Hi, I'm working on a project of dc coupled jfet preamplifier.
One aspect of fet i dont love is the parameter too much variable, non contrasted by the Gm relatively low instead of bipolar hfe, so I have not much experience using fets in audio amplifier application.
Anyway, before chose the components to use, and successive real measurements, i traing some simulation with Tina-ti or LTspice.
My dubts is about bias and zero tempco conditions, basilar for the DC coupling.
From many sites i read "zero tempco is realized subtracting at VGSoff about 0,63V..
Simulation software I used permit only a variation of ambient temperature, end models have generic temperature values, the same for more device.
I tried a simulation a simply circuit to verify if this teory is verificable, a simply schematic with a jfet with his power supply source and current meter on Drain Source loop, and a triangular wave generator in the gate drain loop.
Running the the analisys in parameters temperature step mode, the results is not choerent with the teory.
With more jfet models used, in the simulation windows the lines cross each other (Zero tempco) in correspondence of a value of Vgs near 1 volt from VTO, and anyway far from the 0,63V of the theory.
Have someone an explanation? Is the jfet parameters too much generic? Changing the ambient temp is not like change junction temp? What other?
thanks
fabio
One aspect of fet i dont love is the parameter too much variable, non contrasted by the Gm relatively low instead of bipolar hfe, so I have not much experience using fets in audio amplifier application.
Anyway, before chose the components to use, and successive real measurements, i traing some simulation with Tina-ti or LTspice.
My dubts is about bias and zero tempco conditions, basilar for the DC coupling.
From many sites i read "zero tempco is realized subtracting at VGSoff about 0,63V..
Simulation software I used permit only a variation of ambient temperature, end models have generic temperature values, the same for more device.
I tried a simulation a simply circuit to verify if this teory is verificable, a simply schematic with a jfet with his power supply source and current meter on Drain Source loop, and a triangular wave generator in the gate drain loop.
Running the the analisys in parameters temperature step mode, the results is not choerent with the teory.
With more jfet models used, in the simulation windows the lines cross each other (Zero tempco) in correspondence of a value of Vgs near 1 volt from VTO, and anyway far from the 0,63V of the theory.
Have someone an explanation? Is the jfet parameters too much generic? Changing the ambient temp is not like change junction temp? What other?
thanks
fabio
Last edited:
don't know offhand if the temperature coefficient effects are included, modeled right in generic Spice Jfet models
the next question would be if you can get any model with realistic values even if the effect is included in the template
you should be able to run repeated sims while stepping the global Spice temp value - ugly but that's the legacy way Spice works
my recollection is that people say the zero tempco bias point of most fets is too low for most audio application due to poor noise and gain
the next question would be if you can get any model with realistic values even if the effect is included in the template
you should be able to run repeated sims while stepping the global Spice temp value - ugly but that's the legacy way Spice works
my recollection is that people say the zero tempco bias point of most fets is too low for most audio application due to poor noise and gain
You might be right, as I just wrote in I have some difficulty in the use of FET for these reason, but there are many commercial amplifiers that use FET, though, reparing more than one, I have often found that the bias is controlled by current source with bipolar and zener.
Anyway measure the Idss and VGSoff is quite simple without sophisticated equipment, and to avoid measures loops in change of temperature that you describe, I was trying to know if the axiom subtract 0.63 from VGSoff had a confirmation with a sofware simulation.
As you say the values of spice models are very rough but can help to understand whether a type of component can be right for a purpose.
Some models contain only two parameters related to temperature, Vtotc and Betatce, but are always constant for all devices, 2,5m and 500m, and I dont know this this is real or an approximation.
This is what is available in this model LEVEL, Other LEVEL can have more parameters but are not compatible with free software as LTspice Or Tina-ti.
(my english is very bad, I'm sorry but i hope you can understand me)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.