[puppyslugg]

***

 

[ppl]

while most Designers would rather swamp parrisetics rather than utilize them still in no way changes dislike of unnessary compensation capacitence. it is possible to get no ringing a good phase margin and stable operation if the Active Devices are properly chosen and the layout is done with attention to Critical Nodes with additional attention paid to The placement and the souroundings of the gain stages Inverting (-) Input. for simplicity sake Op Amp = gain stage as treated hear.

Eric if you really want your amp to reproduce a 40 KHz sq wave perfectly the you are goiungf to have to use somthing different than a meta board.

 

[eric343]

ppl- yeah, I know, I know... but hey, if I can get close, than that's worth something, isn't it?

Incidentally, I'm not using a META42 board, I'm using an ETA42:

 

[jeffreyj]

Quote:

Originally posted by eric343 Merton- Yes. Jeffrey - what might also be contributing to the ringing/overshoot is that I'm running it into a 1M scope probe as a load- not the <50ohms that the EL2009 requires to be run into if you don't want to use a snubber. So I think I may need to use a snubber... (incidentally, do snubbers go inside or outside the feedback loop?)

What?! All of this time the amp has been unloaded?!

Okay, first off, the scope probe won't cause ringing unless it is not compensated properly - so it presents a capacitive load - and it is applied to a node where stray inductance is the dominating parasitic (forming a tuned circuit is necessary for ringing to occur; always.)

With your amp not terminated into a load - either headphones or a resistor - the only things there are for it to drive are parasitics! The 17nH/in. inductance of the pc board traces and the who-knows-how-much pF/cm^2 combine to make a.... tuned circuit!

Unwittingly, your testing methodology has created the ringing problem when it otherwise might not exist. Go back and test the amp with a load connected - preferably several different pairs of headphones, but resistors of values between 33 and 100 ohms can be used (note that resistors are a much "nicer" load as far as the amp is concerned, so not the best way to test stability).

If the amp still rings, then add a small pF value capacitor across each 4.75k Rf.

section below was edited after re-reading eric343's original post

Snubber network? This is only necessary for the buffer if it will be driving reactive loads (either inductive or capacitve). Unterminated cables are especially troublesome because of their high capacitance which results in a loss of phase margin.

 

[jeffreyj]

Quote:

Originally posted by ppl ... it is possible to get no ringing a good phase margin and stable operation if the Active Devices are properly chosen and the layout is done with attention to Critical Nodes with additional attention paid to The placement and the souroundings of the gain stages Inverting (-) Input. for simplicity sake Op Amp = gain stage as treated hear.

I agree completely with this!

But, like the hoary old cliche, "no battle plan survives contact with the enemy," I believe that no schematic survives contact with the Eric343's of the world! (sorry Eric, couldn't resist

) Which is to say, you can carefully specify the op-amp, layout, passive component selection, etc. to the nth degree - the amp is truly a work of art, now, with all its parasitics balanced or minimized - and then someone comes along, yanks your carefully selected op-amp out, puts a socket in, and starts testing other op-amps. Maybe even "upgrades" the decoupling capacitors from lossy X7R ceramics (lossy=good here because they can't form a high-Q resonant circuit with the pc board trace inductance!) to polypropylene films! Disaster is just around the corner. Now the power supply is being fed by an extremely high-Q resonant network and the op-amp, put into an unplanned-for socket, has several pF of stray capacitance between its (-) and (+) inputs, etc... No wonder the circuit rings like a church bell on a Sunday morning!

This is not an extreme example, either, ppl, and you know it! But it illustrates well why my philosophy is to limit bandwidth, swamp strays and parasitics, etc., so that the circuit trades ultimate performance - yes - for better resiliency and behavior in the face of unanticipated conditions. Note that I do not necessarily believe that my way is "better" than your way, just different.

 

[ppl]

jeffreyj> Your Theroy is classic textbook and is what EE students are taught in school. We must consider all the possible unknowens. unfortunatly DIY Designs do get cfhoped up by various builders and Op amp Rolling that is so popullar on these Forums. can and dose degrade alot of the designs that were perfectly Designed with ample consideration for the Knowen parrisitics.

With opamps of similer Bandwidth and open loop gain no problems uasualy happen. and if the same type of input stage is used Bipolar or FET the Input Capacitences are again probaly similar. In these cases the swap should not result in any problems of a well designed circuit.

The opereative word hear is should not. Murphy's Law always comes into play on technical matters and comes under what can go wrong will go wrong. I have been a long perponent of using spicific part number"s for IC's in a Amplifier The chosen devices having been tested under all possible operating conditions in the Actual circuit (Prototyping)

With any Amplifier Topology one can only asume that hopfully the Designer did a good job on selecting the Given device base both technical Performance and the suitability for the Application In addition to sound quality.

Any change from the builder from these will affect both the performance and sound of the Amp. I have sugested from time to time that in my designs only approved devices be used. This message has met stiff resistence from the Everyman's Amp Crowd.

