Need help with JFET phase shift oscillator

[c36041254]

Hi there,
I been out of the forum for long time for some reasons and I had to drop my FM transmitter project idea,..... I'm back on it. I successfully designed BJT oscillators (all kind) and hartley and colpitt with JFETs ( only series fed still struggling with parallel). What I'm much confused with is how to design phase shift oscillator with JFET. What I do with BJT is first design a simple amplifier with unity gain (emmiter by pass cap will increase gain) and with almost half of source voltage across R2 (Base to GND resistor in voltage divider config)so that even a small noise signal can pass and that does the job , But, the same technique is not working with JFET. I'm not putting any ckt because I just do what I explained for BJT oscillator (PhaseShift) and there is nothing but a simple JFET amplifier with phase shift network , so if any one have any working phase shift oscillator with JFET then pls post it.
BTW : I know that the method I explained is not mathematically right, it should be that the gain of transistor >29 and all that math work but I never got where that 29 comes from. So whatever if any one can explain PS osci. with JFET for now, I'll be very grateful. DON"T CARE ABOUT FREQUENCY ANY WILL DO.

 

[on1aag]

Hi Chris,

A phase shift oscillator with a fet is a bit more tricky
because it's difficult to get the required gain at a
low supply voltage.

on1aag.

 

[c36041254]

Thanks! but what is this J310, I tried with 2N4393 but didn't work. I think I should avoid to make PS osci with JFET, even over the internet there are not any PS osci. with JFET I tried a lot, they all just explain that simple phase shift network and gain theory but no practical example.But I still would like to make it just for once, please tel me what type of JFET that is. Thanks again!

 

[on1aag]

Hi Chris,

The J310 is a fet with a high gfs, the J310 is the plastic TO-92
version, the U310 is the metal TO-72 version.
The circuit should work with most fets with a high gfs, like
J308, J309, J310, half of J430 and J431 (these are dual fets),
and of course also all the U versions.
I got curious and built the damn thing, it didn't work, but
I got it working !
I increased the supply voltage to 30 volt and lowered the
source resistor to 3k3. It comes down to a drain current of
about 1 mA and a drain to source voltage of about 12 volt.
I tested ten J310's from Motorola and the output voltage of
the oscillator was between 8 Vpp and 16 Vpp.
I also tried a couple old Siliconix J310's and they worked
without changing the original schematic.
I also tested a couple of old Siliconix J300's and they looked
very similar to the Motorola J310's.
I even tested a couple of 2N4393's in a plastic TO-92 housing
and got them working too.
Give it another try and let us know how many went through the
ceiling.

on1aag.

 

[c36041254]

Hi there!
I don't have that much big DC supply so I tried it in simulator (circuit maker) and it works, in fact it also works with 9V source but the amplitude is only few mV, but even with 30V it takes time to start oscillation. Thanks for help!

BTW: Won't it be temperature sensitive ? Having such a low drain current will make it very unstable oscillator, won't it ???

How did you find values of Rd and Rs, imean where did you start from ?

 

[c36041254]

I was trying to make gain = 1 but that was only peak to peak that's why it was not working so I got that the method is the same with BJT but still don't get why the Vgs ~ 1V necessary, if decreased from that then the output goes to uV level ????

 

[naveen5511]

haiiiiiiiiii
this is a great model.

 

[benormous]

Wow, thanks, this is a great example. Do you have any tips on how to clean up the signal? I am looking to produce a clean sinusoid.