Mosfet Amp [Simple

[Steben]

Since it is a class A follower, you "only" need to beef up the input voltage. Class A sucks power all the time, so there is no need to worry that much about even "hotter" disasters. heatsinking...sinking there's that submarine again! 

[petemoore]

  Yupp, I got it running loud and cool.
  Doesn't get warm.
  Resistor does 'just...
  And Cool sounding amp too.
  I tightened the bass up with some speaker enclosures, paper mache', I started with a plastic bowl, large enough, should have oiled it, cut the 1/4''+ thick mache' to get it off the bowl, this sealed around the speaker easily [about an hours messing, 8 hours drying, mostly dried so I didn't wet the speaker with it] just glued right in, I softened the edges...
  The other side had room so I cut a styrofoam bucket's top 3'' off [blade mounted to block, flat surface kept the blade cutting nicely around the sides, this one weather stripped and pressed in from the bottom of the bucket!
  Big difference in the foreward projection and bass response.
  I looked up all over for a transistor signal _/+ swing splitter, didn't really see anything much, but did find tube configurations which could be easily subbed with Jfets or other I believe, just put equal value R's on the S/D as in the phase inverter circuit of a triode...same schematic but small signal device and power supply voltage...I must have gone around the pages that show this or there is some other issue with using a chip to get _/+ signals.
  Seems simple enough and I might just try it...take a signal from the D to one amp, the S to the other...then the one amp would be invertedly adding gain to the signal swings to the other, speaker between them...I'll start with a nold speaker..lol.
  I'm trying to think about what I need to think about grounding the two amps and the inverter...not much I guess, just ground 'em, maybe star ground the signal/power.
  Upgrading the PS current is the first business of upgrade I would need to attend to.

[mac]

Found this, not  a class a but maybe usefull

http://members.tripod.com/~gabevee/sstubamp.html

mac

[petemoore]

  The input side..looks like equal value S/D resistors would divide the signal into a version and inverted version of itself, the 1 resistor under the 2 resistors under the input Q sources...I guess to accomodate source bias.
  Input resistor on the gate of Q1.
  That's about it for DC in the first '1/3' of the circuit except..and I don't know how..
  Actually I don't understand how that could work.
  The second transistor input' would have to be...
  Some kind of feedback amp, at least the first two sections before the top/bottom output transistors.
  It looks like Signal in to Q1, out Q1's drain to the top biplar's input, the amplified version of that collector output to Q2 mosfet Gate input...providing some type of source feedback...is the only output connection I can see Q2 as having to the signal path.
  The output transistors appear to have their drains connected, and their inputs connected/inputs connected to both bipolar collectors AC wise.
  I'm trying to sort out how it could work and would have to trust it does and build a model to figure it out better, with no values stated tho..
  It looks like an interesting enough topology, I just can't follow exactly what the path is, I can't quite tell what's going on there, what Q2 would be up to other than feedback, and how it's push pull when it looks like what's introduced to the last 2/3rds [bipolars and fet's] is one version of the whole ac signal. It looks like it could be split by the first fet, but the source signal side would have to go through 2 source resistors into an fet source and out the gate, then into the combined collectors ! that can't be it.
  Somehow the collectors are intended to drive the second fet's gate, it's source somehow having influence on the input fet's source...go figure !
  Barring the first 2/3rds, the output section {I'm not really that advanced to match an input 'pre' to it..} looks like it might be workable.
  Well...it'd need to have the drains 'opposite'...using inverted inputs to top and bottom gates...that'd mean a 'fancy' pre-amp...probably what he's got right there, - the values.

[alextheian-alex]

Just thought of something... don't know why it slipped my mind before, but since the voltages and currents that you are running are so low, you can just rig up an LM317 as a current source, and probably boost your efficiency a bit by using that instead of a resistor... just 3 parts, 317, current set resistor and a capacitor to bypass the regulator to prevent oscillation.  Should be a simple mod to try out, and it could probably be used just as well in a "ZEN" style amp running at slightly lower voltages and currents.

