F5 output MOSFETS matched or not?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
valvesound

valvesound

Member
Joined 2011
Sorry guys, I'm sure most of you regulars here are thinking "oh no! another question on this, isn't there enough info here?:rolleyes:" but please, I have searched and read and cannot get a straight answer.

Q: Does the output MOSFETS IRFP240/IRFP9240 have to be complementary matched (N to P)?

Mr Pass has articles mentioning the inputs should be matched, and the outputs should be matched if an amp has N's in parallel and P's in parallel. The F5 output is one complementary pair only as you know.

I am sure in a perfect world they would be matched, but what can I expect if they are not?

Thanks
Willem
 
jackinnj

jackinnj

Member
Joined 2002
tech-diy.com
I just responded to a fellow who asked the question via email -- I just built a couple more F5's and the last batch were F5-R1's -- the Vgs of the opposing pairs were 4.30 and 3.82 -- with the 200 ohm trim-pot and experimenting with the output source resistors i was able to get the THD% down to 0.0022@1W 1Khz. I think that the range which NP described was 0.001% to 0.004%

fwiw, i start with source resistors of 0.68 ohms and parallel in values to determine the optimal value -- this pair of amps used 0.68 ohms on one leg and 0.34 ohms on the other.
 
h_a

h_a

Member
Joined 2006
Due to the very different transconductances of the IRFP240 and IRFP9240 (and most other vertical fets) it is even in theory pretty pointless to attempt N-to-P matchings.

These complimentary fets are never the same, no matter if VGS is matched or not.

VGS matching in the output stage is anyway only done to ensure equal turn-on and not for any sort of distortion cancellation. For current sharing there are the source resistors.

Just keep reading Nelson's articles - everything practically needed is in there ;)
 
jackinnj

jackinnj

Member
Joined 2002
tech-diy.com
h_a said:
VGS matching in the output stage is anyway only done to ensure equal turn-on and not for any sort of distortion cancellation. For current sharing there are the source resistors.
Click to expand...

I disagree -- when you adjust the source resistors (mindful of keeping the bias current the same as your starting point) the THD% can be manipulated downwards. The amplifier to which I referred earlier started out at 0.0088%, and was tweaked down to 0.00022%. You need a fair degree of patience to do this since the amp has to reach a thermal stasis before you start changing out the one of the source resistors, the bias will have to be reset, the offset nulled and the 2nd harmonic nulled each time.
 
h_a

h_a

Member
Joined 2006
Hi Jack,

thanks for the numbers! Nice to hear there's some room left for tweaking.

As you stated previously it is not directly related to VGS matching which is what is commonly done. My post refered to that only, as I understood the original question this way.

I guess some gains can almost always be had by matching the outputs for distortion. However, that works best only for single operating points and is, of course, time costly. (And one needs to have the gear to measure reliably at these low numbers.)

And even then, the distortion generated by normal transducers is one magnitude or two larger than your largest cited THD-number, so at least I leave it at that.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom