evil inductances
Hi Bonsai,
The funny thing is that some people around here can only hear a coil when it's placed inside the amp, but not when placed outside, disguised as piece of wire, cross-over filter or voice coil.
But let's be serious. Not only things like topology, fT and compensation method determine the the stability of the OP stage, but also the leads or traces to the power transistors have a marked influence, which -I'm afraid- is overlooked by most people.
If the OPS makes havoc, it's usually around 10MHz. A piece of wire of say 5cm, has a self inductance of about 5nH and an impedance (at 10MHz) of 0.314 Ohm. Given the high currents, large enough to compromise the stability.
BTW, that's why the PCB of my next amp will have 4 layers, so I can make the traces to the OP devices short and wide. Equally important, decoupling of the power rails (with low ESR caps) is also easier and more effective when using 4 layers.
Cheers,
Edmond.
Bonsai said:
Hi Bonsai,
The funny thing is that some people around here can only hear a coil when it's placed inside the amp, but not when placed outside, disguised as piece of wire, cross-over filter or voice coil.
But let's be serious. Not only things like topology, fT and compensation method determine the the stability of the OP stage, but also the leads or traces to the power transistors have a marked influence, which -I'm afraid- is overlooked by most people.
If the OPS makes havoc, it's usually around 10MHz. A piece of wire of say 5cm, has a self inductance of about 5nH and an impedance (at 10MHz) of 0.314 Ohm. Given the high currents, large enough to compromise the stability.
BTW, that's why the PCB of my next amp will have 4 layers, so I can make the traces to the OP devices short and wide. Equally important, decoupling of the power rails (with low ESR caps) is also easier and more effective when using 4 layers.
Cheers,
Edmond.
john curl said:
Mr. Curl,
I cannot help but understand the feeling of frustration underlying your post.
But then please do understand that within a strongly technically oriented playground like this of electronics design, it is very hard for us to surrender reason and take as gospel what an otherwise undisputably proficient designer has to say.
Mix into this an allowance for human error on both sides, and add the fact we must necessarily figure in the influence of eventual vested interests behind - until proved nonexistent if it were possible - just for the sake of completness, and you will not be surprised to meet challenge.
Rodolfo
john curl said:
And in this regards, 'extra special' is in the eye of the beholder. This industry has ample proofs of this fact, otherwise one could not understand why is possible coexistence of disparate paradigms as class A, vacuum tubes, no feedback, high feedback etc. all of them backed by frequently fundamentalist supporters, and all for the same purpose of audio amplification.
Rodolfo
Re: evil inductances
Yes, I have seen the "10nH per cm" rule quoted here frequently, but this only applies to a 0.25mm wide track. Make the track 10mm wide and the inductance is a great deal lower.
BTW, for my 100A plus output power amp, I’m hard wiring the 20 pairs of output devices in lieu of a PCB, using aluminium bus bars for the power rails (25mm X 10mm), output and ground return (50mm X 5mm). 😀
Cheers,
Glen
Edmond Stuart said:
Yes, I have seen the "10nH per cm" rule quoted here frequently, but this only applies to a 0.25mm wide track. Make the track 10mm wide and the inductance is a great deal lower.
BTW, for my 100A plus output power amp, I’m hard wiring the 20 pairs of output devices in lieu of a PCB, using aluminium bus bars for the power rails (25mm X 10mm), output and ground return (50mm X 5mm). 😀
Cheers,
Glen
Attachments
nH/cm
Hi Glen,
That's precisely the reason why I will make the traces as wide as possible.
Thank for the formula anyhow.
Cheers,
Edmond.
Hi Glen,
That's precisely the reason why I will make the traces as wide as possible.
Thank for the formula anyhow.
Cheers,
Edmond.
Re: Re: Re: evil inductances
Complete? No. But I have a photo of my new workshop for which the remaining 9 meters by 5.5 meters of the 15 meter long floor was poured on Tuesday. And until I get my pedestal drill press and workbench dyna bolted to the floor, I won't be able to drill the many mounting holes in the bases of the two massive 1500mm X 220mm X 85mm heastsinks either.
All will be revealed on my website, eventually........
PMA said:
Complete? No. But I have a photo of my new workshop for which the remaining 9 meters by 5.5 meters of the 15 meter long floor was poured on Tuesday. And until I get my pedestal drill press and workbench dyna bolted to the floor, I won't be able to drill the many mounting holes in the bases of the two massive 1500mm X 220mm X 85mm heastsinks either.
