Just a quickie,

I've just dug up 2 pairs of these Hitachi MOSFETs. I would like to build an amplifier from them in a couple months maybe. My question is: Is it worth it? I know they are elderly by current standards, are there any good circuits to use them in Hi-Fi Land? It just seems a shame to waste them...

Many thanks.

These MOSFET's are classics, that are very hard to find.
I'm just in the process of designing a integrated amplifier with 10 complementary pairs for each channel.
They have a Rds of 0,7 ohm so you need a couple of them to drive a loudspeaker unless you use NFB and I would like to keep NFB below 6dB if possible.
They are termally stable without any resistors at the source and that's why I love them.
I looked at N. Bayer's design and it seems OK.
Not the best in the world, but OK for nice sound.


since you only have two complementary pairs the ouput impedance is to high to be used with loudspeakers,
but you can use them for a headphone amp if you want.
If you run 20 mA through the driverstage you don't even need to include them in the NFB circuit if you have one.


There is no need to give up that quickly. Dig into some Elektor designs of yesteryears. Some of the designs that come to mind are the AXL, Mini Crescendo and Crescendo designs. The last one is supposed to have a reputation as an oscillator but with my own PCB design I have had very little trouble getting this amplifier going. Then there is the 1KW Elektor design, but then you need four pairs of Mosfets.

Alternately, have a look at Russel Brenden's article entitled 'Class B in a new Class' in Electronics World of the year 2000 (I forget which month exactly). Use the same circuit design and substitute the Mosfets suggested by the author with the ones you have. I can assure you that it has a brilliant sound. It is a Class B design with a modified feedback loop that makes the amplifier sound like single-ended Class A. I have used the 2SK1058/2SJ162 pair with this design with much success.

Another great design using exactly the same Mosfet pair that you have, is an Electronics World design by John Linsley Hood. It appeared sometime during 1995, 1996. According to J.L.Hood, that design sounds very open and least offensive. Electronics World has a back issue service and they can help you with photostat copies of the original articles.

So persevere and all the best,

I'm afraid you're getting your articles mixed up. Russell Brenden's article was entitled 'A new 100W Class-B topology' and appeared in EW June 2000. The 'Class-B in a new class' article was by Mike Renardson and was in EW April 1998. Details of this amp can be found at Mike's website:


As for the JLH designs, an index of EW articles (along with a means of purchasing copies) can be found here:

http://www.softcopy.co.uk/electronics world index.htm


Don't forget about the mfr of the MOSFETs: Hitachi. They had a few simple schematics for good performing amps in their databooks on the MOSFETs. Something like 5 transistors + the MOSFETs.

Hitachi even produced an amp based on this design a a number of years.


Thanks for the correction. It was indeed Russel Brenden's article that I was referrring to and got the title of the article mixed up.

By the way, in that article the author refers to using Vertical FETs but aren't the ones used in the circuit Lateral types. If I can recall correctly, it was Helix who in this forum mentioned that he used the latest Mosfets from Exicon in this design. I have used 2SK1058/2SJ162 successfully after burning up a pair of IRF9540/540.

Thanks again.
yeah the original hitachi designed amp used lateral mosfets but vertical ones can be used instead with very few changes.

1st the difference between the 2......

lateral mosfets are primarily for audio and thus exibit very low distortion.

vertical mosfets (aka. hexfets) are primarily for switching and so can have a higher distortion .... but they also have higher transconductance and lower Rds(on) figures which is a plus and they cost generally less than 1/5 of lateral types.

if you use vertical mosfets in the circuit that is designed for lateral types you will need to add a Vbe multiplier based bias circuit.

The only article I have from EW March 1999 is the 'Class A to 300W' article by Colin Wonfor. Email me if you would like a copy. If there were other articles in this issue that you are interested in, let me know which and I will have a trip to a (relatively) local library and obtain a copy. It might take me a few weeks to get there though.