FIRE
What I worry about now is that even if I put the Zobel in the amp how can I be sure when the cap is aged or not ? and sounds like not if but when it will burn my house down
Chris
What I worry about now is that even if I put the Zobel in the amp how can I be sure when the cap is aged or not ? and sounds like not if but when it will burn my house down
Chris
chris ma said:
Well, obviously this isn't what I'm saying directly, but allow me to clarify for those who might take my off-hand remark a bit too literally 😉
To generalise hugely, and ignoring all other variables like standard of construction, etc, think about the heat that is constantly being created by a class A amplifier. As you no-doubt know, heat is the enemy of reliable operation, and that's why we all worry so much about heatsinking. Roughly speaking, reducing the operating temperature of a semiconductor junction by 10 degrees C will double the lifespan of the component.
(Yes, I know the above is an oft-quoted cliché and doesn't withstand rigorous analysis, but for a first approximation it's not a bad thing to bear in mind)
So, a class B amplifier ought to be more reliable on these grounds, assuming the basic design is sound and the protection works well enough. Bear in mind that most people don't play music at high levels continuously, although for PA use this point is mute. But then we're discussing class A amps, so I guess that means we're not talking about PA applications… 😉
Note also that lowering operating temperature improves the chances for all the other components in the amp - particularly electrolytic capacitors. You've only got to consult the manufacturer to see how temperature affects the lifespan of these, and you'll find no shortage of horror stories from people who've witnessed these things "letting go"!
Without doubt, the biggest challenge in building a class A amp is dealing with all that heat (stating the obvious 😉), but also important is anticipating the eventual failure modes and building in protection circuits to stop the thing catching fire or destroying your speakers. This latter point applies equally to any amplifier, and can be extended to include anything mains powered you build ever 😉
Cheers,
Mark 😉
On the other hand, I'm not saying that you shouldn't implement Zobel in your circuit, if you feel that it's required.Peter Daniel said:
For that reason I find posts like the one form head_spaz very entertaining😉
Re: FIRE
Hi Chris,
I'm not sure if you're joking or not 😉
But if you're really worrying about the Zobel cap, rest assured because it won't be an electrolytic capacitor. And non-electrolytic capacitors tend to be very, very reliable 😉
Cheers,
Mark
chris ma said:What I worry about now is that even if I put the Zobel in the amp how can I be sure when the cap is aged or not ? and sounds like not if but when it will burn my house down
Chris
Hi Chris,
I'm not sure if you're joking or not 😉
But if you're really worrying about the Zobel cap, rest assured because it won't be an electrolytic capacitor. And non-electrolytic capacitors tend to be very, very reliable 😉
Cheers,
Mark
Re: Re: This thread is an embarrassment.
If you days back is ages ago to you, we can assume that you may well exaggerate about other issues.
If you look at the current board, you'll notice that some of your trivial suggestions were implemented ( I actually jumped on implementing them, right after you posted). Some other less trivial or ones that dodn't make sense, were not implemented.
Would you agree with that?
As to the treatment on the forum, I've been done much worse, yet I don't complain about it.
mhennessy said:
If you days back is ages ago to you, we can assume that you may well exaggerate about other issues.
If you look at the current board, you'll notice that some of your trivial suggestions were implemented ( I actually jumped on implementing them, right after you posted). Some other less trivial or ones that dodn't make sense, were not implemented.
Would you agree with that?
As to the treatment on the forum, I've been done much worse, yet I don't complain about it.
Zobel required for Chip Amp
Peter:
If you are concered about the need for a "Zobel" network on the output of the amp try the following-
Connect the amp with very short wires to a high quality decade cap or a series of caps starting at .01 uF and increase until you get to about 5 uF. Drive the amp to a 1V P_P square wave and watch the output with a scope. Look for ringing and oscillation. If you see none then its not necessary. (Most amps have enough bandwith to start oscillation around .1 uF to 1 uF.)
The "Zobel" network (really an RC damper in this case) combined with the series inductor/resistor in the output keeps the feedback loop from exceeding 180 degree phase shift and inducing oscillation. Any audio amp will have an output impedance that rises with frequency (like an inductor). Combining this with a capacitive load and enough feedback you make an oscillator. (This is simplified but I hope clear.) The two networks serve to isolate any capacitive load and keep the phase shift at the output of the amp less than 180 degrees.
However the series RL will effect the output impedance and may effect the frequency response at the speaker which may be audible.
Integrating these components on the chip is both very difficult and costly AND they need to be optimized for each gain configuration. If you follow National's configuration exactly (and don't test it for power bandwidth) you will probably never encounter oscillation. If your design needs a different gain, or other changes then all of the stability issues should be looked at. If your load never becomes dominantly capacitive (Low C cables, dynamic speakers, well designed crossover) you may not need the networks.
I have built and shipped wideband feedback amps without but its a calculated risk.
Again I suggest testing.
-Demian
Peter:
If you are concered about the need for a "Zobel" network on the output of the amp try the following-
Connect the amp with very short wires to a high quality decade cap or a series of caps starting at .01 uF and increase until you get to about 5 uF. Drive the amp to a 1V P_P square wave and watch the output with a scope. Look for ringing and oscillation. If you see none then its not necessary. (Most amps have enough bandwith to start oscillation around .1 uF to 1 uF.)
