Amplifier with jfets and laterals / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Amplifier with jfets and laterals

previous 1 2 3 next

Move Thread

LAN_403

Mattski

Thu Oct 27 2011, 10:57PM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

Inducktion wrote ...

Experimentonomen wrote ...

You gotta be kidding me! Its one of the simplest things you can do in diy electronics next to a lamp and a battery.

Just to prove how simple it is, heres a schematic:

That doesn't seem simple.

There exist simpler amplifiers than this one. This one here, if I'm reading it correctly, is a differential high-gain op-amp type amplifier with negative feedback for high linearity, and a high current output stage. You can also put together a simple amplifier with nothing more than a single transistor, a few resistors, and some capacitors, though it won't have the bells and whistles that this one does.

Experimentonomen

Fri Oct 28 2011, 08:40AM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

I do NOT like chip based power amplifiers, even though they offers nice low THD, they have too much internal compensation stuff that makes makes the sound tireing to listen to.

As i said, an amplifier cannot be made any simpler than my schematics and still offer good performance.

To Dr.Kilovolt, these mosfets are the lateral type which compensates themselves without the need for a Vbe multiplier, thus only a simple pot or even a fixed resistor is needed to set the bias to around 50-100mA.

Experimentonomen

Sat Oct 29 2011, 06:50PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

Some pics:

Power stage:

Frontend and drive board:

Experimentonomen

Sun Oct 30 2011, 02:49PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

It's DONE!! :D

A video:

Steve Conner

Sun Oct 30 2011, 04:24PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Nice job, and it's cool that you managed to get hold of some original 2SJ50/2SK135 lateral MOSFETs.

Just one quibble, from your schematic it looks like you have the output stage wired as a CFP. So the idle current is determined by the Vbe of the driver transistors, not the threshold voltages of the FETs. And so strictly speaking you do need the thermal compensation.

Nelson Pass's Zen amps are simpler, but not everyone would consider their performance "decent". Needs moar power!

Experimentonomen

Sun Oct 30 2011, 08:25PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

The fets were given to me guy a guy that scrapped a large 800W amp that had been retired after 30 some years in a theatre.

This CFP thing is a bit weird, some say the Vbe limiter(thermal compensation) should be placed on the driver transistors and some say it should just be placed on the board like any transistor(which has the same effect as my current pot) as placing it on the main heatsink would overcompensate a CFP design and sent it into class b with following crossover distortion as a result.

So far i havent had any thermal runaway, each channel can idle around 100mA per output transistor = 200mA for the whole channel and still only get luke warm. When cold during tests it could start out at say 75-90mA for all output transistors together and a hour later around 150mA total. I guess the self compensation of the laterals still prevent thermal runaway as long as the idle current doesent exceed what the heatsink can handle.

We'll see, if it turns out to be a problem, i could always add the Vbe limiters, or even just a transistor wired as a diode in series with the pot.

Steve Conner

Sun Oct 30 2011, 09:09PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes, the Vbe limiter is supposed to be placed on the driver transistors. The lateral MOSFETs lose their temperature compensating ability in the CFP. (by the same argument, the CFP output stage needs emitter ballast resistors even though the lateral MOSFETs normally don't)

If the circuit is designed so that the drivers don't dissipate much heat, then just putting the Vbe limiter on the board will work the same.

I've got Bob Cordell's new book, and he loves to give designs made with dual JFETs that you can't get any more.

Experimentonomen

Sun Oct 30 2011, 09:26PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

There should be suitable substitutes.

Another fun thing with this amp is that i can run it at almost no bias current and it still sounds good, i cant hear any ounce of difference between 15mA and 75mA.

I've always been told that mosfet amps want 100+ mA before the crossover distortion even begin to disappear.

I wish i had one of those THD meters like the one you have to see if this thing is anywhere near what the simulation claims and see if thetre is a diff in THD with CFP and EF, sim claims CFP gives like 5x lower thd than a EF, 0.0005% for CFP and 0.001% EF.

Steve Conner

Sun Oct 30 2011, 10:23PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

It makes sense that the CFP gives lower distortion. When the FETs are used as emitter followers, the distortion comes from them having a gain less than 1. It takes something like 8V to make one of those lateral FETs conduct 6A.

But in a CFP the MOSFET is inside a feedback loop that hides its behaviour and the whole circuit block behaves like a super BJT emitter follower. It has gain closer to 1, so there's less distortion. Douglas Self wrote about this, and he also showed that it needs less idle current than an EF output stage. So I guess by putting feedback round the MOSFET you also hide its need for a high idle current.

The THD results from Spice are always a bit optimistic. One reason is that it doesn't include the circuit noise, but real THD meters do. OTOH I think Cordell showed that it overestimates the THD for MOSFET amps because the stock FET model doesn't model subthreshold effects. Real FET output stages have a smoother crossover region than the model.

I'm restoring an old Electrocompaniet just now, it uses Matti Otala's original circuit.

Experimentonomen

Sun Oct 30 2011, 11:02PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

I think i need to buy more 0.1ohm 5w resistors, running low. Or maybe 0.22 or 0.47 ohms is better.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob