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:  

"Crossover notch" biasing - my final answer!


 
8/31/2000 11:58 PM
Randall Aiken
email
"Crossover notch" biasing - my final answer!
Okay, guys, I've posted a tech paper on my website, which illustrates what I see as the problems inherent in the "crossover notch" method of biasing. It can be seen at:  
 
http://www.aikenamps.com/CrossoverNotchBiasing.html  
 
I would appreciate comments on this paper from others, in particular, from Randy Jamz, who is the most, shall we say, "outspoken" proponent of this method. :)  
 
Please don't take this personally, Randy. It is not my intention to engage in a flame war, nor an attempt to belittle anyone, or feed my ego. As anyone who knows me will tell you, I try to avoid flame wars at all costs, and just stick to the technical facts. I don't buy into "mojo","hype", or doing things "like they did in the old days" without a shred of technical evidence to back it up. Instead, I like to see proof, or at least sound technical reasoning, before I recommend anything. I admit that I got a bit upset at you when you jumped all over Mr. BluesLyne, accusing him of "neurotic silliness" and "not knowing how to use a scope", following a "fad", etc., in an attempt to support your opinion, and for that, I apologize, even though I still feel you were out of line.  
 
I truly want to discuss this topic only from a technical point of view, and see some sort of technical proof from you or someone else on the suitability of the "crossover notch" method of biasing. I'm man enough to admit I'm wrong whenever presented with a valid technical proof to the contrary. So, let's see it!  
 
Thanks,  
 
Randall Aiken
 
9/1/2000 12:15 AM
Matt W.
email

Hey, I like those scope traces. How did you make 'em?  
 
Matt
 
9/1/2000 12:54 AM
Randall Aiken
email

Those are PSpice simulation outputs of the 2-EL34 output stage shown. Much quicker and easier than taking scope photos, and accurate enough to get the point across!  
 
Randall Aiken
 
9/1/2000 12:22 AM
Randy Jamz
email

I don't like to argue either, unless both parties agree to argue facts, and settle on the outcome, regardless of who was 'right'. Sure I was out of line jumping on BluesLyne, but sometimes a 'tough love' approach works best. At least that what I'm told.  
I've also been told the crossover notch method isn't the end-all-be-all method, but it works the best until you have developed a good ear. I am always told in a tough love way that if I can't hear a crossover notch in an amplifier to get get out of the audio business. With current draw ONLY, who knows when the notch disappeared? You'd have to go back and forth until it sounded right.  
I'll read your paper this weekend and get back to you. If I ask questions along the way, bear with me, but I find most of your technical paper a little over my head. I like that though, it takes me out of my comfort zone and back to reading my college books and others from my library. Right now I'm reading a marvelous piece on the pros/cons of single vs. separate cathode bias resistors in a class A amplifier.
 
9/1/2000 12:53 AM
Randall Aiken
email

With current measurement only, you do indeed have the problem of not only not knowing when the crossover notch is gone, but you also don't know if you are at a safe dissipation level. You have to know both current and plate voltage, along with class of operation, and calculate the average plate dissipation to see if you are in the "safe" zone. Just specifying a current number is indeed meaningless, as you suggest.  
 
The point is that the crossover notch is not the end-all to good tone, in fact, if the plate voltage is high enough, you may *have* to live with some crossover distortion in order to keep the tubes alive, unless you wrap enough negative feedback around the amp to remove it. The other point is that if you bias a low plate voltage amp using that method, it will end up with no crossover distortion, but it will be biased far colder than it should for best tone.  
 
There is no one "correct" bias point. Anywhere between full cutoff and the max point not exceeding plate dissipation at any operating point is fair game, and subject to taste. As you well know, adjusting the bias not only sets the idle current, it sets the "headroom" of the output stage as well, and can have a great effect on the tone and overdrive characteristics of the amp. Restricting yourself to the spot where crossover distortion just goes away may not put you in the best spot for tone for your particular style of playing.  
 
Having said that, I have never seen a "unified bias theory" which takes into account plate voltage, plate current, plate dissipation, and reflected load impedance, in order to come up with a hard and fast maximum allowable plate dissipation at worst-case operating conditions. Generally, you see people quoting a figure of between 50% and 70% of max dissipation as the uppermost figure to shoot for. It has been my experience that 50% is a bit conservative, and 70% is closer to the max (also, in looking at the numbers quoted in the RCA receiving tube manual for max and idle plate currents, an even higher point is sometimes used in class AB1 operation). In order to come up with the correct number or formula, the load line and characteristic curves of the tube have to be taken into account, since it is the average dissipation that counts, which is the sum of the static bias dissipation and the average of the superimposed AC waveform, which, in a class AB amp, is greater than zero, and dependent upon the amount of time the tube is in cutoff on each side. One of these days, I'm going to investigate this and try to come up with a better "rule of thumb".  
 
Thank you, Randy, and I look forward to a good discussion!  
 
Randall Aiken
 
9/1/2000 11:22 AM
PeteR
email

Randall,  
Great explanation and logic!  
I always bias by directly measuring the plate current and checking the plate voltage. This gives me a true idea on how the valves are cooking.  
Then I connect a CRO (scope to you guys), measure and calculate the output power at clipping and check for any "parasites" and imbalances. It doesn't take long to do. If there is a crossover problem it will be apparent in this process and dealt with.  
Like me, I'm sure you guys guarantee your work. To let an amp out to someone who is going to electronically "flog" it, without being sure the amp can take it is not only foolhardy but also bad business practice.  
Imagine a doctor treating someone for a heart condition and not bothering to take their pulse and blood pressure. To me the pulse and blood pressure of an amp is the B+/plate voltage and plate current. The palor and appearance is on the CRO. Does this make sense?  
Now where did I put that scalpel?.......Nurse!!!!  
Pete.
 
9/1/2000 2:13 PM
Dave H.
email

quote:
"Those are PSpice simulation outputs"
 
 
Thatís cheating! Itís not the real thing. Itís like playing through a Line6. The simulations are very good. Thatís exactly what I see on a real scope. Iím gonna have to drag myself into the 21st century and learn about this Spice stuff. I thought PSpice was Victoria Beckham.  
 
The EL84s in my homebrew AC30 (350V B+) cut off before it gets to clipping but even without feedback I canít see a crossover kink on the scope until thereís a big flat on the top of the sine wave. On my 5F6A (480V B+) with feedback I can see the crossover kink well before clipping and I canít get rid of it even by biasing the tubes into meltdown so I bias at 70% and live with the crossover. I canít hear it but thatís OK because Iím not in the audio business. (Only joking Randy :)) Why do the two circuits behave so differently? Iíd think it was because the AC30 has nearly 200mA bias and I canít bias at more than 100mA on the 5F6A without melting the tubes. They are both driving a 4k p-p loads so the AC30 is operating closer to ClassÖ.er has both sides conducting for more of the time. Is that the reason 500V amp in the simulaton has more crossover than the 350V amp or is it something else.  
 
Thanks,  
Dave
 
   Page 1 of 7 Next> Last Page>>