Hi Pete,
I'm sorry I missed responding to this post about my transistor models. Thanks for pointing out that pairing discrepency among the 2N5089, 2N5087 and 2N5210. I was wrongly assuming that the 2N5087 and 2N5089 were designated as complements, and overlooked the max voltage ratings. I fell into the trap of often assuming that closely numbered NPN and PNP devices of the same type, e.g., low-noise, are often used as complementary pairs.
The BD139/140 are modeled as the mid-range hfe classification, in accordance with the plots shown in the datasheets.
Sometime soon I hope to post the EKV models for some of the power MOSFETs that currently have VDMOS models posted.
Cheers,
Bob
Low THD amps
Hello Bob
I seem to recall that you prefer single pole compensation for your amplifier topology consisting of input stage and EC output, the reason being that different speaker loads can have an effect on the poles of an amp (effecting stability) and with first order compensation you get a good phase margin and a better chance that the amp is still stable into different loads.
Is this the jist of what you have said on this matter.
Regards
Arthur
Hello Bob
I seem to recall that you prefer single pole compensation for your amplifier topology consisting of input stage and EC output, the reason being that different speaker loads can have an effect on the poles of an amp (effecting stability) and with first order compensation you get a good phase margin and a better chance that the amp is still stable into different loads.
Is this the jist of what you have said on this matter.
Regards
Arthur
dynamic loudspeakers aren't "difficult" loads anywhere near typical power amp O(1 MHz) gain intercept frequency
cable C is more relevant to amp stability that high - and uH output series L protects well
2-pole compensation will usually have the 2nd pole gain break 5-10x below the loop gain intercept and should only "cost" 5-10 dgrees of phase margin at the gain intercept
the ESL inspired uF capacitive load test is simply badly concieved - the step-up xmfr leakage inductance resonates with the panel C below 50 KHz from plots I've seen - to the degree that added series R between the amp and xfmr would be a good idea to control the resonance peaking
the EC amp has local feedback around the output stage - the stability has to be evaluated "inside" all of the feedback loops enclosing the output devices - you will already see "2-pole" loop gain with single pole global compensation
adding 2-pole global compensation to Bob's EC amp would make the loopgain measured around the output devices have a 3rd order gain region - inviting nonlinear instability in clipping
cable C is more relevant to amp stability that high - and uH output series L protects well
2-pole compensation will usually have the 2nd pole gain break 5-10x below the loop gain intercept and should only "cost" 5-10 dgrees of phase margin at the gain intercept
the ESL inspired uF capacitive load test is simply badly concieved - the step-up xmfr leakage inductance resonates with the panel C below 50 KHz from plots I've seen - to the degree that added series R between the amp and xfmr would be a good idea to control the resonance peaking
the EC amp has local feedback around the output stage - the stability has to be evaluated "inside" all of the feedback loops enclosing the output devices - you will already see "2-pole" loop gain with single pole global compensation
adding 2-pole global compensation to Bob's EC amp would make the loopgain measured around the output devices have a 3rd order gain region - inviting nonlinear instability in clipping
Hi Pete,
I did my patent myself. Over the years I have always worked very closely with the corporate patent attorneys at Bell Labs, Bellcore, etc. in preparing my patents, and from that I learned a lot. It took about four years to get through the patent office, however. The patent office can be a frustrating experience, as you often have to teach the examiner much of the art, since a finite number of examiners must cover a great deal of territory. I think it cost about $300 to file and then about $1000 at issuance.
Good luck!
Cheers,
Bob
Hi Arthur,
The amplifier with error correction really doesn't need higher-order gobal FB compensation to obtain exceptionally low distortion, so a first-order roll-off makes sense. I always use a L-R coil to isolate the amplifier from the load at high frequencies. This is especially important with EC.
For amplifiers without EC, I like TMC, which of course is not simple first-order compensation.
Cheers,
Bob
That is interesting, we basically wrote up our patents, handed them off to a review board who then handed them off to an outside legal firm. We supported the attorney handling the patent but they did all the work to push it through.
Hi Pete,
My interaction with the attorneys involved participation in doing the search, writing the spec (the text) and in writing the claims. They took care of all of the paperwork and interactions with the patent office. However, whener there was an office letter asking for changes or questioning novelty or raising other prior art issues, I would get involved.
Cheers,
Bob
Well, I'm pretty much an electronics/audio newb, although I have some basic understanding of things. I'm just wondering how the heck u get any kinda stability without the use of negative feedback? Also, does using negative feedback actually color the sound all that much? Maybe I should read ur book and find out?
