Thanks for your kind words about the book and your understanding that the class D section is just a starting tutorial at best. I have learned a lot more about class D in the three years that have passed and continue to seek more input and suggestions on that section. I do wish someone had written a good class D book before I wrote the section in my book (at least one that I was aware of), because I had to do quite a lot of research and learning to write the amount that is there. Any references on class D are very, very welcome.
I'll definitely add some material on the current feedback topologies and try to list some good references that people can go to for further reading.
Your idea about including an option to purchase a pcb of an example amplifier as it evolves is a good one.
Cheers,
Bob
Not this side of $ 500, no.
PDF editing is very hard and the software to do it will cost you.
Adobe CS versions do it.
You'll find many offerings on the 'net but they are very limited; I've tried a few but was quite disappointed.
jan
That's a shame. I would have thought that wholesale replacing one font for another should be straightforward. Does the PDF format not essentially say "draw these characters using font "x""? If instead of full-blown editing I had some way of changing "x" from, say, Times New Roman to Palatino, that would be a good start.
Well you can if you open the PDF in a text editor, but that's soooo nerdy
jan
Ghostscript lets you get at fonts somewhat. A simple global substitution should be doable. There are lots of copyright pitfalls here, there are only so many license free fonts.
Fonts and font facilities supplied with Ghostscript
Hello,
Try converting the entire file to word format.
Next make your changes.
Then export back to PDF.
Conversion programs are available for free.
myhrrhleine, That is what I was thinking. If you already have a copy of Microsoft Office or even Open Office, though I think that name has been changed to Libra Office, you should be able to change to any one of the fonts that you already paid a license fee for from Microsoft or the free versions in Open Office.
But what if your PDF is a graphic with some text in it? Does that process really work? For example, can you change the font used in the attached PDF, or change the dash characters (that MATLAB erroneously uses for negative numbers) to the correct minus-sign character?
Myhrrhleine, I have a hard time envisioning someone in the healing arts picking up a soldering iron and going at a bunch of op-amps, though stanger things have happened. You only list cooking and sewing on your web site as hobbies, is audio a secret interest?
Hello,
Silly me. I thought audio as an interest was obvious on this forum.
I got interested in audio long before healing.
A MOSFET Power Amplifier with Error Correction* Paper
Bob,
I wonder if would you be kind enough to clarify a little section within your paper. This is the end of section 2.2 on page 12. It is where you discuss the use of shielded wire with a 10R resistor at the driver end and a ferrite bead at the MOSFET gate.
My questions are as follows:
1) Is this method ok to use with Lateral MOSFETs (I suspect it would be ok)
2) What do you mean by local bypass ground? Is this the ground used for the decoupling caps? Is so, does this not cause a ground loop?
3) What is the spec of the shielded cable?
4) What is the spec of the ferrite bead?
It would be very much appreciated if you could find the time to assist with these questions or point me in the direction of other references.
Many thanks
Paul
Bob,
I wonder if would you be kind enough to clarify a little section within your paper. This is the end of section 2.2 on page 12. It is where you discuss the use of shielded wire with a 10R resistor at the driver end and a ferrite bead at the MOSFET gate.
My questions are as follows:
1) Is this method ok to use with Lateral MOSFETs (I suspect it would be ok)
2) What do you mean by local bypass ground? Is this the ground used for the decoupling caps? Is so, does this not cause a ground loop?
3) What is the spec of the shielded cable?
4) What is the spec of the ferrite bead?
It would be very much appreciated if you could find the time to assist with these questions or point me in the direction of other references.
Many thanks
Paul
Hi Paul,
Wow, you read the paper pretty closely! Long time ago that was written. I had to go back and review it to answer your question. I had long forgotten my mention of that gate drive approach.
I implemented the technique during the prototyping of the amplifier, but eventually went with the combination of a gate stopper and a gate Zobel network. I seem to remember that in the prototype there were 3 or 4 inches of signal path between the driver and the output MOSFETs, hence the concern about the inductance.
The technique should be definitely compatible with lateral MOSFETs, but I strongly doubt it is needed for them, since laterals have significant internal gate resistance and are not nearly as fast as vertical MOSFETs.
In that amplifier I defined only one ground, so yes, the interconnections of the two coaxial cables to the gates, each connected to the same ground at each end, created a ground loop. I can only say the I do not recall the ground loop creating a problem.
I think the coax was 50-ohm RG-174. Nothing special, but nice and small and flexible. I have never really liked bulky RG-58 and don't even use it for BNC cables in my lab.
I don't recall the spec of the ferrite beads - probably whatever I had at hand at the time
.Cheers,
Bob
Bob,
Thank you for the very quick response. I have read that paper many times
The good news is I have by chance some miniature 50 ohm coax and some ferrite beads of unknown spec. So will be experimenting.
The problem I have is trying to parallel two pairs of Exicon dual die (16A) Lateral MOSFETs. Want to be running gate resistors of 100R and 150R (or lower). With one pair I have no problems. With two pairs I get a high (ish) frequency oscillation of around 10Mhz with 4V peak to peak amplitude. Increasing gate resistances to 220R and 330R the oscillation stops. But this slows the MOSFETs down too much and Phase margin of the amp suffers according to simulation.
Have followed the data sheet / application notes regarding PCB layout with local decoupling for each FET. Also have gate zobels currently 330+100pF and 220+100pF depending on whether N channel or P channel. The output stage is on a separate board to the drivers.
I need MOSFET speed and not sure how achieve it. I know that the gate resistors need to be small. The idea of 10R gate resistors appeal. have read that laterals have higher gate resistance so should the resistor at the driver match this? Have a suspicion parasitics are to blame for the oscillations just a little stuck on how to approach this problem.
Paul
Thank you for the very quick response. I have read that paper many times
The good news is I have by chance some miniature 50 ohm coax and some ferrite beads of unknown spec. So will be experimenting.
The problem I have is trying to parallel two pairs of Exicon dual die (16A) Lateral MOSFETs. Want to be running gate resistors of 100R and 150R (or lower). With one pair I have no problems. With two pairs I get a high (ish) frequency oscillation of around 10Mhz with 4V peak to peak amplitude. Increasing gate resistances to 220R and 330R the oscillation stops. But this slows the MOSFETs down too much and Phase margin of the amp suffers according to simulation.
Have followed the data sheet / application notes regarding PCB layout with local decoupling for each FET. Also have gate zobels currently 330+100pF and 220+100pF depending on whether N channel or P channel. The output stage is on a separate board to the drivers.
I need MOSFET speed and not sure how achieve it. I know that the gate resistors need to be small. The idea of 10R gate resistors appeal. have read that laterals have higher gate resistance so should the resistor at the driver match this? Have a suspicion parasitics are to blame for the oscillations just a little stuck on how to approach this problem.
Paul