| till |
"When the amps is done properly, Zobel represents no real advantage."
what leads direct to the question if chipamps are done properly. Maybe they are only a compromise to get 50W from a single die. Some kind of comparator trimmed down with loads of feedback to behave like an amp. |
|
|
| Fred Dieckmann |
And now..... the rest of the story.
From the same LM3875 data sheet:
"External Components Description
(Figure 1 and Figure 2)
Components Functional Description
1. RIN Acts as a volume control by setting the voltage level allowed to the amplifier’s input terminals.
2. RA Provides DC voltage biasing for the single supply operation and bias current for the positive input terminal.
3. CA Provides bias filtering.
4. C Provides AC coupling at the input and output of the amplifier for single supply operation.
5. RB Prevents currents from entering the amplifier’s non-inverting input which may be passed through to the load upon power-down of the system due to the low input impedance of the circuitry when the under-voltage circuitry is off. This phenomenon occurs when the supply voltages are below 1.5V.
6. *CC Reduces the gain (bandwidth of the amplifier) at high frequencies to avoid quasi-saturation oscillations of the output transistor. The capacitor also suppresses external electromagnetic switching noise created from fluorescent lamps.
*Optional components dependent upon specific design requirements. Refer to the External Components Description section for a component function
description.
Components Functional Description
7. Ri Inverting input resistance to provide AC Gain in conjunction with Rf1.
8. *Ci Feedback capacitor. Ensures unity gain at DC. Also a low frequency pole (highpass roll-off) at:
fc = 1/(2π Ri Ci).
9. Rf1 Feedback resistance to provide AC Gain in conjunction with Ri.
10. *Rf2 At higher frequencies feedback resistance works with Cf to provide lower AC Gain in conjunction with Rf1 and Ri. A high frequency pole (lowpass roll-off) exists at:
fc = [Rf1 Rf2] (s + 1/Rf2 Cf]/[(Rf1 + Rf2) (s + 1/Cf (Rf1 +Rf2))].
11. *Cf Compensation capacitor that works with Rf1 and Rf2 to reduce the AC Gain at higher frequencies.
12. *RSN Works with CSN to stabilize the output stage by creating a pole that eliminates high frequency oscillations.
13. *CSN Works with RSN to stabilize the output stage by creating a pole that eliminates high frequency oscillations. fc = 1/(2πRSN CSN).
14. *L Provides high impedance at high frequencies so that R may decouple a highly capacitive load and reduce the Q of the series resonant circuit due to capacitive load. Also provides a low impedance at low frequencies to short out R and pass audio signals to the load.
15. *R
16. CS Provides power supply filtering and bypassing.
*Optional components dependent upon specific design requirements. Refer to the Application Information section for more information.
Attached image below
FIGURE 2. Typical Single Supply Audio Amplifier Application Circuit
The optional external components shown in Figure 2 and described above are applicable in both single and split voltage supply
configurations."
Please read the WHOLE data sheet, there is more on stability and layout. |
|
|
| jean-paul |
OMG, again a repetition of that datasheet. We can read and we can see the Typical application WITHOUT Zobel and we know it can oscillate and we know you can add optional components like a Zobel.
We know, all people know on the forum, even the tube guys started using Zobels afraid of having oscillations. I added a Zobel in my TV intimidated by the horror stories about oscillations of the picture tube. My bicycle has a Zobel too to avoid fluctuations and ringing of the lightbulb.
Please Sy, please. |
|
|
| till |
"BTW, who was or is Mr. Zobel? Anybody who knows?"
was a Mr. Dr. Zobel at Bell Labs according to http://www.rane.com/par-z.html
Jean-Paul if you can´t stand it, why do you want to close the thread? you always have the option to stay out. |
|
|
| Fred Dieckmann |
| quote: | Originally posted by Peter Daniel
Perheaps Pedja's post in the other thread was to subtle for you to comprehend?
So here I'm putting it out for you in layman terms. If you still argue that National recommends Zobel for LM3875, I might suspect that you are doing it for reasons other than your usual care for members trying to "understand some basic amplifier issues".
I assume you know what word "Typical" means. |
This is the most disingenuous thing I have even seen you do. Pick out one figure and conveniently ignore the rest of the data sheet. Are you trying to KEEP members from understanding some basic amplifier design issues. At least you have let go of me and are dragging yourself through the mud. Do you really think this is wise? I really don't understand what you are trying to accomplish. Would you tell us please?
Will the moderators please split the thread so we can explore this topic which is off topic for this thread? |
|
|
| Peter Daniel |
| quote: | Originally posted by Fred Dieckmann
Please read the WHOLE data sheet, there is more on stability and layout. |
I can also read: "*Optional components dependent upon specific design requirements. Refer to the Application Information section for more information."
My experience show that they are not required in my application. You have no experience, as you've built only one amp.
EDIT: I would have post it in the other thread, but then it would be out of context. I leave it to moderators to use the most suitable approach. |
|
|
| Christer |
| quote: | Originally posted by till
"BTW, who was or is Mr. Zobel? Anybody who knows?"
was a Mr. Dr. Zobel at Bell Labs according to http://www.rane.com/par-z.html
Jean-Paul if you can´t stand it, why do you want to close the thread? you always have the option to stay out. |
Whoops, that the info I linke to previously. I missed it said
anything about Zobel himself. Still, it says nothing about
what he did, but but, that's maybe not interesting anyway.
One point in closing the thread is that it is now suddenly
about Zobel filters, something which is already being
discussed in two other threds which are dedicated to that
topic. |
|
|
| till |
"One point in closing the thread is that it is now suddenly
about Zobel filters,.."
Thats no reason for closing it, as the main topic is not finished.
Its a good reason for splitting it and clean it up. But this obviosly would be more work for the moderators. And me thinks they may have some interest in the main topic not discussed anymore.
O. J. Zobel, "Theory and Design of Uniform and Composite Electric Wave Filters",
Bell Tel. Sys. Tech. J., vol. 2, pp. 1-46 (1923 Jan).
from http://www.trueaudio.com/st_zobel.htm |
|
|
| Christer |
| quote: | Originally posted by Peter Daniel
My experience show that they are not required for my application. You have no experience, as you've built only one amp. |
There is a very big difference between doing empirical evaluation
on a limited number of load situation and doing a proper
theoretical worst-case analyses. Many have found that out
the hard way, both in electronics and other areas. I think that
is what Fred is trying to tell. One of the most famous cases
is a missing blank in a Fortran program that caused a
satellite to crash. But no big deal, that was only about a
billion dollars thrown in the trash can. :) |
|
|
| Peter Daniel |
| quote: | Originally posted by Christer
There is a very big difference between doing empirical evaluation
on a limited number of load situation and doing a proper
theoretical worst-case analyses. Many have found that out
the hard way, both in electronics and other areas. I think that
is what Fred is trying to tell. One of the most famous cases
is a missing blank in a Fortran program that caused a
satellite to crash. But no big deal, that was only about a
billion dollars thrown in the trash can. :) |
I still feel I have the right to be able to do it my way, not according some guy on a forum I have never met before. Even if I have to learn it hard way. |
|
|
| Christer |
| quote: | Originally posted by Peter Daniel
I still feel I have the right to be able to do it my way, not according some guy on a forum I have never met before. Even if I have to learn it hard way. |
Sure you have. I actually think Fred is just trying to be helpful,
warning you of potential future disaster with complaining
customers. |
|
|
| Pedja |
Moamps,
I am not sure I have understood… It is not a problem if there are a few ways in which Zobel works and a few posts back I just have wrote that.
My last post was intended just to inform what I have found to be the logic of the National's datasheet writer. It was just an effort to clarify “Read the datasheet!” hassle.
Recommendations found in the datasheets may and may not have something with the internal circuit, but I can not comment (or surmise) on this since none TDA2030 datasheet I have (4 pcs) shows what is inside.
Btw, as can be seen from your screenshot, it is claimed that the ineffectiveness of this network will deteriorate stability under an inductive load. True, the comment for the C7 (that is the Zobel cap) can be understood as it refers to either load but it also may refer to not load at all. If I recognized well, that is from the datasheet from 1998 and here the Typical Application circuit uses single supply. Note that in the datasheet from 2000, in which the Typical Application uses split supply (though it comprises the band limiting inside the feedback loop which certainly results in the nicer phase margin), has next comment for the situation in which this cap is smaller than recommended: “higher bandwidth”.
Pedja |
|
|
| fdegrove |
Hi,
| quote: | | When the amps is done properly, Zobel represents no real advantage." |
Sorry but that's just not true.
This is precisely what Fred and others have been trying to put across, you can have a very stable amp and still run into trouble given extreme circumstances.
Heck Peter, you should even recall the trouble someone had with a certain speaker cable and a transformer coupled tube amp not too long ago....
Not that a Zobel network behind the OPT could have saved that one IMO but still...
Better safe than sorry?
At the end of the day it's up to you to include it or not, it doesn't cost an arm and a leg and hey, you can always include a switch. If you gain enough market confidence you could eventually leave it out of the production units, no? Not that I would but I suppose you catch my drift, right?
Just my two lousy cents,;) |
|
|
| Peter Daniel |
| If I encounter a single case when Zobel is required, THEN I will include a switch. Everybody on a forum will know about it too.;) |
|
|
| Christer |
| quote: | Originally posted by Peter Daniel
If I encounter a single case when Zobel is required, THEN I will include a switch. Everybody on a forum will know about it too.;) |
Yes, but what if it is too late? What's you budget for refunds
for customers blown amps and speakers? No, I am not asking
you to reveal business secrets, just make you ask yourself
that question. |
|
|
| Peter Daniel |
| I'm not really concerned with blown amps (as they are easy to fix). I will not comment on blown speakers, as some might take advantage on my approach. But I would definitely do whatever I can to keep everybody happy. |
|
|
| fdegrove |
Hi,
| quote: | | I'm not really concerned with blown amps (as they are easy to fix). |
From a business POV that is akin to commercial suicide if you don't mind me saying so... :whazzat:
Let's see if I can come up with a consultant at a reasonable fee for you...:D
Cheers,;) |
|
|
| Christer |
| quote: | Originally posted by fdegrove
Hi,
From a business POV that is akin to commercial suicide if you don't mind me saying so... :whazzat:
Let's see if I can come up with a consultant at a reasonable fee for you...:D
Cheers,;) |
I think we are all speaking for deaf ears. Maybe it's better to
just let him find out the hard way. If he's lucky he get no
problems. If not, he might get himself into deep trouble.
Too bad. He seems to be a nice bloke. |
|
|
| fdegrove |
Hi,
Christer,
Since everyone's going OT these days I'll have a little fun too:
Lesson in Dutch/Flemish #1:
| quote: | | I think we are all speaking for deaf ears. |
Tot dovemans oren spreken.
Lesson in Dutch/Flemish #2:
| quote: | | Maybe it's better to just let him find out the hard way. If he's lucky he get no problems. If not, he might get himself into deep trouble. |
Philosophical conclusion of the above:
Wat helpen kaars en bril als de uil niet zienen wil.
(Old flemish popular expression).
If you manage to guess it right you'll notice the etymological closeness of the European languages...Have fun.
Cheers,;)
P.S. Only trying to help, Peter...by no means patronizing you...Although I'd dare...;) |
|
|
| netgeek |
| quote: | Originally posted by Christer
Edit: I also believe these chips are mainly intended to be
built into integrated amp-speaker systems where the designer
has full control over the load condition. They probably never
were intended for general amplifier usage with arbitrary
unknown loads. |
Perhaps Christer has hit upon the "real problem". National wants to provide guidance for potential designers so they know what to expect - and to help them take care of the unexpected. I'll bet that many of the parts from the "recommended" (i.e. application note) schematic can be deleted in *many* applications - but we're talking about a "general purpose" and "unknown environment" in the case of DIY gainclones aren't we? I'd get alot of satisfaction as a DIY-guy if I had a PCB which included layout options for ALL of the app-note components and I only used 1/2 of them (with the rest unpopulated) - but ONLY if it was because I *understood* the circuit, had *tested* the parameters, and finally determined that MY application could do without some of the components - OR - that I'd done enough work (and the math, and the testing) to figure out how to optimise the values for MY application.
Given that the majority of hobbyists either lack the skills and/or the test equipment to make such determinations - Fred is ABSOLUTELY RIGHT in calling for the minimum of safeguards required to ensure that MOST applications will work reasonably (and safely). Why are we still bashing this subject? It's the responsibilty of those who know to look out for the interests of those that don't know - and that's exactly what Fred is doing. He should be thanked - and so should Peter - and so should PA - and so should anybody else who has gone to the trouble of thinking about, building something and trying to share it with the community.
I hope all of this discussion has served to enlighten the non-technical DIYers rather than frighten them off...... :) :)
All the best,
Bill |
|
|
| Peter Daniel |
| quote: | Originally posted by Christer
I think we are all speaking for deaf ears. Maybe it's better to
just let him find out the hard way. If he's lucky he get no
problems. If not, he might get himself into deep trouble.
Too bad. He seems to be a nice bloke. |
Maybe I'm lucky, or maybe I just know what I'm talking about. Why are you trying to do my work for me? It's not only because you just like to howl with the wolves, is it?
And I thought you were nice bloke too;) |
|
|
| jam |
| You a digging yourself a bigger and bigger hole....................most of the things that you have posted have errors or are ideal (not real world cases). You might want to stop and listen for a while. |
|
|
| Peter Daniel |
| quote: | Originally posted by jam
You a digging yourself a bigger and bigger hole....................most of the things that you have posted have errors or are ideal (not real world cases). You might want to stop and listen for a while. |
I know at least one other person who's digging much bigger hole for himself. Too bad he decided not to speak;) |
|
|
| SY |
| quote: | Philosophical conclusion of the above:
Wat helpen kaars en bril als de uil niet zienen wil.
|
Oehoe zegt de uil. |
|
|
| jh6you |
PD, you started this "Looking for successful Zobel cases." Did you start this thread not to discuss technical question, but just to appraise your commercial product? Very disappointing... If your product is ok with no RC compensation (Zobel network) even in case of potential oscillation, just shut up and keep up your self-confidence.
JP, are you sure that you are a Moderator?
JH |
|
|
| fdegrove |
Hi,
Kwak zegt de eend, kwaak zegt de kikker (Kermit), miaauw zegt de kat....Still teaching young Jimmy languages, he?
Good on you...:cool:
| quote: | | JP, are you sure that you are a Moderator? |
Why, does he remind you of another moderator too?:D
Oh, wait....Was he wearing his police cap at that time....checking.............
Cheers,;) |
|
|
| SY |
Since I have no idea of how to pronounce the Flemish words, I don't think you'd recognize what he's saying. He IS picking up quite a bit of Japanese from Iron Chef.
If amps blowing up were anything important in the High End, Jadis would have been out of business decades ago. |
|
|
| fdegrove |
Hi,
| quote: | | If amps blowing up were anything important in the High End, Jadis would have been out of business decades ago. |
That'll teach 'em....
They forgot to put those Zobel networks in there, didn't they? [joke] (Jocko Homo style, IP, TM, Copyright and Patent applied for, soon to be marketed at your local forum.)
At least they put in their own "poppers" (read fuses) in the cathode circuit now to protect the output tubes.
Mind you, the Higher End the amplifier is supposed to be the more keen on blowing up it seems to be too....
Something must have been lost in the translation from the RDH...
Not that speaker designers are entirely blameless either...
Progress by ommission...kinda sums it up nicely...:cannotbe:
Cheers,;) |
|
|
| Pedja |
The LM3875 and the GC story are not at their beginnings neither the knowledge about them starts these days. Some posts here are driven by the oversight of the fact of the existence of huge experience of pretty well behavior of LM3875 in almost any situation it has been put in (a few years and hundreds, maybe thousands of reports).
Some posts here are caused barely by the need for giving “advices”. In that regard, their authors should accept warning about lack of understanding of the game very seriously.
It is safer to recommend building of minimal GC than to recommend building of any class A amplifier or tube (high voltage) amplifier.
At the other hand, my impression is that moamps deserves the prize, but though I consider him quite competent in the audio electronics, a jury should check his supply capacitors layout (JLTi thread). If it is at all the same amp, of course. The jury yet should keep in mind that the Zobel has that good feature to act as a low pass filter. Many “successfully solved problems by Zobel” are bare consequences of this fact.
Another few issues…
Regarding the capacitive load, the Zobel network is only a compromise solution between the stability and low output impedance. If one is concerned about such load, then damped inductor, or low value series resistor are more adequate solutions. Damped inductor + RC Zobel, assuming nicely designed amp, will probably ensure unconditional stability.
Bear in mind that some circuits (amps output stages) as such ask for the circuit usually recognized as Zobel. In fact these circuits need low passing (HF shunting) of the currents running through them.
Closing thoughts:
I would personally rather put some safety belt on the amp, but if Peter has decided to give 10 years warranty for the amp without any load protection, this fact should be respected.
Pedja
PS: It is funny to see again and again statements like “I believe LM3875 is probably intended for…”, so I guess this will be present hereafter. I just do not know who can warn all their genius authors. |
|
|
| moamps |
| quote: | Originally posted by Pedja
...... but though I consider him quite competent in the audio electronics, a jury should check his supply capacitors layout (JLTi thread). If it is at all the same amp, of course. |
Hi,
It is not same amp. Anyway, what is to you problem with my amp?| quote: |
The jury yet should keep in mind that the Zobel has that good feature to act as a low pass filter. |
Yes. It's something normal for RC network.