The "Zobel" network (really an RC damper in this case) combined with the series inductor/resistor in the output keeps the feedback loop from exceeding 180 degree phase shift and inducing oscillation. Any audio amp will have an output impedance that rises with frequency (like an inductor). Combining this with a capacitive load and enough feedback you make an oscillator. (This is simplified but I hope clear.) The two networks serve to isolate any capacitive load and keep the phase shift at the output of the amp less than 180 degrees.
However the series RL will effect the output impedance and may effect the frequency response at the speaker which may be audible.
Integrating these components on the chip is both very difficult and costly AND they need to be optimized for each gain configuration. If you follow National's configuration exactly (and don't test it for power bandwidth) you will probably never encounter oscillation. If your design needs a different gain, or other changes then all of the stability issues should be looked at. If your load never becomes dominantly capacitive (Low C cables, dynamic speakers, well designed crossover) you may not need the networks.
I have built and shipped wideband feedback amps without but its a calculated risk.
Again I suggest testing.
-Demian
I agree it is a calculated risk, but again, i go with it.
I also don't try to impose any views or practises, I merely report my approach. Anybody's milage may differ😉
What I don't understand are those knuckleheads, who jump on that, like I'm committing a crime😉
I also don't try to impose any views or practises, I merely report my approach. Anybody's milage may differ😉
What I don't understand are those knuckleheads, who jump on that, like I'm committing a crime😉
Calculated Risk
Just tell your customers not to connect high Q caps or cables (Goertz) directly to the outputs. What we did (and Spectral is still doing) is only warrant the product used with the recommended cable, MIT. We knew the cable had enough isolation that the amps would not oscillate with them in the chain.
-Demian
Just tell your customers not to connect high Q caps or cables (Goertz) directly to the outputs. What we did (and Spectral is still doing) is only warrant the product used with the recommended cable, MIT. We knew the cable had enough isolation that the amps would not oscillate with them in the chain.
-Demian
Peter, you became a criminal as soon as you became a vendor, you didn't see that attitude shifted from your buddies? give me a break😀
Mark, thanks for the reassurance regarding non-electrolytic capacitors. BTW, I was not joking when house fire involved.
Regards,
Chris
Mark, thanks for the reassurance regarding non-electrolytic capacitors. BTW, I was not joking when house fire involved.
Regards,
Chris
Re: Re: Re: This thread is an embarrassment.
I'd like to comment, but I don't understand the first half of this sentence.
I've been looking out for a recent version of the board, but unless I've missed it, I haven't seen anything since last week.
Umm, no. Right after posting, you jumped on me and dismissed every suggestion
http://www.diyaudio.com/forums/showthread.php?postid=391595#post391595
You only started to consider implementing anything after I expressed my indignation. I wonder if you would do that now?
To you, maybe....
I'd like to see a more recent board. I'd also like to (sincerely) thank you (again) for taking me more seriously later in the thread - just in case you think there are still hard feelings - which isn't the case on my part 😉
Yes, but by a Moderator? As discussed in the other thread, your status meant something to ordinary members, whether you believed that or not. I would have worried much less about that sort of treatment from a peer, but I believe I was right to stand up to you then.
However, things have changed. We can argue on a more equal footing now. Although I don't particularly want to argue with anyone
Best regards,
Mark 😉
Peter Daniel said:
I'd like to comment, but I don't understand the first half of this sentence.
I've been looking out for a recent version of the board, but unless I've missed it, I haven't seen anything since last week.
Umm, no. Right after posting, you jumped on me and dismissed every suggestion
http://www.diyaudio.com/forums/showthread.php?postid=391595#post391595
You only started to consider implementing anything after I expressed my indignation. I wonder if you would do that now?
To you, maybe....
I'd like to see a more recent board. I'd also like to (sincerely) thank you (again) for taking me more seriously later in the thread - just in case you think there are still hard feelings - which isn't the case on my part 😉
Yes, but by a Moderator? As discussed in the other thread, your status meant something to ordinary members, whether you believed that or not. I would have worried much less about that sort of treatment from a peer, but I believe I was right to stand up to you then.
However, things have changed. We can argue on a more equal footing now. Although I don't particularly want to argue with anyone
Best regards,
Mark 😉
chris ma said:
No problem. And yes, you're right to ask - especially when safety is concerned 😉
Cheers,
Mark 😉
Re: Re: Re: Re: This thread is an embarrassment.
That attitude explains quite a lot. I will not comment on that, however.
PS: It should read: "few days back"
mhennessy said:
That attitude explains quite a lot. I will not comment on that, however.
PS: It should read: "few days back"
I was wondering about that too Peter. Why didn't anyone bring
the Zobel subject up when you first posted This is not another Gainclone? (Circa December 2002!)
In fact, i never saw it mentioned in any of the other forums or
websites of all the other people building GC's before you ever
posted your schematic.
It's like something gets popular (Like Grey's Aleph-X)... and then it
must be torn down.
I hope this doesn't put a damper on the next guy who wants to
post an idea or project that just might get popular........
the Zobel subject up when you first posted This is not another Gainclone? (Circa December 2002!)
In fact, i never saw it mentioned in any of the other forums or
websites of all the other people building GC's before you ever
posted your schematic.
It's like something gets popular (Like Grey's Aleph-X)... and then it
must be torn down.
I hope this doesn't put a damper on the next guy who wants to
post an idea or project that just might get popular........
- Status
- Not open for further replies.