best thing for a guitar amp!
The tone will be different - it's all taste. Different PI schemes are part of why a 5D3 Tweed sounds different from a 5E3, or a DR for that matter...
|2/10/2005 3:29 PM|
Any phase inverter you make with a 12AX7 has pretty much the same gain. When you are dealing with a differential amplifier or one that goes from single ended to differential, you have to add the outputs from both sides so, the cathodyne PI really has a gain of (almost) 2. Put a common cathode stage in front of it and you're close to any other PI. The main difference is the driving impedance is lower, the headroom is less, and feedback is kind of a pain. Ever see anyone use a presence control a cathodyne PI ? It would take a 250 ohm pot and a 10uF cap. Not completely impossible.
The PI like the 5C3 probably has the most distortion and poorest balance because the lower side is just inverting the top side with no local feedback (this may be a good thing). The 5D3 balances a little better because of the local feedback. The LTP and cathodyne PI's have the best balance mostly dependent on resistor tolerance. I like Aiken's approach but I power the tail resistor from the negative bias supply and blow off the input capacitors like some of O'Connor's designs. The next time I build one, I may use an active current source for perfect balance.
|2/10/2005 6:20 PM|
I've seen them vary from about .8 to each output, to 125+ to each output (which I thought was pretty impressive for a 100-mu tube).
|2/10/2005 9:39 PM|
You can't measure the gain of a PI when is is inside a feedback loop unless you measure differentially across the inputs and outputs. Try this: Connect an amp to it's intended load impdeance. Apply a sinewave to about 1/4th the rated output and measure the signal at the grids of the output tubes. Now change the load to 1/2 the rated impedance. The signal at the output tube grids will get bigger even though the signal at the load gets smaller. Did the gain of the PI change ? No, feedback is trying to make the same voltage across the lower impedance and that takes more drive to the output tubes.
In most power amps that use an LTP PI with feedback, the signal between the grids of the PI tube is only about 1/10th the amplitude of the input applied from the preamp (at low frequencies).
|2/11/2005 6:42 AM|
Well, you can compare the single-ended input signal (referenced to ground) with the single-ended output signal (referenced to ground), and this will give you your gain across a single half of the PI, NFB or no NFB. Obviously individual stage gains should be measured with any NFB loop(s) open, as should the open-loop gain of the circuit to which NFB will be applied, and then your closed-loop gain and stability margins calculated with NFB applied (assuming one wants to use a NFB loop at all, which I surely don't).
How did we get started on NFB and loop gains, anyway? I was talking about the very wide variation in 12AX7 PI-circuit gains, and after doing some checking I found that at least one single-12AX7 PI can provide a voltage gain of almost 800 narrowband, and 250+ wideband (bandwidths not specified, but I'm assuming "wideband" refers to the standard 20-20kHz hi-fi response).
|<<First Page||<Prev||Page 2 of 4||Next>||Last Page>>|