Please calculate F(-3dB).
Please guys, (like Fred says) read some (simple) books.
Leach, Slone, Self .....;)
Regards |
|
|
| Christer |
| quote: | Originally posted by Peter Daniel
Maybe I'm lucky, or maybe I just know what I'm talking about. Why are you trying to do my work for me? It's not only because you just like to howl with the wolves, is it?
And I thought you were nice bloke too;) |
Well, I think we are nice blokes, or we wouldn't persist and
wouldn't even have cared in the first place. It is naturally
still your choice, but the question is, do you really understand
the choice you are making?
Frank,
While it is often possible to understand quite a bit of flemish
if one knows german, english and swedish, those examples
you posted beats me. :) |
|
|
| peranders |
| quote: | Originally posted by Peter Daniel
Since it seems like Zobel issue becomes more and more controversial, I'm looking for cases when implementing a zobel network in a chip amp helped to stop oscillations.
Please post here a link to any posts (on this forum, or the other, Chip Amps forum) that reported fixing the oscillation problem when Zobel was added.
I'm offering Brian's board to anybody that submits such link. |
Peter, I saw a really good explanation of this, with formulas and all... but where?
Still, I think the recommended values are a good choice. |
|
|
| UrSv |
| quote: | Originally posted by fdegrove
-- snip --
Kwak zegt de eend, kwaak zegt de kikker (Kermit), miaauw zegt de kat....Still teaching young Jimmy languages, he?
-- snip -- |
| quote: | Originally posted by Christer
Frank,
While it is often possible to understand quite a bit of flemish
if one knows german, english and swedish, those examples
you posted beats me.
[/B] |
Quack says the duck, ribbit says the frog, meow says the cat...
The only input I have, sorry. |
|
|
| peranders |
| quote: | Originally posted by Fred Dieckmann
I built one pair point to point with copper clad unetched PCB board segments for low inductance ground. I have never laid out a PCB for one or sold one. |
| quote: | Originally posted by Fred Dieckmann
Let see if I understand your question......... after over a decade and a half as a telecom design engineer and designed audio for 25 years you want to know if I have I have ever laid out a PCB?
After my head spun completely 360 degrees around on my shoulder like that poor little girl in The Exorcist, I thought myself "You melon head .....are you out of your mind PA?" Then it dawned on me that had to be humor! No other explanation would be possible.
I have lost count of the number of boards I have laid out with both CAD designers for and by myself for audio products. |
Sorry, Fred. I missed the word "one". I gather that you never have made any pcb for anyone but yourself? Is that correct? If that is the case, why don't you put something together? |
|
|
| Christer |
| quote: | Originally posted by UrSv
Quack says the duck, ribbit says the frog, meow says the cat...
The only input I have, sorry. |
Yes, I understood that one. It was the first two I meant, but
maybe that was in one of the other Zobel threads. One
almost feels like having split personalities when the same topic
is discussed in two or three parallel threads. :) |
|
|
| Christer |
| quote: | Originally posted by peranders
Sorry, Fred. I missed the word "one". I gather that you never have made any pcb for anyone but yourself? Is that correct? If that is the case, why don't you put something together? |
My interpretation is rather that he has spent 25 years laying
out commercial PCBs for the telecom company, which I presume
he doesn't own since they fired him.
(No, I don't mean he only worked with PCB layout) |
|
|
| jam |
Peranders,
You tend to make pointless attacks attacks on someones ability when you are backed into a corner. There are several members here that have for more experience and ability than you imagine...........stop acting like a child.
Regards,
Jam |
|
|
| runebivrin |
Not that P-A needs my defending him, but I don't think he's attacking anyone. He's genuinely misunderstanding Fred.
P-A: I think Fred's saying he's never made a GainClone PCB, but certainly has a great deal of experience with PCB-s otherwise - both commercial and personal.
Rune |
|
|
| jean-paul |
| quote: | Originally posted by jh6you
JP, are you sure that you are a Moderator?
JH |
Why are you asking ? |
|
|
| jh6you |
Your post #112 makes me put my question.
What is the real meaning hiding behind the "maybe"? |
|
|
| carlosfm |
| quote: | Originally posted by Peter Daniel
To make it clear for you, I'm looking for actual cases where Zobel was find useful, not theoretical. I don't know why this bothers you. |
Late arrival, but still in time?;)
Peter, the only way I could put back my Kimber 4TC speaker cable back in my system was to put two paralleled 0.47 ohm 3watt carbon resistors in series with the output on my GC power amp.
As you know, the Kimbers are mid/high capacitance cables.
The difference in sound was from almost no treble to a fantastically open and detailed treble.
Also, without the resistors it seamed that my amp was not man enough for my speakers, bass was affected too.
To get to this final result I tested almost all kinds of combinations possible with Zobels, with no satisfying results.
You can imagine this was intensive work, and always listening to the results.
After this, I always use these resistors on the output of my GCs.
They play much better with whatever cable and on whatever system you connect it.
Unfortunately, I can't prove this by words, so you'll have to trust what I say.:D
Or catch the first plane to Lisbon and I'll show you.:cool:
BTW, my final solution, after all, is to put a resistor (2.35 ohms seams spot on) on the chip's output, as you always do with op-amps.:eek: :D |
|
|
| Peter Daniel |
| quote: | Originally posted by carlosfm
To get to this final result I tested almost all kinds of combinations possible with Zobels, with no satisfying results.
|
So you mean you tried Zobel as well, but series resistors at the output worked better, right? |
|
|
| carlosfm |
Yes, that's right.
Much better.
Try it.
Try two 0.47 ohm resistors in parallel.
No white coffins, please.
Carbon or film.
