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 25th June 2002, 03:59 AM   #21
diyAudio Member
 
Circlotron's Avatar
 
Join Date: Jun 2002
Location: Melbourne, Australia
Default Hexfet thermal path

Copper is 2 inches by 2 inches, 3/16 inch thick.
Attached Images
File Type: jpg fet mounting.jpg (49.8 KB, 2070 views)
  Reply With Quote
Old 25th June 2002, 01:51 PM   #22
diyAudio Member
 
mrfeedback's Avatar
 
Join Date: Feb 2002
Location: Perth, Australia.
Default Pratt Truck

Hi Graham,
Nice looking truck !.
The copper sub plates are the only proper way to do it IMO.
This will also have good sonic benefit I expect.
If you consider that the audio waveform is assymetric, the power supply imbalance is probably a good thing if the speaker polarity is assigned appropriately.
Continue the good work !.

Regards, Eric.

BTW - Did you get my second transformer email ?.
  Reply With Quote
Old 28th June 2002, 08:14 AM   #23
diyAudio Member
 
Circlotron's Avatar
 
Join Date: Jun 2002
Location: Melbourne, Australia
Thinking of temperature compensated bias here, and particularly the limitations of the "amplified diode" setup that is commonly done by means of a pot across a small temp sensing bipolar. I did mention on another thread that I think that way is wrong because if you adjust the pot to get a different voltage then you also change the gain of the cct so the mV per deg C changes. The quiescent setting might be ok for the temp that you set it at, but as the temp goes up the amp will either "dig in" or "dry out". The exact voltage is important but so is the rate of change of voltage with temp. You need to be able to adjust both, not just the one.

The cct I am thinking about now has both these parameters separately adjustable. On the diagram they are referred to as "offset" and "slope". Here's how it works. The 4 temp sensing diodes have a certain voltage drop at 20 deg C, i.e. cold heatsink. Say 1.2 volts total for the 4 schottky's. Make the TL431 voltage reference equal to 1.2 volts. Actually I just realised it won't go below 2.5 volts so you will have to use a different type of reference. Now as the diodes heat up the LHS of the slope pot remains at 1.2 volts (as in 20 deg C diodes) but the RHS of the pot starts to decrease at 4 times 2.2mV per deg C. If we adjust the pot we can get a temp coefficient anywhere from zero to 8.8mV per deg C.

Turn on the amp. Straight away while it is still cold, with the slope pot fully ACW, dial up the offset pot to whatever it needs to set the cold bias current to the figure you desire. Done. Now watch and wait. As the amp warms the current begins to creep up. You have to set the slope pot for a suitable down-slope of bias vs temp. This is easy. When the temp and therefore the current has crept up a bit, dial the slope pot to bring the current back to where it should be. That's it! Give the amp a good rev up till it gets nice and warm, maybe even get it artificially warm, and then trim the slope pot a little to get the current just right. Now the current should stay fairly constant all the way from cold to hot.

GP.
Attached Images
File Type: gif bias.gif (9.3 KB, 1882 views)
  Reply With Quote
Old 28th June 2002, 09:04 AM   #24
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Default Circlotron amp using N-channel mosfets

Hi Circlotron,

Pretty neat idea! Having the benefit of hindsight, I was wondering:
If you just put the slope pot across the diode string, you get at the slope pot wiper a variable DC that is temp dependent and can be dialled to any slope from, what was it, 8.8mV/deg C downwards. Then you add the offset for the bias set point. So, it seems you don't need the regulator. What do you think?

Cheers, Jan Didden
  Reply With Quote
Old 28th June 2002, 12:26 PM   #25
diyAudio Member
 
Circlotron's Avatar
 
Join Date: Jun 2002
Location: Melbourne, Australia
Thumbs down Slope pot becomes pot of slop.

If you had the patience to set it correctly it would work ok. The trouble is you set the offset first, then when hot you set slope, but as you change the slope pot it would introduce a whole lot of offset as well. Think of it this way - at 20 deg C the slope pot should have no effect because the diode drop is the same as the reference, but the way you are suggesting even at 20 deg C you have 1.2 volts along the length of the slope pot. Of course if there is a simpler way of doing it I'd love to know!
  Reply With Quote
Old 29th June 2002, 10:31 AM   #26
djk is offline djk
diyAudio Member
 
djk's Avatar
 
Join Date: Feb 2001
Location: USA
Jan Didden, I remember your 200W cascode amp in AA a few years back. A brief description to jog peoples' memory. MJE200/210 65Mhz output transistors cascoded with 2N3773. If you want to build a no tears Circlotron with a single pair of 450W 37N50 use the following: A BurrBrown OPA2604 driving MJE200 cascoded with the 37N50. That is it. You're done. Patent #4,229,706 shows all the power supply and bias details. Patent #5,376,899 shows the feedback loop for the Bongiorno amp in figure 7 and how to wire it like the Pass amp in figure 9.

