MF500 | My 500W/4Ω Hi-Fi All-FET Amp Design - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 20th July 2012, 08:22 PM   #11
diyAudio Member
 
Nico Ras's Avatar
 
Join Date: Nov 2005
Location: East Coast of South Africa
MagicBox, why not save the schematic as a PDF, then it can be viewed any size the reader wants also save it for later reference.

I take it that your aim was to lower the total noise in the bunch of parallel sources by root 8 due to the summing of non-coherent noises else you would have used single devices, correct?

Interesting design my friend, a little complex but clever.
__________________
Kindest regards
Nico

Last edited by Nico Ras; 20th July 2012 at 08:35 PM.
  Reply With Quote
Old 20th July 2012, 08:53 PM   #12
diyAudio Member
 
Join Date: Sep 2011
Hi Nico

Did you mean the sources of the output transistors or the ones in the current mirrors? I assume that you mean the transistors in the currentmirror, though. I'd say no, I didn't choose multiple transistors to reduce noise, although if that is a side results, that's just another free benefit

I did it because I primarily wanted to mirror the full swing vas rail so I would have a second rail whose voltages I could divide freely, thus being able to position the gain/bias element anywhere between supply rails. With a small scale amp, a 1:1 mirror would probably be sufficient.

This amp however sports 200V across its voltage output rail and to maintain a high bandwidth the quiesent current through this VAS rail is relatively high, chosen at 45mA. If it were 50mA, the rail as a whole would dissipate 10 watts. That current mirrorred 1:1 would still be 45mA. A quite high current for the gain devices and another 10 watts of dissipation doing nothing but accepting controlling currents within a free-to-pick voltage range. So, that's why I added the devices, to explicitly make use of the current multiplication property that the Wilson Mirror provides and 'move' current from the mirror input to the output. With a control input of 6.43mA quiesent current into both legs that amounts to 51.43 mA which is a little over 10 watts of dissipation, barely a bit more than the power I need to drive the output stage.

Because the mirror itself is not supposed to act as a current gain stage, the multiplication shouldn't be too high either. Just high enough to balance the input and output rail currents so that most of the dissipation goes into the VAS output rail. Then we have the practical number of transistors, thus multiplication factor and boardspace to consider, which together made me strike a compromize at a factor of 7 for a total of 8 transistors per mirror.

Oh as for turning the schematic into a PDF, I'll do that when I am going to draw the schematic into PCAD for PCB development and output to PDF from there, then the supplies and other sub schematics will end up in there too. For now you could save the schematic image at the bottom of the first page or rightclick it and "View link in another tab". That will show the image directly into your browser and you can then zoom in/out and scroll.

I hope it makes sense what I say.. if not just say so, I'm open to positive critics
__________________
Negative Feedback: The Need for Speed

Last edited by MagicBox; 20th July 2012 at 08:58 PM.
  Reply With Quote
Old 30th July 2012, 07:29 AM   #13
diyAudio Member
 
Join Date: Sep 2011
Hi,

I've now started to breadboard pieces of the concept and after initial troubles (the high bandwith current test source) and so far it works out well. Cascodes and a reasonable matching are required for the mirror elements, but they'll do the job perfectly. Once I've scaled it up to 7 mirror devices and completed the whole circuit I'll try to run an AC analysis on it.
__________________
Negative Feedback: The Need for Speed
  Reply With Quote
Old 30th July 2012, 08:10 AM   #14
diyAudio Member
 
Join Date: Sep 2011
PS using BS170 and BS250 as replacements for the ZVP/ZVNs. I'm almost inclined to go SMD and put the SMD mirror transistors very close together and thermally couple them with a very small heatsink, more to keep them thermally equal rather than a need for cooling.
__________________
Negative Feedback: The Need for Speed
  Reply With Quote
Old 30th July 2012, 02:07 PM   #15
diyAudio Member
 
Join Date: May 2001
Location: Bangalore, India
Matching SMDs is that much more difficult.

