Joining 2N3819s to form a dual gate mosfet ?

[Jolly 7]

I was wondering if connecting the two drain leads together and the two source leads together from a pair of 2N3819s would create a dual gate mosfet ? Or is this idea unscientific and unworkable ?

[Radio Wrangler]

The 2N3819 is a junction FET, not an insulated gate FET And it's a depletion-mode device as well, so there are significant differences to start with.

However, you can stack them, drain of the lower device to source of the upper device to make a JFET cascode circuit.

The dual gate MOSFET is essentially an integrated cascode of two MOSFETs Their benefits are good gain retained to high frequencies due to reduction of the Miller effect, and good reverse isolation.

Real dual gate mosfets have long been cheap, but now types are vanishing rapidly, most of the newer types for 5v operation have gone. The era of rolling your own may be upon us.

David

[Maarten]

A dual gate MOSFET is certainly impossible since they are JFETs. I'm not sure whether there are any designs out there using dual gate JFETs but it might be possible to use it as a substitute circuit.

[G6Tanuki]

It seems a bizarre way to go about things: surely it would make more sense to use a pair of 2N3819 in a proper 'cascode' circuit where you get to control the gate-biases as need be?

[remember that the likes of the 2N3819 have wide production-tolerances so 'tweaking' the bias is almost essential for optimising gain/linearity in amplifiers]

See https://cdn.eeweb.com/articles/app-n...1302505476.png or

https://www.qsl.net/va3iul/Homebrew_...ascode_LNA.gif

[Argus25]

Originally Posted by Jolly 7

I was wondering if connecting the two drain leads together and the two source leads together from a pair of 2N3819s would create a dual gate mosfet ? Or is this idea unscientific and unworkable ?

I assume you are talking about this sort of design that has cropped up in transistor radio projects:

http://theradioboard.com/rb/viewtopic.php?t=2228

It is simply cascode, except that the upper fet's input, instead of being clamped to a DC and grounded to RF reference, has the local oscillator injected there. So to some extent the overall arrangement behaves as a two gate Fet. The upper input is called the high level input, the lower input the low level input. This is because different signal levels are tolerated at those "ports" prior to device saturation.

If you have a look on page 7 of this article below (on the Hero Jr robot and its radio board), there is the once ubiquitous RCA dual gate Fet, the 40673. As you can see from the circuit there, the L/O is mixed into one gate and the received signal into the other. If you want to play with dual gate fets for your AM radio applications, the 40673 is the one to get, very easy to bias as shown and use as a mixer:

http://worldphaco.com/uploads/SAVING...dio_Board..pdf

(You might also be interested in the 455kHz IF in that Hero Jr circuit, because it uses a ceramic filter and has an IF and AGC amplifier IC in it, the Motorola MC1350, all which would work well for an AM radio project as does the standard arrangement of IF transformers)

[Jolly 7]

Thanks everyone for the cascode circuit idea. At first I thought it might be a mis-spelling, but it is indeed spelt cascode and not cascade. I have learnt something new today.

@Argus25: have built only one circuit ever using the brilliant 40673 mosfet, an amplified AM loop antenna, which uses a hoola hoop with enamelled copper wire inside as a frame. It works marvellously well and I can rotate it via a geared motor too.

The 40673 has been described on another website as a 'zombie component that refuses to die' ��

[Radio1950]

Hi Jolly
I have a vintage coms radio which uses that 40673 device, and I happened to read that "Zombie" article a week ago, whilst researching low noise FETs.
I wouldnt pay much attention to that "opinion" if I were you.
Seems the AD8307 is rubbish also.

[Synchrodyne]

The cascode jfet was used commercially. In part that may have been because it became available maybe a couple of years before dual-gate mosfets, and even then, some makers waited a bit longer again (about another year I think) until protected-gate dual-gate mosfets became available before using them.

FM tuner RF amplifiers was an early application for the cascode jfet. Scott was one of the first, with a jfet shunt cascode RF stage in its top-end FM front end towards the end of 1965. Heathkit used a jfet shunt cascode RF stage in its AR-15 of late 1966, and I think stayed with jfet-based FM front ends for about a decade, even though it had adopted dual-gate mosfets for AM in 1969.

B&O reworked the front end of its Beomaster 5000 tuner to use a jfet series cascode in place of the original germanium bipolar RF amplifier.

In general though, jfets gave way to dual-gate mosfets for FM front-end applications by the end of the 1960s. Economics may have been a contributing factor, in addition to relative performance. One dual-gate mosfet probably cost a bit less than two jfets.

There were also commercial applications. The GEC RC410R HF receiver (1967) had two tuned RF stages, each using a jfet series cascode with agc bias applied between the source and gate of the lower unit. Apparently the original RF amplifier design was bipolar, but with this the desired noise factor and intermodulation performance could not be achieved at higher frequencies, so a change to jfets was made.

The Marconi Hydrus point-to-point ISB receiver (1968) used series-cascode jfets in the RF amplifier, mixer and IF amplifier positions. In the RF (and I imagine the IF) case, agc bias was applied to the gate of the upper unit. For the mixers, the signal was applied to the lower unit gate and the oscillator input to the lower unit source. Thus both the signal and oscillator were looking into a cascode. Given that the first oscillator input could be as low as 41.5 MHz and the IF output was at 40 MHz, and wideband (± 0.6 MHz), my inference is that it was thought desirable that the oscillator input as well as the signal input be screened from the output in order to minimize regeneration opportunities.

The Eddystone EC964 marine spot-frequency SSB receiver used a triple jfet array for its 2nd mixer, said to be chosen for good signal handling in a position were low noise was not essential. This could be seen as a jfet series cascode with the upper unit replaced by a source-coupled pair. The oscillator went into the lower unit gate, the signal into the left-hand upper unit gate and the output was taken from the right-hand upper unit drain.

Thus were there various ways in which the jfet cascode was used. Whilst the signal input was (mostly) the lower unit gate, agc bias and oscillator injection points varied. With dual-gate mosfets, there seemed to be more uniformity, with signal nearly always on gate 1, for amplifiers, agc bias on gate 2 and for mixers, oscillator on gate 2. An exception to the latter was in VHF TV tuners, where signal and oscillator both went into gate 1 (It took me quite a while to work out why that was.)

There were also hybrid cascode combinations, In the RF amplifier of its EC958 HF receiver (1969), Eddystone used a jfet in the lower position and a single-gate mosfet in the upper position, series-connected, this for greater robustness against RF voltage spikes than would be obtained with a gate-protected dual-gate mosfet. In the late 1970s the combination of a jfet in the lower position and a bipolar in the upper position (previously seen in HF applications from c.1969) was adopted for car radio AM RF amplifiers in the US at least, and was carried over into high quality AM stereo tuners in the 1980s. It was said to be preferred over the dual-gate mosfet.


Cheers,