negative feedback strategy in tube line amps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

funkydiplomat

Well-known member
Joined
Jun 6, 2004
Messages
76
Location
Roanoke, Virginia. USA.
hey guys. i've been reading up on preamps and such. I've got a couple of books and try to read most of the threads on here pretty regularly, i've scetched a couple of circuits of my own, but i'm looking for a bit more about how people employ negative feedback in their amp circuits. NYdave scanned in an article a while back about it. I read it. It makes sense when i read it, but i'm curious how you people or other designers put it into action. i often see it expressed as a percentage. Is there a common formula or strategy than most use when working it into their circuits? If anyone has any suggestions on a few other articles i could read, or a different text i could pick up, i certainly would apreciate it. Or... if you think you could explain it with a post, that would be really cool too. c'mon and learn me somethin'!
 
Feedback can be a pretty deep subject--read papers by Nyquist, or H.S. Black. But I've found that following a few simple guidlines can keep me out of trouble most of the time when trying to apply negative feedback to my designs.

First of all, it's generally better to start out a design with feedback in mind, rather than trying to graft it on to an existing circuit. The paper I posted addresses the issues of paying attention to your phase shifts and open-loop frequency response to ensure stability when negative feedback is applied.

I like the open-loop gain of my circuit to be at least 20dB higher than the desired closed-loop gain. Although it's not conclusive that it applies in all cases, there is some published evidence that using too little negative feedback can cause problems more objectionable than those it solves.

I make an exception to my self-imposed "20dB rule" is guitar amps, since in that case I'm after a certain tonal character and yes, interesting distortion, rather than absolute accuracy. A few dB of feedback can give a guitar amp circuit some useful characteristics. I'll get into that later.

As you know, the amount of negative feedback is usually expressed as "so and so dB." This refers to the reduction in gain with feedback applied. For instance, suppose I have an amp with an open-loop gain of 1000 (60dB). If I add enough negative feedback to reduce the gain to 10 (20dB), then I have 40dB of negative feedback.

I try to pay attention to the number and location of poles in the feedback network. I aim for an open-loop frequency response that rolls off gracefully outside the frequency range I want to amplify (again, refer to the paper I posted). Rolloffs that start about an octave above and an octave below the desired frequency range have been working well for me so far. Oftentimes, the input capacitance of the active devices (e.g, Miller effect) takes care of the high end rolloff for you, so then it's just a case of sizing your coupling caps appropriately. Sometimes you need to add R-C networks to increase the HF rolloff. It really depends on the circuit and the devices you're using.

I generally avoid inductors or transformers inside the feedback loop. When including a transformer inside a negative feedback loop (such as in the case of a power amp), you may have to compensate the network. Voltage leads current in an inductor, and current leads voltage in a capacitor, so sometimes a small cap is connected across the feedback resistor to try to even it all out. This isn't always necessary; it depends on the transformer. Many guitar amps use cheap output transformers and uncompensated -fb loops and get away with it.

Gain controls (such as a volume control between stages) should not be used inside a feedback loop. Altering the volume control alters the open-loop gain and frequency response of the circuit. Ditto with tone controls.

The feedback network acts as a load on the output of the circuit, and it should be accounted for when calculating or measuring the open-loop gain of the circuit. Ideally, the feedback resistors should be sized so that they do not load down the output of the circuit. This isn't always easy when parts of the feedback network serve another purpose, such as setting DC conditions (e.g., emitter or cathode resistors).

Negative feedback reduces output impedance, but don't be fooled into thinking that it will allow a circuit to give a greater output into a given load. The maximum output is limited by whatever current the tube/transistor/whatever can supply. The benefit of the lower output impedance occurs mostly at small signals; the absolute clipping point will stay about the same, although distortion will be lower with negative feedback until the clipping point is reached, where the transition is fairly abrupt.

This happens to be why I like guitar amps with negative feedback around the output stages; the transition is more abrupt, so I can go from clean to heavy metal with just a manipulation of the volume control on my guitar. No-feedback guitar amps have a very wide "in between" region, which is great for blues, but not so great for some other styles.
 
NYD

I did not see UL in your post. UL with no global feedback is kind of like a 3rd basic guitar amp. I understood you post about liking feedback amps with guitars.

I think we kind of think the same, it is the whole system when it come to using or not using feedback If I am not reading to deep into your post
 
That's the Bunny!------->
rdh4_ch7logo.jpg



Download-right click-save link target as-

http://vacuumbrain.com/The_Lab/TA/RDH4/CHAPTR07.pdf

Warning! 5 mb pdf.

Let me know if too slow and I will zip it up.

cj
 

Latest posts

Back
Top