If you have high capacitance cables to try, better still.
Kimber 8TC is even more capacitive than the 4TC.
:angel: |
|
|
| Peter Daniel |
I tried Mills 0.22, 5W resistors at the output, but I didn't notice any improvement, except for additional sonic signature of the resistors. My cables are easy load though.
Although your entry is not exactly what I was asking for, it's better than nothing and actually proves my point (that Zobel is not a universal fix). So if you don't have any of those (Brian's) boards, send me your address and I'll send you one;) |
|
|
| Pedja |
| quote: | Originally posted by moamps
It is not same amp. Anyway, what is to you problem with my amp? |
To me it is not any problem. I do not know if your amp has problems with a bit strange cabling around though. Positive supply cap oriented thus to enable at least 5-6cm longer traces… Then bypass caps mounted to the strange positions, are you bypassing the inputs? You have to bypass the chip inputs, not the PCB inputs.
http://www.diyaudio.com/forums/show...5696#post255696
| quote: | Yes. It's something normal for RC network.
Please calculate F(-3dB).
Please guys, (like Fred says) read some (simple) books.
Leach, Slone, Self .....;) |
What is the point is that the improvement is not necessary a consequence of the protected stability under the load, just a simple fact of low passed signal.
I am not sure what and why am I advised to calculate and read???
Like Fred…?
Pedja |
|
|
| Jax |
Um, the Zobel is as far as I know not intended to compensate an amp for capacitive loads. (?)
:2c: |
|
|
| Christer |
| quote: | Originally posted by Peter Daniel
Although your entry is not exactly what I was asking for, it's better than nothing and actually proves my point (that Zobel is not a universal fix). So if you don't have any of those (Brian's) boards, send me your address and I'll send you one;) |
Is that your point? Then why didn't you say so in the first place.
I think everybody would have agreed with you then. I thought
your argument was that the Zobel filter is never a (required) fix.
Of course a Zobel filter cannot prevent all imaginable problems.
Carlos,
Did all the Zobels you tried cause more oscillation than the
series resistor solution, or are you talking about sonic
preferences? |
|
|
| platenspeler |
| quote: | Originally posted by Peter Daniel
I tried Mills 0.22, 5W resistors at the output, but I didn't notice any improvement, except for additional sonic signature of the resistors. My cables are easy load though.
Although your entry is not exactly what I was asking for, it's better than nothing and actually proves my point (that Zobel is not a universal fix). So if you don't have any of those (Brian's) boards, send me your address and I'll send you one;) |
In post 59 I gave my experience with Zobel, and to my surprise it actually worked to get my amp stable.
I guess my entry is exactly not what you were asking for
:rolleyes: as it does not prove your point.
Too bad, apparently those entries do not get the boards :xeye: :bawling:
Maarten:D |
|
|
| jam |
Maarten,
They should................;). Don't you think so Peter. |
|
|
| jean-paul |
| quote: | Originally posted by jh6you
Your post #112 makes me put my question.
What is the real meaning hiding behind the "maybe"? |
The "maybe "was an answer to a post of Till. Since 2 threads were merged my post which was an answer to the post in the other thread got accidentally moved to this one. It was no answer to the previous post. I deleted it for clarity now. |
|
|
| moamps |
| quote: | Originally posted by Pedja
.....problems with a bit strange cabling around though. Positive supply cap oriented thus to enable at least 5-6cm longer traces…
|
Hi,
My whole PCB is 11cm long; where did you find those 5-6cm?
Look to PCB design below.(Elektor design with Zobels)| quote: | | Then bypass caps mounted to the strange positions, are you bypassing the inputs? You have to bypass the chip inputs, not the PCB inputs.] |
My intention is caps bypassing.
Also, this isn't HF design.
Also, this isn't design witch I want to sell to anybody.
Are you educated PCB designer?
Regards |
|
|
| moamps |
| quote: | Originally posted by Pedja
I am not sure what and why am I advised to calculate and read???
|
from book
M.Leach "Introduction to electroacoustics and audio amplifier design" |
|
|
| Peter Daniel |
| quote: | Originally posted by Christer
Is that your point? Then why didn't you say so in the first place.
I think everybody would have agreed with you then. I thought
your argument was that the Zobel filter is never a (required) fix.
|
It was my point all the time, as I stated it already in this thread and in the other threads. Zobel might be of help, but in case of LM3875 chip, when it oscillates, other "fixes" are much more effective. Those include better layout, better grounding and series, output resistor (sometimes with parallel inductance) at the output.
| quote: | Originally posted by platenspeler
In post 59 I gave my experience with Zobel, and to my surprise it actually worked to get my amp stable.
I guess my entry is exactly not what you were asking for
:rolleyes: as it does not prove your point.
Too bad, apparently those entries do not get the boards :xeye: :bawling:
Maarten:D |
I guess it wasn't what I was aking for. I was specifically asking for links to previous posts on this (or the other) forum showing that Zobel was helpful in reducing oscillations. Only one post in this thread qualifies for that so far, and it is a post #20. So don't get upset. I'm showing again my original request, so you can see for yourself I'm not making anything up:| quote: | Originally posted by Peter Daniel
Please post here a link to any posts (on this forum, or the other, Chip Amps forum) that reported fixing the oscillation problem when Zobel was added.
I'm offering Brian's board to anybody that submits such link. |
As to the Carlos prize, I said it's better than nothing, and since it again proves my point why not distinguish that attempt.
But send me your address and I'll forward a board for you as well.;) |
|
|
| RHosch |
| quote: | Originally posted by Peter Daniel
I guess it wasn't what I was aking for. I was specifically asking for links to previous posts on this (or the other) forum showing that Zobel was helpful in reducing oscillations. |
OK, so he didn't post a link to a previous post but only gave testimony to his experience of a Zobel being helpful in reducing oscillations.