[petemoore]

you can just rig up an LM317 as a current source,
  I have some LM317 Rigs, I used them to regulate on Muamp starving and other VR needs I had.
  When you say LM317 as current source...
  as in splice the regulator between the PS WW and the circuit, a little voltage drop'd be fine, I should try regulatin' anyway.
  317 data, then I'll measure actual voltage again on the noisy one.
  ................
  I mightta overcurrented it? dag if I didn't figure out at one point running a 'plenty 'o current'...what was it [I measured so many V's yesterday] I think it was 32Vdc, with the gate pulled all the way toward V+...I think it'd be easier for me to replace that Mosfet and see than it would for me to figure out if it Should take that...I reason that 'giving way' is something that'd be heard from a transistor..IOW...something 'gave way'...as in the 'bigger' noise didn't actually begin until after the circuit had run like that for about 4-5 seconds, plus that power supply transformer / Br. Rect. /filter with AC120v input was letting AC in the DC, it ran perfectly quietly when I stepped up voltage from 5.6VAC to whatever it was...~18vdc.
  So...I know it's in the PS or the circuit, probably the active if it's in the circuit.
  I had it running very quiet, and it's just reglar noise level, usable but unacceptable since I know it's getting lowered.

[Steben]

Found this, not  a class a but maybe usefull

http://members.tripod.com/~gabevee/sstubamp.html

mac

Ah, alwys the same. "I use MOSfets because of their "even harmonic content". and "I'm going to put in some AC feedback to reduce distortion"...  yeah, 

[alextheian-alex]

you can just rig up an LM317 as a current source,
  I have some LM317 Rigs, I used them to regulate on Muamp starving and other VR needs I had.
  When you say LM317 as current source...
  as in splice the regulator between the PS WW and the circuit, a little voltage drop'd be fine, I should try regulatin' anyway.
  317 data, then I'll measure actual voltage again on the noisy one.
  ................
  I mightta overcurrented it? dag if I didn't figure out at one point running a 'plenty 'o current'...what was it [I measured so many V's yesterday] I think it was 32Vdc, with the gate pulled all the way toward V+...I think it'd be easier for me to replace that Mosfet and see than it would for me to figure out if it Should take that...I reason that 'giving way' is something that'd be heard from a transistor..IOW...something 'gave way'...as in the 'bigger' noise didn't actually begin until after the circuit had run like that for about 4-5 seconds, plus that power supply transformer / Br. Rect. /filter with AC120v input was letting AC in the DC, it ran perfectly quietly when I stepped up voltage from 5.6VAC to whatever it was...~18vdc.
  So...I know it's in the PS or the circuit, probably the active if it's in the circuit.
  I had it running very quiet, and it's just reglar noise level, usable but unacceptable since I know it's getting lowered.

Just a quick and dirty way of replacing the fancy CCS in the Zen, or the simple resistor in the amp posted above.  Since 317 are available ar Rat Shack, I just buy them by the handfull, and whevener i need to regulate voltage or current, I'm just like ... 'well, throw a 317 on it and see what happens'.  If it works out, then I'll move on to something fancier.

[evertonberger]

Hi pete you have an idea off how much watts in your project??
Schemtics?? some?
Thanks
Nice work

[petemoore]

  <2watts
  Enough to power an 8'' alnico 'nicely'...with the preamp of course, the 'amp' is a source follower, which will drive the speaker/output to what the voltage swings of the input are...very small if just guitar, much bigger with a gain stage @9v, @18v on the 'preamp' [transistor gain stage] starts getting enough clean drive to begin getting loud enough to distort the amp.
  For me that is two power mosfets on the R side of stereo, a power mosfet with a 2n7000 mosfet 'pre' on the L side.
  MAC
    http://members.tripod.com/~gabevee/sstubamp.html
  The topology of this amp, even without the resistor values...it just looks marvelous...I like the idea of a channel 'above and below' the output [schematic left side/power output], the symmetry of the diagram is alluring, no parts values is...disconcerting though.
  All that said and done...I'm ordering some TDA/Lm3886 chips mahn, looks like one of those takes alot of the gruntwork and cajoling out of the output amp 'design-build-test' process.