In conclusion I can easaly understand the rational of swamping with additional compensation capacitence. From purly technical considerations this is safe. For my self, I like to optimise the Circuit to perform and sound good without having to resort to external compensation. In a properly funtioning Amplifier these caps always seriously degrade the sound and for some strange reason also negitivily affect the Bass. Now why a few pF of capacitence can inpact frequencies that low is beyond me but i have on nurmerious times heard this effect. On a non properly performing Amp similer to eric's Silbelense will result from an underdampened condition with hash and grain resulting from Gross overshoot.

In Eric's case He is still learning and is incredibly sharp in comprehention and can understand things very well. I think that elimitating the ringing from his Amp under discussion hear will pay big rewards later on since he will then reconise the symtom and will have formed a method of attacking the issue should it come up again later. IMHO people like Eric should be encouraged to persue perfection and not settle for the Easy answer but futher refine there technique.

 

[jeffreyj]

ppl: Yep, I admit that my recommendation is one you'll find in most any engineering textbook. It's also a lot more practical than making a dozen or so pc boards to get the layout just right for one op-amp/buffer combination.

 

[eric343]

Quote:

Originally posted by ppl In Eric's case He is still learning and is incredibly sharp in comprehention and can understand things very well. I think that elimitating the ringing from his Amp under discussion hear will pay big rewards later on since he will then reconise the symtom and will have formed a method of attacking the issue should it come up again later. IMHO people like Eric should be encouraged to persue perfection and not settle for the Easy answer but futher refine there technique.

Thanks, man, that made my day!

 

[ppl]

jeffreyj> Yes I admit that my prototyping is more involved than what is normaly considered reasonable. However ,This is the principal effect that will arise from the quest for true fidelity. I feel rather silly redirecting your attention to the title of this site. We are far from attaining this goal ,but the knowledge that is required is not complete in the text books of our time. So I would put to you this question....When and where will this new defenition of Hi fidelity arise.

Eric> Glad it made your day. Just the facts, good job

 

[jeffreyj]

Quote:

Originally posted by ppl jeffreyj> Yes I admit that my prototyping is more involved than what is normaly considered reasonable.

We weren't discussing what method of prototyping was best for you, ppl, we were discussing what method was best for Eric. I think it is questionable, at the very least, to advise a "beginner" in any endeavor to discount the acquired wisdom of all who came before and instead re-invent the wheel as many times as is necessary to "get it right." Your advice to Eric to ditch the shunt capacitors was impractical and I called you on it.


Quote:

I feel rather silly redirecting your attention to the title of this site.

So let me get this straight: you state that ringing results in "grain and harshness," so, presumably, my "textbook advice" which eliminated it would be right and proper in the overall aim of achieving high fidelity. But then you imply that it is not compatible with high-fidelity in the above quote. So which is it, ppl?

Quote:

We are far from attaining this goal ,but the knowledge that is required is not complete in the text books of our time. So I would put to you this question....When and where will this new defenition of Hi fidelity arise.

So you are single-handedly extending the state of the art beyond what is in "text books of our time", eh? Uh-huh. Right. You, for example, are more of an authority on analog circuit design than Bob Pease, then? Can out do Douglas Self when it comes to designing amplifiers? Move over Horowitz and Hill, ppl's got a new Art of Electronics he'd like to share with the world?

Yeah, right.

A lot of the people here won't know any better when they see, for example, the multiple current sources feeding the power-on LED in the PPA. I recognize it for what it truly is, ppl: needless complexity in the hopes of making a unique circuit, not necessarily a better sounding one.

Regards
The Objective One

 

[eric343]

If you don't mind my butting in here, since ppl's given me an ego boost I'll return the favor and save him some time rehashing what he's already said...

Quote:

Originally posted by jeffreyj So let me get this straight: you state that ringing results in "grain and harshness," so, presumably, my "textbook advice" which eliminated it would be right and proper in the overall aim of achieving high fidelity. But then you imply that it is not compatible with high-fidelity in the above quote. So which is it, ppl?

I believe the answer would be neither; namely, to design the board and circuit in such a fashion that there's no ringing, even without a compensation capacitor.
Quote:

So you are single-handedly extending the state of the art beyond what is in "text books of our time", eh? Uh-huh. Right.

More or less

The problem here is that textbooks ain't everything... some things come from experience. Keep in mind that textbooks are written with the goal of teaching electronics, not how to design high fidelity audio circuits... It's a common trap; having learned a whole bunch of stuff from one source, you then treat it as the be-all end-all Spring of Knowledge against whose authority none may go without incurring your wrath.

Except they AIN'T the Spring of Knowledge, and there's stuff to know that ISN'T in them.
Quote:

You, for example, are more of an authority on analog circuit design than Bob Pease, then? Can out do Douglas Self when it comes to designing amplifiers? Move over Horowitz and Hill, ppl's got a new Art of Electronics he'd like to share with the world?

Acutually, that would be quite cool; I'd love to see ppl write a book on electronics

In any case, I don't see anyone claiming to out-geek any gurus, nor do I see those gurus taking sides in this debate. Quote:

Yeah, right. A lot of the people here won't know any better when they see, for example, the multiple current sources feeding the power-on LED in the PPA. I recognize it for what it truly is, ppl: needless complexity in the hopes of making a unique circuit, not necessarily a better sounding one.

