2SK170 limit

[revintage]

Have not worked with 2SK170 at high B+ before. For a future Zen-inspired preamp-project I intend to use it like below with B+=90-100V, Vd=35-38V, Id=8mA. This is quite close to the limits of 40V/10mA.

Will there be any immediate risk of frying it due to the high B+?

[AndrewT]

Hi,
the high Vds of 35V combined with the medium Id current of 8mA requires a junction dissipation of 285mW. That's 71% of maximum rating.
Far too hot for continuous duty.

[revintage]

Far too hot for continuous duty.

Hey Andrew,

Is this something you have found out empirical as this info is nowhere else to find? Toshibas datasheet only states 400mW at 25 degrees. Found another typical JFET where dissipation derating above 25degrees was close to 3mW/degree.

Though my main concern was running B+ of two times max Vd, .

[revintage]

To clarify, my concerns are what happens at "power up" and "power down".

[juma]

Originally Posted by revintage

To clarify, my concerns are what happens at "power up" and "power down".

It's very easy to find out - eventual sacrifice of one JFET will not cost much and you'll know for sure.
OTOH, cascode is an easy way around the Pd problem (with additional benefit of characteristic's linearization and larger swing capability).

[revintage]

Originally Posted by juma

It's very easy to find out - eventual sacrifice of one JFET will not cost much and you'll know for sure.
OTOH, cascode is an easy way around the Pd problem (with additional benefit of characteristic's linearization and larger swing capability).

Hey Juma,

Good idea, this is cheap stuff.....

Maybe one should try a MOSFET capacitance-multiplier with long time-constant to get around the evntual start up problem and also get better PSRR.

Will go singlestage Zen-style(not BOZ), cascode is two stage.

About linearity and swing this one with a single 2SK170 will put out ca 25Vrms into 10k(Zout below 500ohm) at 11dB gain with below 0,1% THD. This as shunt feedback is a lot more effective than current feedback commonly used here.

Question though is if it will sound as good as the figures indicate?

[iko]

Originally Posted by revintage

To clarify, my concerns are what happens at "power up" and "power down".

Valid concern IMHO. FWIW 2sk117 gives you another 10V to max Vds. Another way to deal with this would be to go the cascode way, by using a jfet with larger Vgs(off) on top. I use a 2n3819, but many others would qualify. My 2c.

[juma]

Originally Posted by revintage

Hey Juma,
...Will go singlestage Zen-style(not BOZ), cascode is two stage.

No, it's still one gain stage with two active devices

Originally Posted by revintage

... with a single 2SK170 will put out ca 25Vrms into 10k(Zout below 500ohm) at 11dB gain with below 0,1% THD. ....

So, you want that poor JFET's Drain to sit at 40V and swing 35V (peak). It means that you'll have Vds=75V during the positive peak. Doesn't seem viable to me.
Zout will be Rd (6k8) || RLoad.
THD figure is unrealistic too.

[revintage]

Originally Posted by ikoflexer

FWIW 2sk117 gives you another 10V to max Vds. Another way to deal with this would be to go the cascode way, by using a jfet with larger Vgs(off) on top. I use a 2n3819, but many others would qualify. My 2c.

Hey,

Checked 2SK117 and unfortunately it had 25% lower dissipation. Otherwise it look good with much lower Ciss.
About cascode this isn´t an option as I told Juma above that I don´t want to go two-stage. Will anyway sim a cascode side-by-side.

[EUVL]

2SK223. 80V, 400mW.
Don't blame me if you fry one.

Patrick