[mac]

pete,
maybe out of topic but RG pointed to this preamp. if you're looking for one:
http://members.tripod.com/~gabevee/sstubepre.html
http://members.tripod.com/~gabevee/
the author claims even order harmonics and bla bla bla... similar to the blackfire.

[petemoore]

The amplifier as a unit can be used as a stage in a phono preamp.
  Probably works good for bumping signal level up...I think I'll buffer between the drive and output stages on one side first.

[petemoore]

  ANyone?
  Only a few places have them AFAICT.
  NAtl Semi wants 920 for shipping, but that wont work, the 'Buy' button don't work either.
  Futurlec has got badrap on order handlings.
  BD online I can't find online.
  I don't read french.
  NTE doesn't mess with them.
  Natl. Semi wanted the 920 but by the time I got the Visa out, I can't find the dern pay for postage page...they don't want me ordering anymore free samples after the ones they're not sending me, and the 'buy' / 'buy online' buttons don't work, the computer just makes the 'bunk' sound.
  3 hours later I still don't have an order placed...
  any suggestions..these chips are being hard for me to order.

[alextheian-alex]

Do you mean the 3886?  They are all over the place.  partsexpress.com has them for around $7, Digikey.com for around $5-6, and you can find them all over ebay.

[petemoore]

  2 minutes at Parts Exp. did the trick. Order confirmation already !
Monday estimate...thanks Alex !

[alextheian-alex]

  2 minutes at Parts Exp. did the trick. Order confirmation already !
Monday estimate...thanks Alex !

No problem.  If you don'y have it already, here is the super-duper helpful design spreadsheet: http://www.national.com/appinfo/audio/files/Overture_Design_Guide13.xls  it does all the dirty design details for you, and even lets you tweak the freq response, and tells you how much heat you'll be kicking up so that you can pick a heatsink accordingly.

[petemoore]

  Again agin, thanks, just what I just sat down to go hunt up !!
  Word...lug-row...whaddoyacallit...a row of lugs on an insulated board?
  since the construction is pretty simple I'm thinking of doing so layout drafting today, and Sunday maybe get most of the circuit built up, heatsink, box 'n stuff, there's alot can be done and take some time to do it before..hopefully monday the chips'll be in.

[evertonberger]

Hi pete... very cool and we are waiting for news now... and a schematic soon!!! hehe
I will waiting your project be done to begin to make my own... for now I want to build one of this thing for a long time now...
And sorry my bad english a live in brazil and dont have much time to speaking English with others persons!!

[evertonberger]

Uohwww I forgot One link:
http://audioclone.free.fr/Ampli-LM3886.html

Prety cool design 2 LM3886 in bridge...

[brett]

Nice work Pete
this is a classic.  Not many people manage to start a project simple and keep it simple. 
After a couple of projects similar to this, but using a constant current source (CCS) instead of that 10W resistor, I concluded that the CCS was overkill and may have impaired the sound quality. 

When using the CCS I tried a trick to prevent nasty clipping: a pre-amp that clipped (diodes to ground) just *before* the MOSFET clipped.  It worked quite well.  I was using simple, two JFET pre-amps.  The first JFET boosted by about 15x, then diodes clipped the signal (to 0.7V), then a volume control, then the second JFET boosted again by about 15x.

Pete, you might want consider how much current is needed from your power supplies.  At 24V, my amps were soaking about 0.7 A.  That 30V 0.5A supply would be mighty stretched, even at 24V.  The amp won't quite swing rail to rail, but 24 V peak-to peak sine wave = 8.5 V = 1.1 A into 8 ohms, so you can see that these are power-hungry little buggers.

thanks for cool work and updates