2n3819 bias problem.

[SteveG]

Hi,

A while back I built a FET based guitar pre-amp using the 2N3819 device. It's a very simple single stage common-source amplifier powered by a 9v battery, 4k7 load (actually a pot) and 470R source resistor to set bias. Sounds great, etc - so far so good!

My problem is in trying to build more of them. I can't seem to find 2N3819's that bias the same. The 470R source resistor has to be increased considerably to get Vd to be 1/2 supply voltage. That in turn eats into the headroom to a point where the design isn't viable.

Anyone know where I can obtain 3819's that will bias like the first one I tried? So far I have 2 devices that work, and 35 that are unusable. Both the ones that work are Vishay Siliconix, but I have ten that don't. None of the 25 Fairchilds I've tried work. (Work = work in this application).

Thanks in advance!

Steve

 

[Gus]

I get about 4 fets per 100 that will work the way I want for microphones.

IMO the best way to build stuff with fets is to buy then in large amounts(>100) and pretest them.

 

[AMZ-FX]

Try a J201... Small Bear sells them in small quantites and so does Mouser. I find they are more consistent that many other jfets. The 2SK117 is a good one as well.

regards, Jack

 

[PRR]

> where I can obtain 3819's that will bias like the first one I tried?

2N3819 is a very loosely-specified part. Vgs(off) could be as high as 8V or as low as... hmmmm, they don't even give a minimum Vgs(off)! Idss could be 2mA or 20mA. You need only about 1mA, which any 2N3819 will pass, but high-Idss and high Vgs(off) parts could take over 6V to cut them back to 1mA, leaving nearly nothing left from a 9V battery.

In short, you may need much-more than 8V supply voltage to set an operating point for any 2N3819.

I think the real use of a part like this is a high-current amplifier with no DC resistance in the load: a tuned radio amplifier. In that application, you generally start with a LOT of current for best weak-signal gain and noise, then apply AVC grid-bias to turn it down for typical stronger signals.

Try 2N5484: it is more the size you need for 1mA current at 0.5 to 2.0V of bias, and you are likely to find more workable units per baggie. This is selected from Process 50, the workhorse JFET wafer, so should be very common.

EDIT: Jack has a good point: J201/J202 is a smaller Process 52 part aimed at low-current audio. J201 maxes out at 1mA so may not be best for 5K load on a 9V battery, J202 may be a better bet here. If you play with various JFET projects, it might be wise to get a baggie of each, J201 and J202. Mouser has them at about $0.40 each, about $8 for a bag of 25. (A dime each if you buy 5,000....)

 

[SteveG]

Cheers for the replies.

No sign of the 2N5484 or J202 in any of the catalogues I have here (I should explain that I'm in the UK, which is a bit like having one arm tied behind your back). They do list the J201 though, but I wouldn't want to run that at max current.

Looks like another of life's necessities that I'll have to have shipped from thousands of miles away. :roll:

Thanks again,

Steve

 

[CJ]

:guinness:

 

[PRR]

> No sign of the 2N5484 or J202 ...I'm in the UK

Go to http://uk.farnell.com/ Farnell InOne, search for "J202", check the box "Include Newark InOne Direct Ship..."

I wound up at this monster URL, which may not work for you.

After wading past J201-numbered PC-toys down to Transistors, JFET, I got:

Order Code 4475215
Part No. J202 VISHAY SILICONIX
In Stock £0.24

Can't complain about the price. (Except you need almost 100 to make a minimum order.)

> another of life's necessities that I'll have to have shipped from thousands of miles away.

All these parts are made in the furthest (and cheapest) corners of the earth. They sure don't make them in Newark NJ USA (which at one time was a vacuum-tube hub). (Anyway, I think Newark Electronics is in Chicago now, and they don't make JFETs there either.)

What is Newark InOne Direct Ship?
Newark InOne Direct Ship is exactly what it says. You can now order products on this site from our sister company, Newark InOne, and have them shipped directly to you from the USA, within 3 days.

So you pay with your local money, taxes are handled, and you might have it in about the same time it would take to find a moment to pop into the local parts-shop.

> They do list the J201 though, but I wouldn't want to run that at max current.

No harm will happen, it will just sit there saturated. Change the load resistor to 10K, maybe half of J201 will work well enough for jazz.

 

[PRR]

> I thought they were just a run of the mil current source.

They are. But the third leg makes them a voltage-variable current source. And that's an Amplifier.

Steve's design needs a device from nearer the LOW-voltage end of the run of the mill, not the high-current ample-voltage end that is popular in RF work. Process 52 is small, good only for low current, what we want. The lowest end of P52 is sold as J201; Steve needs a little more so J202 is more his size, but unfortunately not popular enough to stock locally. (You KNOW there is someone like you with a 2,000-bag of J202 in his county, but where?)

 

[SteveG]

PRR and CJ - Thanks again for the replies.

I obviously didn't look hard enough in the Farnell catalogue. :roll: (Edit: Ah, It's from Newark - that explains it)

IMO, there's nothing like a single FET pre-amp for driving the front end of valve guitar amplifiers. Kind of like having an extra 1/2 an ECC83 in there.