All will be revealed on my website, eventually........
Attachments
Re: Re: Re: Re: evil inductances
Talk about doing things from scratch. You even build a new house to be able to build your amp. 🙂
G.Kleinschmidt said:
Talk about doing things from scratch. You even build a new house to be able to build your amp. 🙂
cooling
Hi Glen,
Don't forget to build a swimming pool as well. You will need the water for your next project: a liquid cooled ultra high power amp. 🙂
Cheers.
Hi Glen,
Don't forget to build a swimming pool as well. You will need the water for your next project: a liquid cooled ultra high power amp. 🙂
Cheers.
john curl said:
The problem here is that you want to have it both ways, John. You want to be able to jump in with a comment about coils that suggests that anyone who uses them is an inferior designer or does not have the secret sauce. Then when questioned or challenged, you can't or won't back your claims up. That is what is really frustrating. It is certainly not just less experienced designers who ask follow-up questions to your assertions that you would prefer not to answer.
Bob
Re: evil inductances
Hi Edmond,
These are good points, and it is important to recognize that an output capacitance not only can de-stabilize a global feedback loop, but it can also de-stabilize an emitter-follower output stage that has no feedback around it.
Cheers,
Bob
Christer said:
Edmond Stuart said:
Hi Edmond,
These are good points, and it is important to recognize that an output capacitance not only can de-stabilize a global feedback loop, but it can also de-stabilize an emitter-follower output stage that has no feedback around it.
Cheers,
Bob
Talking about the importance of short PCB tracks, there is AN-1135 from IR. This app-note is for classD, but I think the basics are also applicable to classAB amps. We should minimize signal current loops (that enters and leaves via gate to source, the importance of decoupling capacitors for this loop, not just for power supply decoupling), like fig.7, fig.8. Putting attention to this loop and minimizing physical loop area, sufficient track width results in better audible sonics.
Would anyone like to discuss something something instead of output coils? I don't use them anymore, and I don't have much to say, except you might look at the 20KHz damping factor at 20KHz of the JC-1 power amp. It is better than most, due to the fact that I don't use an output coil. Is it more prone to oscillation? Maybe, but not in my experience with 1,000's of units built over the last 15 years with a similar circuit topology.
Already discussed but about RF getting into preamp enclosures and Bob's practice of terminating input RCAs with 50ohm res and cap. If there is an impedance mis-match at RF would that not stop RF from going into the enclosure. Does it matter that it is reflected back and forth in the cable? Two less components.
I just now had a problem with a vintage Sony TA-N7 (remember those, output stage a cascode of fast bipolars with VFET on top?) that played a local radio station with an unterminated input. Sometimes even breaking through with terminated input.
This amp has dual inputs, one direct and one via a blocking cap. Just to try it, I connected 150pF from each input RCA to a common 50 ohms to ground, directly at the input panel. Gone was the local station. Thanks Bob!
Jan Didden
This amp has dual inputs, one direct and one via a blocking cap. Just to try it, I connected 150pF from each input RCA to a common 50 ohms to ground, directly at the input panel. Gone was the local station. Thanks Bob!
Jan Didden
john curl said:
What exactly did you want to discuss ?? Since you are so tight lipped about your own secret designs what is the point of your contributions to this thread ?? For all of the misleading information that you have offered on this thread you may as well talk about the price of fish on Mars !!
You remind me of someone who likes to keep the top shelf whiskey to himself whilst handing out the cheap casket wine to his guests 🙁
Yes, Snoopy, I don't PUBLISH schematics. That is the only thing that I keep to myself. I design audio schematics for a living, yet you would insist that I let that go also. I am impressed.
To be honest with everyone, it is a shame that so many people are closed minded about even successful audio design concepts. I attribute this to the 'sophomore' effect, where a little knowledge in a field, brings out a lot of opinion, without any real understanding.
I am only here to help, but to little avail, it would seem.
To be honest with everyone, it is a shame that so many people are closed minded about even successful audio design concepts. I attribute this to the 'sophomore' effect, where a little knowledge in a field, brings out a lot of opinion, without any real understanding.
I am only here to help, but to little avail, it would seem.
- Status
- Not open for further replies.