Vishay acquired IR, last time I read what Papa said, their P channel do have the anomaly. However, I would not consider negative phase second harmonic as problem. On the contrary, it is a feature I (and also a few others) prefer to have in an amp at some, preferably adjustable, small amount. But there are also other ways to have it.
Thank you Papa, I'm rather thick early dawn.🙂
Yes Marconi, since you need 2 pairs of J201/K1530 for a stereo F5 build, they will be about $70-80 for matched NP pairs, a bit "easier" compared to $120 for the 6 matched NP pairs of J313/K2013 you need for similar level of performance otherwise.
Yes Marconi, since you need 2 pairs of J201/K1530 for a stereo F5 build, they will be about $70-80 for matched NP pairs, a bit "easier" compared to $120 for the 6 matched NP pairs of J313/K2013 you need for similar level of performance otherwise.
A single pair of 2sj201/k1530 for an F5 don't need to be matched.
4 150W, 12A mosFETs at standard prices makes up the output stage of a stereo build.
To match that thermal performance from 25W 1A mosFETs one would need 6 paralleled sets in each half of each output stage.
That adds up to 24 matched devices at the matched price.
> A single pair of 2sj201/k1530 for an F5 don't need to be matched.
That depends on whether you care about even harmonics cancellation.
Not for equal current sharing.
Patrick
That depends on whether you care about even harmonics cancellation.
Not for equal current sharing.
Patrick
Yes Andrew, I agree. However, matched output pairs on F5 build will give an advantage of easier adjustment for consistent distortion residual spectra between channels, provided the input jfets are also matched.
I was helping Marconi to save a bit on Juma's F5 with 3 parallel pairs of J313/K2013, your calculation confirm that a single pair of J201/K1530 to replace them will be thermally fine.
You obtain those from a certain member with 4 syllable name selling remaining stock of a currently non active member with 2 syllable name based in balmy city with 3 syllable name. Great are the entertainment power of the pairs performing in a Nelson Pass Diy builds. Absolutely forbidden for sith cloners and dark side copycaters.😀
got mine 2 pairs , not paired , long ago , from 3 syllable city , 2 syllable member
I wish him well , whatever (else) is he doing now
I wish him well , whatever (else) is he doing now
Thank all for your advices
is this basic matching technique valid? (but not ideal):
connect gate to drain
feed drain and source with a constant current psu (lm317+resistor) 200mA for 313 or 1A for 1530
measure the voltage Vds, match them accordingly.
Tried this with 4 pairs ECX10N20 ECX10P20 from a friend. Got consistent values for all P and N, but the value for P and N was different (1.8 and 2.4) . Is this normal? are they maybe fake?
is this basic matching technique valid? (but not ideal):
connect gate to drain
feed drain and source with a constant current psu (lm317+resistor) 200mA for 313 or 1A for 1530
measure the voltage Vds, match them accordingly.
Tried this with 4 pairs ECX10N20 ECX10P20 from a friend. Got consistent values for all P and N, but the value for P and N was different (1.8 and 2.4) . Is this normal? are they maybe fake?
That's perfectly normal. P-N matching is normally difficult to do, but as long as all your P-devices match and all the N-devices match, then you should be fine 😀
/U.
your method actually sets Vdg to zero volts and measures Vgs.
It is better to use Vdg similar to what it would be in the amplifier. That makes the dissipation inside the device MUCH HIGHER. Temperature control then comes to dominate the methodology.
Similarly the current should match the amplifier's typical currents. But they vary from 0Apk to xxApk. This is where curve tracing comes in. A good match is where Vgs is the same at all the currents tested.
It is very usual for the N & P channel to measure differently.
The NFB helps to correct this and brings performance to a very acceptable standard.
The Ns need to match when paralleling.
The Ps need to match when paralleling.
Last edited:
Andrew covered the basics, but your method is still better than no matching. Make accurate note of your measurement to include heatsink temperature and get your feet wet. The notes will help you later should you decide to tweak or upgrade.
They are expected Vgs values at 1A for Laterals. I have never used the Exicons but have read preference for them compared to the Renesas. However, datasheet values shows typical Vgs for N channel at 1A is lower compared to the P. Hopefully somebody with experience with Exicon or Magnatec can share his opinion.
Last edited:
If my conjecture is correct, then all Lateral mosFETS are manufactured in the same Japanese factory.
Only the packaging differs between the various assemblers.
This conclusion followed the complete world wide "out of stock" for ALL types soon after the big earthquake hit southern Japan.
And had to wait till the factory was rebuilt a couple of years later.
Only the packaging differs between the various assemblers.
This conclusion followed the complete world wide "out of stock" for ALL types soon after the big earthquake hit southern Japan.
And had to wait till the factory was rebuilt a couple of years later.
Could be Andrew, but I have the impression there are subtle differences between Renesas' and the rest, could be packaging related. Chuck Hansen measured significant enough difference on P channel parts for him to prefer Exicon's.
If I understand correctly Laterals can't be NP matched but 2SK2013/2SJ313 and J201/K1530 can?
I don't need NN and PP matching because I will use only one pair per channel...
I don't need NN and PP matching because I will use only one pair per channel...
- Status
- Not open for further replies.