-Steve

 

[ajini]

@PRR saved my life... I tested 5 of them and all have Vgs(off) < 300 mv! the range for this transistor is extremely wide and not specified on the datasheet.

 

[gyraf]

Welcome, Ajini!

Note that you're replying to a 20 years old thread.

Now, 20 years later, it's close to impossible to find a real 2N3819 - the market has been flooded by fakes. So you may not even have a 2N3819 at hand..

/Jakob E.

 

[Jarno]

Well, I get where you are coming from in an industrial context, but for hobbyists, there is some low volume old stock here and there (NOT in China though!).
There are a few vendors on eBay which have old stock from places where my employer (Philips) used to be, so France, some places in Germany, UK. They did a lot of electronics assembly there, TV's, radio's and such. And there's often small amounts of stock left, for SMD also, I have some reels of quad opamps BS847 transistors etc. (proceed with caution for SMD, because more often than not, it has been stored in a shed, not a climate controlled warehouse).
Am wondering about the type though "we" did have semicon fabs so there's the "B" range of transistors (BC,BF etc), I think that range also had JFETs.
So, I bought some old ones from France recently, because I did a small on-board preamp buffer for active basses, expecting boards today, so I will let you know if the 2N3819's turn out to be fake

 

[Jarno]

Well, they are JFETS Definitely not 2N3819
I tested a bunch of them on the PEAK Atlas DCA75 (which itself has quite large errors on some measurements).
IDSS in spec for all of them, other parameters, not so much
Also tested a bunch I got locally from a parts store with a lot of old stock. Somewhat larger spread on values, but all pretty much within spec.


So yeah, proceed with caution!

 

[micolas]

Well, they are JFETS Definitely not 2N3819
I tested a bunch of them on the PEAK Atlas DCA75 (which itself has quite large errors on some measurements).
IDSS in spec for all of them, other parameters, not so much
Also tested a bunch I got locally from a parts store with a lot of old stock. Somewhat larger spread on values, but all pretty much within spec.
View attachment 130623

So yeah, proceed with caution!

The ones bought from eBay, have Vgs(off) out of spec, are they fake 2n3819?
Or have you tested them in the circuit and they work reasonably?

 

[Jarno]

They have IDSS within spec, but VGS(off)/gfs/rds(on) are all out of spec (more like, nowhere near the spec), see excelsheet posted two posts back. I think they are rebranded JFETs of another type, I have tried them on two inputs, but ran out of headroom too quickly, so not proper bias (didn't investigate further), had real ones also, so I used those.

Like I said, it is a bit of a crap shoot, there's likely to be old parts stock around, because industrial manufacturing tends to leave a lot of residual parts (well, not a lot in large scale manufacturing terms, but a ton of parts in hobbyist terms), the challenge is in sniffing out the authentic parts. I'll try a few local shops next.

 

[micolas]

They have IDSS within spec, but VGS(off)/gfs/rds(on) are all out of spec (more like, nowhere near the spec), see excelsheet posted two posts back. I think they are rebranded JFETs of another type, I have tried them on two inputs, but ran out of headroom too quickly, so not proper bias (didn't investigate further), had real ones also, so I used those.

This happened to me with the modern CEN 2N3819 from Mouser.
In the U87ai circuit, with Idss of 8...9mA I have disappointingly low headroom, with all the efforts of BIAS on the oscilloscope as correctly as possible.
In the same circuit, NOS Fairchild and Motorola with Idss of 8...9mA sound excellent, with good dynamic range, good headroom.
Unfortunately, we no longer find NOS stocks in the right specifications.

 

[Khron]

So what's wrong with something found in plentiful stock at major distributors, like a J113?

 

[Migs 31]

So what's wrong with something found in plentiful stock at major distributors, like a J113?

I have recently used J113, BF 256b (careful not pin equivalent) and even J111? if I remember well.
But everybody wants the "original" part, original caps, real VF14 etc because they are the ones that have the magic ingredient (read toxic chemicals) that will make my build sound like the 1967/1976 item.

What do I know? I'm just a poor sound engineer DIY'ing along, my eyes are dim, my mind is weak...
But my ears are still working OK, people say so, at least...
M

 

[Khron]

I have recently used J113, BF 256b (careful not pin equivalent) and even J111? if I remember well.
But everybody wants the "original" part, original caps, real VF14 etc because they are the ones that have the magic ingredient (read toxic chemicals) that will make my build sound like the 1967/1976 item.

What do I know? I'm just a poor sound engineer DIY'ing along, my eyes are dim, my mind is weak...
But my ears are still working OK, people say so, at least...
M

I asked what's wrong with the part, not what's wrong with the "vintage fetishists"...

 

[Migs 31]

I asked what's wrong with the part, not what's wrong with the "vintage fetishists"...

Well two birds with one stone.
And I am telling you:
Nothing wrong with that (J113) part or even the poor 2n3819.
What is wrong here, is people's obsessions (includes me at times).
So many times in audio: not the equipment is at fault, its the users.

M

PS: try BF 256b and hear...