. Thinking about it a bit, I can see how the negative feedback might color the sound somewhat while the transistor was stabilizing, but is distortion being added even after everything settles down? If so, it must be minute levels, right? Personally, I'd rather have the stability, or am I missing something here? I am a noob, after all .I also went over to the Dark Side, with four chapters on class D amplifiers
I'm sorry, Bob, but this is unforgivable
. Next thing u know, you'll be asking us to use switching supplies on our high-end amps or something!
Last edited:
"no negative feedback" is a foolish construct of audiophile reviewers and marketing types
they usally mean no Global fedback loop
some insist followers, degeneration aren't negative feedback circuits - despite what those who write textbooks and teach in accredited university EE programs say
bascially all useful audio amplifier circuits use local feedback and most use global feedback
they usally mean no Global fedback loop
some insist followers, degeneration aren't negative feedback circuits - despite what those who write textbooks and teach in accredited university EE programs say
bascially all useful audio amplifier circuits use local feedback and most use global feedback
I think this shows 2nd order loop gain in Bob's HEC amp circuit
http://www.diyaudio.com/forums/soli...terview-error-correction-250.html#post1328496
the behavior of the Tain probe is "interesting"
http://www.diyaudio.com/forums/soli...terview-error-correction-250.html#post1328496
the behavior of the Tain probe is "interesting"
Well, I'm pretty much an electronics/audio newb, although I have some basic understanding of things. I'm just wondering how the heck u get any kinda stability without the use of negative feedback? Also, does using negative feedback actually color the sound all that much? Maybe I should read ur book and find out?
I'm sorry, Bob, but this is unforgivable
Hi AB5NI,
Welcome, newbie's are always welcome. We all started out that way. DIYaudio is a great place to learn. There are a number of good books out there to learn from, and I'd be honored if you read mine and it helped you become more knowledgeable about amplifier circuits.
It is indeed quite difficult to achieve an amplifier with stable dc operating points and gain without negative feedback, but it is important to realize that negative feedback comes in many forms, some a bit disguised. Global negative feedback is only the most obvious. next comes local negative feedback loops within the amplifier. These might take the form of a shunt feedback resistor or a miller compensation capacitor. Then you have truly local negative feedback which includes emitter degeneration and an emitter follower. Amplifiers that claim to use no negative feedback usually do not have any global negative feedback, but virtually always have some form of local negative feedback. Its all a matter of degree, in my opinion.
Cheers,
Bob
That's true of what Bob says... the key is to understand amplifier's and we all started with some sort of amplifier say from text book and things followed on from there...
Even when we think we've good understanding things there's always some 'new' design' that come's along that get's are brains working to understand it..
Even when we think we've good understanding things there's always some 'new' design' that come's along that get's are brains working to understand it..
I'll be sure to pick it up next paycheck, Bob. Can always use another good tech book for the library.
Great info for me, Bob. I guess you could say that, in a perfect world, there would be absolutely no reason for negative feedback, and everything would be stable as hell without the use of it. Actually, I can see this happening in the not-too-distant future, where nano-tech graphene chips are monitoring linearity, with circuitry making minute adjustments to make absolutely sure things stay that way, but that's a ways off yet
. Well, now that i think about it, they'd probably have to use minute levels of negative feedback to stabilize the signal. lmao BTW, are u guys fired up about graphene, and have u considered the ramifications of it's use at audio? I can only imagine what a great amp is going to sound like in the next few years
.Nice typing at ya, Bob, and best regards...
Randy
Well, I am a software engineer by trade, and electronics has always been a hobby for me, picked up from wanting to be like my Dad and be a ham-radio operator
. Been licensed since the 70's now. Wow -- how time flies .Anywho, most of my knowledge has been acquired from the ham mags and from books like: The Art of Electronics, Solid State Design, Radio Frequency Design, RF Circuit Design, the ever popular Amateur Radio Handbook, and various other books that are too numerous to mention. I do realize, however, that because I have this knowledge, it doesn't mean squat in the hifi-audio world
. (I've read Glass Audio and The Absolute Sound on and off for many years, and I know just how picky you guys can be! )Too true, but that's what I really love about this field. Always moving, and if you have a bit of creativity behind you, you might come up with the next do-all, be-all design that rocks the planet
.Anywho, thanks for the warm welcome, guys, and glad that a noob can embrace experts here without feeling alienated.
73 (Best Regards in Ham lingo
),Randy
For sure it is even possible to have both.
The amplifier can use global loop feedback for a DC servo with Fc~300mHz but only use seperate local loops for AF and compensation for multiple stages of amplification, directly coupled.....Would you consider this type of circuit to be labeled as a 'no global feedback' amplifer?