Do you think adding a small value resistor to the Gates of the current mirrors may help in avoiding oscillation?
__________________
Sam
  Reply With Quote
Old 30th July 2012, 02:23 PM   #16
diyAudio Member
 
Join Date: Sep 2011
I've cascoded the output with an IRF9510 picking a zener that causes the exact same voltage drop over the mirror output Vds as the input Vds (= Vgs average of whole mirror block). This one I 'silenced' with a local feedback cap of 22pF. All the mirror transistors have their gates directly connected, no resistor in between. This just degrades performance. A mirror is a pure slave topology with a high bandwidth, if you stabilize the input, the output will follow provided the gates are indeed as directly connected as possible to reduce lag/oscillation in the output response. I knew when things where oscillating on my test setup without even attaching a scope only 2 well placed 22pF ceramics were needed to get it all stable. I'm using 12Ohm degenerative resistors in the source leads btw with a 5mA current per mirror transistor.

Under the cascoded configuration with the equal Vds voltages it's easy to match them and once a good match is made, the mirror is surprisingly stable with temperature.
__________________
Negative Feedback: The Need for Speed

Last edited by MagicBox; 30th July 2012 at 02:26 PM.
  Reply With Quote
Old 30th July 2012, 08:25 PM   #17
diyAudio Member
 
Join Date: May 2001
Location: Bangalore, India
Thanks MB for the reply.
__________________
Sam
  Reply With Quote
Old 30th July 2012, 09:12 PM   #18
diyAudio Member
 
Join Date: Sep 2011
No prolblem.. I meanwhile extended the number of transistors to 4 and still have an exact gain of 4 and that worked too, being stable without any compensation other than the 22pf local feedback on the current source output transistor, a BC550C and one across the opamp output/negative input. There is a reasonable amount of current swing possible without the multiplier drifting much. The more curves are matched, the more this drifting is minimized. Within +/-10% of the qiescent current the multiplier barely changes. The accuracy of the multiplier matters only to the DC bias for the construction, and that's as accurate as your match on the intended current. S far it's all going well and I'm expecting the breadboard version of the input- and driver stage to work as intended, at least, as a functioning negative feedback amplifier
__________________
Negative Feedback: The Need for Speed

Last edited by MagicBox; 30th July 2012 at 09:22 PM.
  Reply With Quote
Old 2nd August 2012, 05:58 AM   #19
diyAudio Member
 
Join Date: Sep 2011
Things going well I've got about 100 each of BS170 and BS250. I've been able to match an 8-transistor P mirror with the BS250's under +-2mV GS difference at the intended operating current, temperature and tension, the result is sweet! I have to match some more and meanwhile wait for an order of a 610/9610 batch, 2SK170's and assorted JFETs for the current sources. But it should then be possible to build an operational front/mid end..
__________________
Negative Feedback: The Need for Speed

Last edited by MagicBox; 2nd August 2012 at 06:12 AM.
  Reply With Quote
Old 2nd August 2012, 06:00 PM   #20
diyAudio Member
 
Join Date: Sep 2011
Dear Mirror, dear Mirror, who's the most gorgeous man...

Oh wait.. I think I can say I finished the north side BS250 mirror 8 Transistors, 8 12 Ohm degen resistors cascoded and all with an input current of 6.43, which gives an approximate current of 45mA. The current stays reasonably stable with temperature. All gates are directly connected; the degens function as NFB too besides averaging out the remaining Vgs differences and lessening their impact some more. Bandwith exceeds 2MHz easily, the current source into which I injected a 2MHz test signal suffered severe rolloff, but modulating right into the mirror let it through full power. The mirror has at least 20mA lineair swing too, it's great. Together with accurate and secure assembly, this is going to be a great building block for this amp
__________________
Negative Feedback: The Need for Speed
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
200w/8Ω smps & class d module debugging authlxl Class D 24 8th June 2010 03:30 AM
fet amp design revamp unclejed613 Solid State 2 14th January 2007 08:26 AM
Fully complementary FET-in FET-out design Bensen Solid State 36 7th January 2006 09:21 AM
New 500W reference design from Zetex thecyberdude Class D 1 1st April 2005 10:14 PM
Selfoscillating HalfBridge 500W+500W by IR. Interested? deflorator Class D 0 10th October 2004 03:30 PM


New To Site? Need Help?

All times are GMT. The time now is 08:33 PM.


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