N-Channel Amp with single OP devices

[Samuel Jayaraj]

There have been many who have encountered problems with Anthony Holton's N-channel amp. I too had my share of problems but chose to use lower output rail voltages. Five of these amps are playing everyday in different installations without any problems.

My question is that, do you think I can get away with using IRFP260 or IRFP460 output devices, A SINGLE PAIR for one amplifier, lower rail voltage to +-45volts (100watt into 8ohms, according to Anthony), bridge two of these and squeeze out close to 400watts into 8ohms for PA use?

I'd like to use these for Mid-horns because the mid frequency sonic quality of the N-channel amplifier is magical.

[Kilentra]

I think it's safe to budget 100 watts output for each pair of Mosfets. So if you will be operating a bridged amp at 400 watts then two pairs per side would be safer. Otherwise there is nothing wrong with the idea.

[Samuel Jayaraj]

Parellel mosfet output devices is not the problem with budget, size etc., It is this whole issue of the Driver/Splitter Mosfet not sourcing enough current to drive their(OPs) combined capacitance and hence, asymmetrical turn-on/turn-off times being the real issue.

Any further thoughts?

[darkfenriz]

Hi Samuel Jayaraj
the idea is clever, well-known (cars etc.) and overall ok.
but real life is following laws of Murphy rather than laws of Kirhchoff...
risky...
400W with 4 transistors makes ... say 60 to 140 W of power of heating so up to 35W per transistor... eghm..scorching
by the way... where may I find the schematic of this amp??
Could you attach it or link the page ??
I would be grateful...

back to the topic : try using 2 pairs per amp- this shouldn't affect the sound and you won't smell the burning siliconium...

if you're affraid of the current maybe a simple emitter followers in front of the FETs' gates will do? (I didn't see the schematic but..)

cheers !!!!

[Kilentra]

I do not think it really matters how many paralleled devices are used as long as the previous stage is not driven to distortion. I don't know what the limit is for the N-channel design, but keep in mind that Holton added a source follower after the VAS on the AV800, which uses 7 pairs of output devices. The sym amp works fine with 4 pairs of IRFP240/9240 without any such buffer. Just from my intuitive guess, using one pair only is probably easiest because there's no problem with current sharing and the gate capacitance is lowest. The only disadvantage besides lower power, is higher output impedance. Right?

[Lars Clausen]

I don't know too much about Anthony Holton's output topology. However i know it's usually not twice as safe using twice as many pairs of output devices. Using many paralleled pairs, like 6 or 10 pairs, i would generally find to be a risky approach. One device might have lower Vgs and so get to carry a large portion of the total power. It will fail very quickly.

I think (if properly cooled) the solution with single pairs will run nicely with an output power of 400 Watts RMS. With a bit of "advanced" cooling technique, (mounting the IRF's on a small copper block) i would even find it safe to let one pair (totally) of IRF260 produce 400 Watts RMS of PA power.

I used once an IRF150 as a heater to measure the K/W value of enclosures, and dissipated up to 100 Watts in one single device, and it never failed. The heat sink temperature frequently got above 100 C. They are incredibly rugged, if handled correctly.
The IRF260 has a larger die than IRF150, and so will dissipate even more power with no problem.

All the best from Lars

[peranders]

Lars, you are forgetting safe operating area but I tend to agree with Lars. I manage quite well to have 2 pairs of 2SK135/2SJ50 and got continues power of nearly 500 watts, resistive load and I think normal music and normal speakers won't kill the amp (built in 1984 works still).

I think Mr. Holton and others haven't really optimized the amount of transistor pairs. 10 pairs may not be better (ot more reliable) at all than 1-3 pairs.

[Lars Clausen]

Peranders: Come on, geeks like you and me ... we never forget the Safe Operating Area. <- Joke

A SOA calculation could look like this:

Pmax for IRFP260N is 300 W @ tc 25.
Since it's going to get hot, we can set the tc to 100 C, there is 2 C/W decrease in power dissipation limit for SOA. So we lose 2 x (100 - 25) = 150 Watts of dissipation potential. This means we can only drop off the remainding: 150 Watts / device so with 2 devices : 300 Watts.

Since the highest power loss in a Class A/B amplifier is around 60% of it's maximum output power, (see curve on my hp) this amp with two IRFP260N can safely produce 1.6667 x 300 = 500 Watts. If you could watercool the case of each IRFP260N, to 25 C, then your amp would safely make up to 1000 Watts of power ! But then i guess it would be easier to just use a bigger device like the APT50M50LLL. (www.advancedpower.com) Still in a plastic package a single pair of those babies will produce around 1200 Watts RMS at 100 degrees Tc, before violating the SOA limits. So why even bother to parallel the MOSFETS ?

According to calculations i still think the amplifier with a single pair of IRF260N will safely work as a 400 Watt RMS amplifier.

[Samuel Jayaraj]

Sorry, I am not able to access http://www.aussieamplifiers.com which is the URL for Anthony's great site. I would have liked to attach the schematic so that the discussion becomes even more focussed and fruitful.

[mikeks]

Using all N-channel devices in a linear power is the optimal way to go, ounce the designer has made up his mind to use MOSFETs.

Anything beyond 200W may most reliably and easily be constructed in n-channel MOS silicon.

See power light 9 schematic here:

http://www.qscaudio.com/support/tech...schematics.htm