Well, since his post is now a part of this forum history, I hereby post a link to it in full compliance with your original request. ;)
http://diyaudio.com/forums/showthre...8248#post398248 |
|
|
| Peter Daniel |
| quote: | Originally posted by RHosch
OK, so he didn't post a link to a previous post but only gave testimony to his experience of a Zobel being helpful in reducing oscillations.
Well, since his post is now a part of this forum history, I hereby post a link to it in full compliance with your original request. ;)
http://diyaudio.com/forums/showthre...8248#post398248 |
Clever attempt and technically it qualifies. Send me your address;)
For future reference, I will not accept any submissions with posts made after May 20, 2004;) |
|
|
| fdegrove |
Hi,
| quote: | | So you mean you tried Zobel as well, but series resistors at the output worked better, right? |
How does that qualify as:
| quote: | | OK, so he didn't post a link to a previous post but only gave testimony to his experience of a Zobel being helpful in reducing oscillations. |
Why do I feel as if the marketing department is being overly generous here?:smash:
Cheers,;) |
|
|
| Peter Daniel |
Frank,
RHosch did exactly what I asked for. That he linked a post that is made after I made my initial request is a different story, but since it was a clever attempt I decided to reward it the same;) |
|
|
| fdegrove |
Hi,
| quote: | | Well, since his post is now a part of this forum history, I hereby post a link to it in full compliance with your original request. |
When I than click the link RHosch posted, I stare at post #58 which happens to be Christers'.
Nevermind...Something's obviously escaping me here...:xeye:
Cheers,;) |
|
|
| Pedja |
| quote: | Originally posted by moamps
Hi,
My whole PCB is 11cm long; where did you find those 5-6cm?
Look to PCB design below.(Elektor design with Zobels)
My intention is caps bypassing.
Also, this isn't HF design.
Also, this isn't design witch I want to sell to anybody.
Are you educated PCB designer? |
Hello,
I did not say you want to sell that design. I wanted you to realize that you may have some problem with PS decoupling, yes for the high frequencies. Now I am disappointed if you in fact confirmed you do not give much attention to this issue. Usually it is wrong to think about amplifiers neglecting HF.
If you seriously do not see a way to reduce your PS local decoupling caps traces (seen by the chip) for at least 5-6cm in two steps, I just do not see much reason to talk with you about PCB design. None education could help. As about your PCB, this uneducated PCB designer is just interested to know do you have a driving license?
I do know Elektor’s (7-8/98) “100W single IC amplifier” article and yes, that board is obviously, considering PS caps, quite clumsy designed.
| quote: | Originally posted by moamps
from book
M.Leach "Introduction to electroacoustics and audio amplifier design" |
I can not understand your reason to post Leach in part where he talks some of the things I have already said in this thread. Might be you post just to have a post, but whatever it is, you just do not know what you posted. However, Leach’s recommendation for suited location for Zobel is quite funny. If you find a time, tell some here to mount Zobel on the output terminals because Leach recommends this… (Ok, I have some further thoughts on this but I’ll refrain from offering them as long as I see here play dead approach.)
Pedja |
|
|
| Pedja |
Sorry, I forgot…
It is completely disappointing if we just discovered that you designed your board inspired by that Elektor’s one.
See you. |
|
|
| carlosfm |
| quote: | Originally posted by Peter Daniel
I tried Mills 0.22, 5W resistors at the output, but I didn't notice any improvement, except for additional sonic signature of the resistors. My cables are easy load though.
|
Just to clarify, I have Epos 11 speakers and Kimber 4TC speaker cable.
Not an easy task for a GC, and my power amp has LM3886s.
I tried zobels, several values of resistor and cap across the output terminals and was not satisfied.
I had bought cheap Supra 2.5 cable to use with my GC, because it gave much better results than the better and much more expensive Kimber.
But I don't quit easily and I thought that if I could get my amp to work with the Kimber, I would gain in sound quality, I know this cable and I like it a lot.
Peter, I had tried previously 0.22 ohm resistors on the output and didn't notice any change, maby because they were bobinated resistors.:bawling:
The two paralleled 0.47 ohm 3W was a coincidence, because I couldn't find lower value carbon resistors.:D
I tried it and god:eek: shock, horror, the Krell came back:D , now my system was singing.
Peter, to test this you need (as I have) difficult speakers and difficult cables.
If you get your amp to work well under these circumstances, it will work almost everywhere.;)
| quote: | Originally posted by Peter Daniel
Although your entry is not exactly what I was asking for, it's better than nothing and actually proves my point (that Zobel is not a universal fix). So if you don't have any of those (Brian's) boards, send me your address and I'll send you one;) |
What you were asking for is much worst than what I gave you.;)
I don't have Brian's boards, I have my boards.;)
You can send me one (is it only one channel?:bawling: ), but of course I would be much happier with one of your boxes.:nod: :cheerful: :p :sly: :hohoho: :yikes:
I'll send you a mail with my address.:angel: |
|
|
| moamps |
| quote: | Originally posted by Pedja
...may have some problem with .... Now I am disappointed if you ..... do not give much attention.....it is wrong to think about amplifiers neglecting HF......If you seriously do not see ......... I just do not see much reason to talk with you .....None education could help....... this uneducated PCB designer is just interested to know do you have a driving license.....I can not understand your reason ..........you post just to have a post, but whatever it is, you just do not know what you posted......It is completely disappointing ..........
Pedja [/B] |
We are not in a War, Pedja.
EOD
P.S. Yes, I like Elektor,...since '85. |
|
|
| RHosch |
| quote: | Originally posted by fdegrove
Why do I feel as if the marketing department is being overly generous here?:smash:
|
He was only looking for cases where it was helpful, which Christer found it to be. Neither said the most helpful. ;)
| quote: | Originally posted by Peter Daniel
Clever attempt and technically it qualifies. Send me your address;) |
That's ok... I was just being oddly humorous. Besides, I have no interest in gainclones anyway. ;) |
|
|
| Jamh |
A question:
Doesn't the speaker's crossover circuit act in a similar role as a Zobel? It presents at least a cap and a resistor to the GCs output. |
|
|
| carlosfm |
My Epos 11 speakers only have a cap and a resistor in series with the tweeter.
