diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   How to make a low distortion lateral MOSFET amplifier? (http://www.diyaudio.com/forums/solid-state/243907-how-make-low-distortion-lateral-mosfet-amplifier.html)

danny92 10th October 2013 05:19 AM

How to make a low distortion lateral MOSFET amplifier?
 
2 Attachment(s)
Hi everyone,

I'm projecting a lateral MOSFET amplifier using BUZ900/905 pair, but my intention is to make an amplifier that can deliver 200 Wrms to a 4 ohm load and 100 Wrms to 8 ohm, with 30 dB of gain, and with THD20 < 0.005% and THD1 < 0.0005% (if possible in the ppm region 0.00000x). I'm trying to optimize this circuit for almost one year [a circuit similar to the Bob Cordell's designs and more recently with some improvements of the Honey Badger by Peter V (ostripper)], the problem is that I can get THD20 characteristics lower than 0.02% and THD1 values lower than 0.0000x%, even using advanced compensation methods like TMC. I'm wondering if I should give up from this project, and start a new one from scratch, but with a condition, the output pair will have the BUZ900/905 pair of MOSFETs. I'm thinking maybe in a CFA design, maybe something similar to the circuit design by dadod, but I don't know nothing about CFA :(
There are any documentation about out to develop and test a CFA audio amplifier?
Does anyone with more MOSFET experience give me an idea to what I should do?
I don't know if I should change this design to meet this requirements (if possible), or if I should redraw the circuit from scratch, maybe using another topology.

Best regards,
Daniel

astx 10th October 2013 07:56 AM

Compare your input/vas stage with

http://www.diyaudio.com/forums/solid...ml#post3577892

Or try my asc files and replace the bias stage and bjt ops with your ops. During design phase remove input caps and NFB capacitors.

BR, Toni

mcd99uk 10th October 2013 08:08 AM

If you want low THD you need high open loop gain and use some sort of two pole compensation to get good THD20. The high open loop gain will give large amounts of feedback correction and hence low THD. Amount of feedback correction available = open loop gain - closed loop gain.

danny92 10th October 2013 01:58 PM

Hi everyone, Thank you for helping me,

The solution that you've presented seems very interesting mcd99uk, increasing open loop gain to have more bandwidth, and using TPC to increase the slew rate at higher frequencies. For increasing the open loop gain I have to lower the degeneration resistors at the IPS and VAS, what do you think that can be a good degeneration factor? As about the compensation method I understand better TMC, but maybe TPC has better distortion characteristics, the main reason that leads me to choose TMC is the absence of square wave overshoot.
PS: Increasing open loop gain leads to higher cmiller values, right?

Best regards,
Daniel

moschfet 10th October 2013 03:19 PM

2 Attachment(s)
An other very low distorsion lateral-mosfet-amp.

Edmond Stuart 10th October 2013 07:30 PM

Hi Daniel,

For an ultra low distortion amp (with laterals) see:
Home
PGP Amplifier
PGP (mirror plus more info)

Cheers, E.

dadod 10th October 2013 07:49 PM

Quote:

Originally Posted by Edmond Stuart (Post 3662314)
Hi Daniel,

For an ultra low distortion amp (with laterals) see:
Home
PGP Amplifier
PGP (mirror plus more info)

Cheers, E.

Hi Edmond,
How good measured distortion corresponded to the simulated one, as there are no good lateral mosfet models?
BR Damir

Edmond Stuart 10th October 2013 08:48 PM

Good question Damir.
I just estimated the distortion by filtering out the artifacts (i.e. higher harmonics) caused by the imperfections of the lateral models.
edit: The filtered THD sims were reasonable in line with the real THD figugers, though a little optimistic.

Cheers, E.

dadod 10th October 2013 09:10 PM

Quote:

Originally Posted by Edmond Stuart (Post 3662422)
Good question Damir.
I just estimated the distortion by filtering out the artifacts (i.e. higher harmonics) caused by the imperfections of the lateral models.
edit: The filtered THD sims were reasonable in line with the real THD figugers, though a little optimistic.

Cheers, E.

What models did you use?
BR Damir

bear 10th October 2013 09:27 PM

Daniel, didn't I just see this same front end in another thread?
Pretty sure I did??

_-_-bear


All times are GMT. The time now is 09:34 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2