[jerrygp]

I have built a few Millet amps and now am going to do the Millet Hybrid Minimax. It's been a few years since these builds have been in existence and I cannot find a source for the Fairchild PN4392 JFET's used in BJT Diamond Buffer.. They are now obsolete it seems, and Beezar.com does not carry them either anymore. Can anyone help with a suitable alternative or maybe have a matched pair or two that they would want to part with? I now just enough about these DIY circuits to be dangerous and would appreciate any assistance.
Thanks...

 

[dsavitsk]

I believe Tomb was switching to the J301. You may need to adjust a resistor with the switch. Tomb will be back on Monday and can update you.

 

[jerrygp]

Thank you. I do have a PM waiting for him when he gets back, but I wanted to explore any thoughts the forum might have first. I will mention the J301 and resistor requirement with him at his convenience.

 

[Avro_Arrow]

MMBF4392 is an indirect replacement.
Same part, just different package.

 

[MisterX]

PM me your mailing address and will drop a couple of pairs of them in the mail....

 

[scottie4442]

I am having the same problem, I have all the parts for my minimax, just cannot find the pn4392, did anyone ever figure out what the replacement is for this transistor.
 

 

[scottie4442]

american semiconductor has some of the pn4392's in stock, I ordered 10 just in case
 

 

[zwack]

I couldn't source the PN4392 either. I don't want to order them from the USA. The shipping cost to the Netherlands will be irrational for just a few JFET's. That's why I choose to buy some J301's. 
 
I read in the BOM that any other N-channel JFET with a Idss greater than 50 will do.  Does this mean that I should only use the ones with a Idss greater that 50??
I have bought a few of these J301's and matched them for current. The two JFET's with the closest match have Idss's of 38.3 and 40.0 mA. None of them have a Idss greater than 50. Is it's okay to use these with the lower Idss's and are they a close enough match? 

 

[zwack]

anyone?

 

[tomb]

Quote:

anyone?

I don't think there will be any issue with them.  If anything, there may be more bias current capability that there was with the PN4392.  The original Diamond Buffer for the Millett Hybrid - designed by Steinchen, used the 2n5486.  In the original configuration, there wasn't enough current leverage in the buffer to allow biases much above the turn-on voltage for the output transistors.  That's because the IDSS of the 2n5486 is only about 20ma (whereas the turn-on voltage for many of the MiniMAX output transistors is between 20-30ma.).  You could run the buffer up a bit higher, but the Steinchen DB used doubled-up output transistors in place of single output transistors with heat sinks.  This meant that you got higher overall Class A bias with lower current set in the JFet loop.  Indeed, the overall limit set by Steinchen was 30ma.
 
The MiniMAX used single output transistors with heat sinks and so we were seeking lot higher biases.  We seek at least 50ma or more with the MiniMAX.  We switched to the J301 during prototyping so that we could increase the bias, but later on decided to double the current mirror in the buffer, instead.  This was done by decreasing RB4 and RB5 to 100R.  Since RB6 and RB7 are 220R, this effectively increased the bias by 2.2 to 1.
 
Once we had done that, the PN4392 was the most readily available and dependable JFET at the time and paralleled the diamond buffer in the Tangent PPAV2.  We've used the PN4392 ever since, until they quit making them.

  However, there's no reason the J310 would not work.  Its IDSS is 60ma max, which means it should work for the MiniMAX even if RB4 and RB5 were upsized to 220R.  At 100R, the J310 should allow biases of 120ma (definitely too much for only 1" tall heat sinks).
 
 
P.S. I may have that history a bit wrong in terms of the DB resistors - we may have upsized RB6 and RB7 to 220R instead of decreasing RB4 and RB5 to 100R.  I can't remember whether we started out at 220 or 100.

  It makes no difference, though - it's the 2.2 to 1 ratio that increases the current over what's in the JFet loop.

 

[zwack]

Thanks. I will use the J310's that are matched at 38 and 40 Idss, assuming that the match is close enough.