Actually, I would suspect that the current sources feeding the LED are ideas of Tangent or Morsel, not ppl - they're not designed to be used simultaneously; they appear to be in there for versatility. There's a similar arrangement in the META42 board...

 

[maczrool]

Quote:

I'd love to see ppl write a book on electronics

I don't envy the poor soul who has to proofread/edit it!

Stu

 

[jeffreyj]

Quote:

Originally posted by eric343 The problem here is that textbooks ain't everything... some things come from experience. Keep in mind that textbooks are written with the goal of teaching electronics, not how to design high fidelity audio circuits... It's a common trap; having learned a whole bunch of stuff from one source, you then treat it as the be-all end-all Spring of Knowledge against whose authority none may go without incurring your wrath. Except they AIN'T the Spring of Knowledge, and there's stuff to know that ISN'T in them.

I agree completely and haven't stated anything contrary to this. Now, ppl certainly derided my suggestion to you as a "textbook" solution, but that was his opinion, not actual fact.

His suggestion to keep tweaking the board layout until the ringing went away was technically sound, but not terribly practical with respect to you, was it? That was the whole reason I suggested you put a capacitor across the feedback resistor: it was guaranteed to work, quickly, if not optimally. It is my opinion that it is best to start with a working circuit and then optimize layout; trying to go about it the other way around is like looking at a map after you've reached your destination.

 

[Tomo]

Hey,

Testbooks aren't everything. I had to read all the "texts" available to me. ie Website, Journals, Articles, Scraps. You can learn extraordinary quantity of knowledge from meticulous reading.

I have collected over 100 MB of documents on electronics over few years. I have hardcopy of about a quarter of them and it weighs like 6 kilograms, almost 10th my bodyweight. Yeah, even a dumb guy like me can accumulate load of intels. I think significant amount of them cannot simply obtain by few years of experience (which is tiny in comparison to Walt Jung or Jan Didden who are professionally involved.)

So I wouldn't say experience is everything. Heck I don't wanna build amps if I had to train for 20 years. But fact is that you can learn great deal from materials without actually experiencing.

I think we are too much into simply experience oriented. We could use some orientations where we can collectively learn stuffs. Perhaps invite some DIY masters like Mr. Pass to give a talk few times a year? Perhaps young ones (me!) to post some paper? ... DUDE I SOUND LIKE A GRAD STUDENT ...

Tomo

 

[binary_digit]

Quote:

Originally posted by jeffreyj We weren't discussing what method of prototyping was best for you, ppl, we were discussing what method was best for Eric. I think it is questionable, at the very least, to advise a "beginner" in any endeavor to discount the acquired wisdom of all who came before and instead re-invent the wheel as many times as is necessary to "get it right." Your advice to Eric to ditch the shunt capacitors was impractical and I called you on it. So let me get this straight: you state that ringing results in "grain and harshness," so, presumably, my "textbook advice" which eliminated it would be right and proper in the overall aim of achieving high fidelity. But then you imply that it is not compatible with high-fidelity in the above quote. So which is it, ppl? So you are single-handedly extending the state of the art beyond what is in "text books of our time", eh? Uh-huh. Right. You, for example, are more of an authority on analog circuit design than Bob Pease, then? Can out do Douglas Self when it comes to designing amplifiers? Move over Horowitz and Hill, ppl's got a new Art of Electronics he'd like to share with the world? Yeah, right. A lot of the people here won't know any better when they see, for example, the multiple current sources feeding the power-on LED in the PPA. I recognize it for what it truly is, ppl: needless complexity in the hopes of making a unique circuit, not necessarily a better sounding one. Regards The Objective One

I happen to know for a fact that ppl was refering to the eric343s of the world in his reference to re-defining the state of the art. To make such a statement when refering to oneself would be pompous and arougant. Those who have met with ppl in person can say he is far from from that.

I own one of ppl's earlier amps and have asked him about the current sources for the LED. It is quite simple really . The zener constant current source keeps the LED at a constant brightness from full charge to a pre determined low voltage shut off point apposed to a constant slow dimming as depletion occures. At this point the batteries are getting to low and should be plugged in. If you don't then you will begin to experience clipping distortion from normal to high volume levels in a short amount of time.

This is a simple way of ensuring a enjoyable hi fidelity listening experience and possibly prevent you from draining your NiMH rechargables to the point of damage. Not just an on / off indicator anymore. A zener and CRD is all that is required. Hardly complex

PPL: Ensuring high fidelity through his patented smart LED technology.

Sorry, couldn't resist

I don't think this many intelligent people would listen to ppl (the P in PPA and Silent P in META 42) if he were simply a gimmic man.

I do not wish to upset anyone but I feel very strongly about feeding energy to those you may depend apon for knowledge that is considerably above your own. PPL is a wonderfull resource for real world highly advanced Audio tech. I am sorry The replies to your post may seem argumentative jeffreyj but I assure you that is not my intention.

Peace All