Re: 2N3819 JFET spice model

from 1998 . . . .

RL

It's never too late to thank your fellow engineers.

.model 2N3819 NJF(Beta=1.304m Betatce=-.5 Rd=1 Rs=1

• Lambda=2.25m Vto=-3 Vtotc=-2.5m Is=33.57f Isr=322.4f

• N=1 Nr=2 Xti=3 Alpha=311.7u Vk=243.6 Cgd=1.6p M=.3622

• Pb=1 Fc=.5 Cgs=2.414p Kf=9.882E-18 Af=1 mfg=Vishay)

from current LTspice distribution.

2N3819 was always poorly defined w/r data sheet. Many vendors, many wafers, many processes were used.

RL

Yeah, using SPICE on a part with a 3:1 range of pinchoff voltage is going to lead to a lot of reasonable-looking wrong answers (my least-favorite kind).

Cheers

Phil Hobbs

Wow, that's an amazingly crappy part, even for a JFET.

The Fairchild datasheet shows a 15:1 range of pinchoff voltages, and that's from just one vendor.

Using SPICE on a part like that is going to generate a lot of reasonable-looking wrong answers (my least-favorite kind).

Cheers

Phil Hobbs

I've always thought of 2N3819 as the JFET analog of the 2N3055 BJT. It's a convenient part number into which you can reasonably throw a whole lot of different JFET products. Test what you fab, and if they don't meet tighter specs for a targeted part number, just call them

2N3819.

They must have made them in saucepans. Actual _good_ JFETs, such as the CPH3910, have much much tighter specs (and much lower noise).

Cheers

Phil Hobbs "40,000 of them in stock"

I always assumed it reflected the manufacturing processes available at the time. By that time, the planar diffused process was pretty common, and it was much better controlled than using the grown-junction or alloy-junction processes. Good engineers learned how to design circuits that could tolerate the variations in those processes and the variations produced by varying operating temperatures and voltages. Ion implantation produced much tighter ranges, but it was still not a common manufacturing process in the mid 1970s.

IIRC the usual planar JFET of that era has a Vp range of around 3:1. This piece of crap's range is 15:1.

Cheers

Phil Hobbs

I think the 3:1 range for one part number was achieved by binning a series of numbers (typically 4 or more) from the same mfring process batch.

RL

My recollection is that there were not very many other commercial JFETs for comparison. And the specs for the MOSFETs were even more poorly controlled because they did not understand the manufacturing hygiene required to control the surface states.

I remember the standard bipolar transistor hfe spec had a 3:1 range. They narrowed by testing and selecting parts, so there were lots of "families" with consecutive numbers and similar specs except for the buckets for the selected parameter (hfe, BVceo, fT, etc).