http://164.195.100.11/netacgi/nph-Pa...1&f=G&l=50&s1='4,229,706'.WKU.&OS=PN/4,229,706&RS=PN/4,229,706

http://164.195.100.11/netacgi/nph-Pa...1&f=G&l=50&s1='5,376,899'.WKU.&OS=PN/5,376,899&RS=PN/5,376,899
  Reply With Quote
Old 1st July 2002, 03:59 AM   #27
diyAudio Member
 
Circlotron's Avatar
 
Join Date: Jun 2002
Location: Melbourne, Australia
Question Cascode?

We must be thinking of 2 different things when you say "cascode". Could you give me a very rough schematic showing the relative connection of the MJE200 and '37N50. My idead of cascode is that they are both in series and the MJE200 would have to take the full current of the main FET, but the '200 is only a little device. Whats the story?

GP.
  Reply With Quote
Old 1st July 2002, 07:13 AM   #28
djk is offline djk
diyAudio Member
 
djk's Avatar
 
Join Date: Feb 2001
Location: USA
No, you're correct. The MJE200 is just a little transistor. Jan used five in parallel in his 200W design, and three 2N3773 for the cascode. The MJE200 is rated at 5A and 15W, the power is soaked up by the 150W cascode device.The Bongiorno Circlotron uses five TO3 in parallel. The point is that by using the Bongorino patent the bias problems are solved. The 37N50 makes a very good choice as the cascode because it needs no drivers and pre-drivers and the one device is rated to handle 450W. The OPA2604 can swing 80V P-P in this circuit with +/- 24V rails. The actual Bongiorno amp used an LF353 with +/- 20V and put out 67V P-P into 4R (about 130W). Should be able to get 200W/4R with the OPA2604. If higher powers are needed a discrete opamp may be built. 400W would be possible with one pair of 37N50 with everything else optimized. Any good BJT with high Beta could be used instead of the MJE200, a single 2SC3281 would be good for 100W/8R.
  Reply With Quote
Old 1st July 2002, 07:47 AM   #29
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Default Circlotron amp

djk,

I think Circlotron is right. My amp is not a circlotron. Have to look at those patents though. Must run now, boss just came in.

Cheers, Jan Didden
  Reply With Quote
Old 1st July 2002, 02:26 PM   #30
diyAudio Member
 
Circlotron's Avatar
 
Join Date: Jun 2002
Location: Melbourne, Australia
djk, I suppose then if you went to a cascode cct and really wanted to use hexfets in combination with bipolars then you could use any old low voltage hexfets with a suitable current rating. As far as the Bongiorno bias cct is concerned, when I was looking at it the other night I think it was a class A, which would indeed make it easy to control the bias, but I want to run class AB so above a moderate volume level the average voltage drop across the source resistors starts to rise and that would cause the bias to back right off. I feel to tired and lazy to get up and have another look at the patent at the moment, so would you do it for me please?

Actually, driving a bipolar above a low voltage low Rdson hexfet in cascode was done sometimes in early SMPS's because on switch-off the voltage was across C-B, not C-E, and transistors that have a Vces of 400v typically have a Vcbo of ~700v so that was good. What's more, with the emitter suddenly open circuited the entire switch current is diverted into the base and out the collector (electron current flow) until the C-B junction has all the charges swept out and it turns off. And it's turns off quick smart!

So I wonder then if a hexfet / bipolar *linear* cascode would persuade a relatively slow bipolar to turn off quicker by reducing the storage time? Any current continuing to be conducted because the transistor is a little slow turning off automatically becomes reverse base current and this should be good to help things move quicker. It would be interesting to see how much difference it makes to a slow old slug like a 2N3055.

GP.
  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
Only N-Channel MOSFETs (NMOS); better Audio from non complements by Audio Power? tiefbassuebertr Solid State 46 27th July 2012 10:09 PM
Replacement transistors/mosFETs for holton n-channel amp hulsebos Solid State 24 12th March 2010 04:56 AM
Push-Pull using only N-Channel MOSFETs Kashmire Solid State 32 6th August 2009 01:48 PM
Circlotron djk Pass Labs 13 20th August 2006 03:19 AM
P-channel mosfets for Aleph Peter Daniel Pass Labs 10 15th August 2002 11:12 PM


New To Site? Need Help?

All times are GMT. The time now is 05:12 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