Nothing more.:eek:
No crossover in the woofer.
That's the Epos way.:cool:
High-End.;) |
|
|
| platenspeler |
| quote: | Originally posted by RHosch
OK, so he didn't post a link to a previous post but only gave testimony to his experience of a Zobel being helpful in reducing oscillations.
Well, since his post is now a part of this forum history, I hereby post a link to it in full compliance with your original request. ;)
http://diyaudio.com/forums/showthre...8248#post398248 |
That's what I thought when pointing to my own post in this thread :D :D
But I think it's generous of Peter to send me a board as well. It's going to be a long summer with all these new goodies.
tx,
Maarten |
|
|
| Jamh |
| quote: | Originally posted by carlosfm
My Epos 11 speakers only have a cap and a resistor in series with the tweeter.
Nothing more.:eek:
No crossover in the woofer.
That's the Epos way.:cool:
High-End.;) |
So, isn't that similar to a zobel right there? Doesn't the extended (including cables and speakers) output circuit of the GC sees these? |
|
|
| BlackDog |
| quote: | Originally posted by Jamh
So, isn't that similar to a zobel right there? Doesn't the extended (including cables and speakers) output circuit of the GC sees these? |
I'm no crossover expert, but I would say that:
1) the capacitor is there to block DC, and to pass only higher frequencies to the tweeter.
2) the resistor is there to attentuate the sign to the tweeter to match the output level with the other driver in the system, ie woofer...
I very much doubt that loudspeaker manufacturers would build Zobels into their products. They would probably expect the amp to behave itself with "their" load. In other words they won't add the Zobel "just in case" the owner has an unstable amp. I wouldn't expect them too either. Furthermore, the effect of any components at the speaker end of the cable is pretty indeterminate - there's too many possibilities !!!
SteveM |
|
|
| fdegrove |
Hi,
| quote: | | I'm no crossover expert, but I would say that: |
Correct on both counts.
| quote: | | I very much doubt that loudspeaker manufacturers would build Zobels into their products. |
I'm not sure how common it actually is but my speakers sure did come with Zobel networks installed right across the speaker terminals on both woofer and midrange sections.
The speakers being 3-way units, the tweeter a purely resistive magnetostat obviously not requiring correction.
In fact that's were I feel these impedance correction networks should belong in the first place.
If everyone would adhere to such a standard, amplifier design would become that little bit easier.
Note that this doesn't solve the speakercable interfacing problem in any way but it helps.
Cheers,;) |
|
|
| carlmart |
| quote: | Originally posted by BlackDog
I very much doubt that loudspeaker manufacturers would build Zobels into their products. |
Did you know there's a speaker cable around (MIT if I am not wrong) that has a built-in Zobel in one end?
Even worst: did you know a patent claimed by MIT on such application (filter on cable) was accepted in the USA?
Patent is probably over now, but what would have happened if someone would have released a similar cable? Someone probably did and nothing happened. In any case nobody seems to hear anything on MIT cables anymore.
Carlos |
|
|
| Christer |
| quote: | Originally posted by carlmart
Did you know there's a speaker cable around (MIT if I am not wrong) that has a built-in Zobel in one end?
Even worst: did you know a patent claimed by MIT on such application (filter on cable) was accepted in the USA?
Patent is probably over now, but what would have happened if someone would have released a similar cable? Someone probably did and nothing happened. In any case nobody seems to hear anything on MIT cables anymore.
Carlos |
Maybe, but just maybe, MIT patented it with no intention to
enforce the patent just to make sure the idea could be used
by anyone? I suppose that wasn't the case, but it is not
unthinkable, though. |
|
|
| carlosfm |
I remember some cable of a brand I don't remember (Oracle?) that had big boxes in the middle of the speaker cables.
Those deam big boxes could fit a big toroid inside:D , so what the hell did they put there?
Just a zobel?:eek:
God, people are paying huge amounts of money for fancy cables, some interconnects I've seen have nice wood clamps before the RCAs and inside there's just a ferrite.:clown: |
|
|
| Sheldon |
| quote: | Originally posted by Christer
Maybe, but just maybe, MIT patented it with no intention to
enforce the patent just to make sure the idea could be used
by anyone? I suppose that wasn't the case, but it is not
unthinkable, though. |
All that's required to prevent someone from patenting your invention is to publish the invention. You can take it a step further and file a disclosure with the patent office, but that's not a requirement.
In most of the world the patent application must be filed before public disclosure and the first to file has priority. In the U.S., first to invent has priority, but the filing must be made within a year of public disclosure (even if someone else disclosed the idea), and you have to be able to document when the invention took place.
Back to your regularly scheduled discussion.
Sheldon |
|
|
| Christer |
| quote: | Originally posted by Sheldon
All that's required to prevent someone from patenting your invention is to publish the invention. You can take it a step further and file a disclosure with the patent office, but that's not a requirement.
In most of the world the patent application must be filed before public disclosure and the first to file has priority. In the U.S., first to invent has priority, but the filing must be made within a year of public disclosure (even if someone else disclosed the idea), and you have to be able to document when the invention took place.
Back to your regularly scheduled discussion.
Sheldon |
Well, you are right, but then somebody might try to patent it
anyway and get away with it. It is my understanding that
there has been a number of such cases in the US, at least.
Of course, you can take it to court, but that costs time and
money, and there is said to be at least one case where the
court did not consider well-known academic textbooks a
sufficient publication!!!
