best match/complement for SFH9240 ?

[janusz]

What would be the best match/complement for Fairchild's SFH9240?

Is it IRFP9240 or IRFP244 or IRFP340 or their Fairchild equivalents or whatever?

I know that N and P devices should be matched within their type but I remember reading somewhere (some technical paper by one of the major manufacturers of MOSFETs) that there were also merits in matching P with N as well.

Looking at data sheets I'd say that its either IRFP240 or IRFP244 but I may be wrong. Has anyone tried to match SFH9240 with any of these or other devices?

Could anyone comment on any merits of matching P with N devices?

Thank you,

[BobEllis]

I don't have the technical background to do a proper analysis, but the Fairchild IRFP240 and SFH9240 look like a good match on transfer characteristics. I used them when I built my A75s and am happy with the sound, although I don't have a distortion analyzer.

IIRC, Nelson Pass once said that the Fairchild pairs are well matched, but IRs are not as good. Someone else mentioned that a good match for an IR IRFP240 would be the IRFP9140 if your voltage requirement could be met.

It seems to me that matching the transfer characteristics of the N and P channel devices in a complimentary follower stage will reduce the intrinsic distortion, meaning that the feedback has less to correct.

Is this what you meant by the merits of matching P and N? or were you looking to match the Vgs of each side? The Vgs of the SFH9240 is slightly higher (around .1 to .2 V) than the IRFP240. In a lot of 100 of each type, all of the 9240s had higher Vgs than all of the 240s, so I don't think it is possible to get an exact match there. It may be possible using an IR IRFP9240, I think that they have slightly higher Vgs than the Fairchilds. But then you may have more of a mismatch on the rest of the transfer characteristic.

[gl]

Hi janusz,

I have paired the SFH9240 with the IRFP240 (IR) with good luck. I matched 150 IRFP240's and 100 SFH9240's and found that the SFH9240's had a tighter grouping of Vgs values. Vgs on the SFH9240's was also somewhat higher IIRC. I will double check this.

You should search for SFH9240 on this site. I have posted several times about this part. I'm still trying to find out exactly where the die design came from within Fairchild. I would like to think that it was the old Harris/Intersil IRFP9240 design but for various reasons I suspect that it is part of the IP they bought from Samsung.

Cheers,
GL

[janusz]

Thanks Bob and Gl,

True, Fairchild's IRFP240 looks like a good match for SFH9240 but it is no longer in production and difficult to obtain. Unfortunately, I cannot use IRFP9140 as no load supply voltage probably will be around +/-60V. Too much for this device.

I purchased SFH9240s some time ago rather than IRFP9240 (also not in production any more) because of problems mentioned by Nelson Pass in his article published in AudioXpress and repeated here.

The paper advocating, among other things, matching of N and P devices mentioned reduction in distortion. I read it long time ago and do not remember much so you must be right about matching transfer characteristics to achieve some improvement in this area. As I haven't much test equipment at home I'll most likely use the method recommended by Elliot from ESP unless you would recommend some other method.

Later I bought 25 IRFP240s so I'll try to match these with SFHs but was also considering other possible options. I remember reading something here about IRFP340 being a better match for IRFP9240 than IRFP240. Looking at data sheets it appeared to me that possibly IRFP244 (which is quite cheap) might be a somewhat better complement to SFH9240 than IRFP240.

I do not know the origins of SFH9240 but Fairchild used to produce IRFP240s and IRFP9240s in the past. I have no idea why they abandoned this production line. So now Fairchild produces only IRFP240B which does not look like an ideal match for SFH9240 and SFH9240 which replaces Fairchild's IRFP9240.


Thank you for your comments,