A somewhat reversed case is that the state of Bavaria has
the copyright for Hitler's "Mein Kampf", with the explicit
intention to make sure that nobody can publish it. |
|
|
| Sheldon |
| quote: | Originally posted by Christer
there is said to be at least one case where the
court did not consider well-known academic textbooks a
sufficient publication!!!
|
I would think that would be a rare instance. I believe it can happen in the case of a foreign publication. But if one was worried about that happening, filing a disclosure would eliminate that possibility. A disclosure filing is inexpensive.
Another interesting item is that holding a patent does not give one the right to make and sell the invention (as in the case of an invention that is an improvement on someone elses patent). It only confers the right to prevent others from producing it.
Sheldon |
|
|
| Christer |
| quote: | Originally posted by Sheldon
I would think that would be a rare instance. I believe it can happen in the case of a foreign publication. But if one was worried about that happening, filing a disclosure would eliminate that possibility. A disclosure filing is inexpensive.
|
There is a possibility that the story isn't true, of course, but it
definitely was a US company patenting something published
in the US. It was a computer algorithm, though, so maybe
that confused both the patent office and the court. As I
remember the story, Lotus got a patent for the Bubblesort
algorithm. It is actually one of the worst sorting algorithms,
so I don't know why they would even bother to patent it,
but it is common in textbooks, so maybe they didn't know
any better one. Anyway, someone else took the case to
court saying that the algorithm was well known since long
ago and was published in Knuth's "The Art of Computer
Programming", which is a standard reference work. The
court did not think that made the algorithm sufficiently
well-known. In the end they agreed howver to restict the
patent to the use of the algorithm in business applications.
| quote: |
Another interesting item is that holding a patent does not give one the right to make and sell the invention (as in the case of an invention that is an improvement on someone elses patent). It only confers the right to prevent others from producing it.
|
What would prevent you from making and selling if you have
a patent? Or do you mean the case where some other patent
overlaps and blocks you from making practical use of it? |
|
|
| RHosch |
| quote: | Originally posted by Christer
What would prevent you from making and selling if you have
a patent? Or do you mean the case where some other patent
overlaps and blocks you from making practical use of it? |
I think that's what he had in mind. If you patent an extension or improvement on an existing patent, and manufacturing the item required use of the prior patent, then you might still be blocked from putting your patent into practical use. All you have done is prevent the prior patent holder from using your improvement. Companies that compete in the same technology field often do this as a way of "hoarding" incremental improvements in SOTA. In many cases, they may not be able to implement their patented technology because doing so would infringe on one held by their competitor, but likewise their competitor cannot implement it either. Kind of a "if I can't do it, neither can you" game. |
|
|
| Peter Daniel |
| quote: | Originally posted by Pedja
__________________
"By pressing down a special key it plays a little melody!" (The Operator, 2 decades ago)
|
Hi Pedja,
I'm not sure if you know, but The Operator has a new album out, on Sony Music label;) |
|
|
| Pedja |
Honestly, I did not know. But checked for it now, sounds like immediate classic. Thanks for the info.
EDIT: I’ve found someone using words “nostalgic futurism” for this. I like it. |
|
|
| Sheldon |
| quote: | Originally posted by Christer
There is a possibility that the story isn't true, of course, but it
definitely was a US company patenting something published
in the US. It was a computer algorithm, though, so maybe
that confused both the patent office and the court. As I
remember the story, Lotus got a patent for the Bubblesort
algorithm. It is actually one of the worst sorting algorithms,
so I don't know why they would even bother to patent it,
but it is common in textbooks, so maybe they didn't know
any better one. Anyway, someone else took the case to
court saying that the algorithm was well known since long
ago and was published in Knuth's "The Art of Computer
Programming", which is a standard reference work. The
court did not think that made the algorithm sufficiently
well-known. In the end they agreed howver to restict the
patent to the use of the algorithm in business applications
What would prevent you from making and selling if you have
a patent? Or do you mean the case where some other patent
overlaps and blocks you from making practical use of it? |
It's hard to comment usefully without knowing the particulars of the case. However, you can get a patent on a new use of an old concept - depends on if it would be judged non-obvious to one skilled in the art. Keep in mind that the "art" here is defined by the field of the new application, not the field of the original idea. The vast majority of patents are just new combinations of previously known stuff. In fact, the best patents are the ones where after reading you slap your forehead, and exclaim: "of course, why didn't I think of that?" But if no one had proposed it before and it wasn't a trivial extension of well known stuff, then it can meet the test of non-obviousness, i.e., just because it's obvious when described doesn't make it obvious in the legal sense. The fact that it hadn't been done before in an area where it's clearly useful can be part of the agument for non-obviousness.
It helps to remember that, although laws use common words, those words when used in legal documents have meanings that are very specific and refined by court rulings. Legal use of a word may be somewhat different (usually more limited) from the meanings they have in ordinary usage.
Rhosch answered the second part.
Sheldon |
|
|
| Christer |
Sheldon,
Thanks for your answer. In this case I guess every programmer
would say it was obvious, but maybe that is why they are
programmers, not patent engineers? :) |
|
|
| Peter Daniel |
| quote: | Originally posted by Pedja
sounds like immediate classic |
I do like it. It has a lot of a character of their old music with a new twist. |
|
|
| carlosfm |
Hi Peter,
I've sent you a mail, have you received it?
BTW, did yoy try the resistors?:cool: |
|
|
| Peter Daniel |
| I got e-mail and will send the goods today. No chance to try resistors yet, but I tried Kimber 8TC on monoblocks and amp works fine. |
|
|
| carlosfm |
| quote: | Originally posted by Peter Daniel
I got e-mail and will send the goods today. No chance to try resistors yet, but I tried Kimber 8TC on monoblocks and amp works fine. |
Yes, my amp worked fine with Kimber 4TC and my brother's Rega Ela speakers.
I mean, test with Kimber AND difficult speakers and you'll see.
Epos?
ATC?:devilr:
Test with something your amp can't handle (I'm sure there's something:) ).
BTW, thanks for the "goods".:cool: |
|
|
|
