Uncommon low noise current source circuit base on serie / parallel circuit

[Overspeed]

Hello

An interesting circuit built for low noise JFET current source with adjustment by series / parallel

Regards
OS

[Kleinstein]

The JFET constant current circuit was already discussed in a different thread about a JFET based voltage reference.
https://www.eevblog.com/forum/metrology/building-your-own-voltage-reference-the-jvr/msg769983/#msg769983
It works resonably stable if trimmed well. For good precision one still needs to keep the DS voltage well regulated, especially with FETs that have a higher threshold. A downside it that the FETs set the current to a rather fixed current (e.g. some 1.1 mA for 2N4391 as a resonable choice as a voltage ref.) and one can only trim the current a little via the DC voltage. With a much different current there is thermal drift.

[TimFox]

I remember an old Tektronix app note for a curve tracer using a storage CRT display.
With a JFET in the test socket, it was sprayed with a cooling spray and the operator noted which curve of constant V_gs did not shift with temperature.

[Overspeed]

The JFET constant current circuit was already discussed in a different thread about a JFET based voltage reference.
https://www.eevblog.com/forum/metrology/building-your-own-voltage-reference-the-jvr/msg769983/#msg769983
It works resonably stable if trimmed well. For good precision one still needs to keep the DS voltage well regulated, especially with FETs that have a higher threshold. A downside it that the FETs set the current to a rather fixed current (e.g. some 1.1 mA for 2N4391 as a resonable choice as a voltage ref.) and one can only trim the current a little via the DC voltage. With a much different current there is thermal drift.

Hello

Thanks , regarding the '' needs to keep the DS voltage well regulated, '' on the italian design they use a stage between the battery and the JFET ( a pair ).

I am not expert in electronic design ,  I try to learn in the same time I do

Regards
OS

[Kleinstein]

The 2 italian circuits are using quite different ideas:
The first version uses the JFETs as constant current referenes. At a part specific current and with fixed DS voltage the current is reasonable temperature stable (the linear TC crosses zero , but there is a 2nd oder TC), especially in combination with a somewhat stable temperature.

I remember an old Tektronix app note for a curve tracer using a storage CRT display.
With a JFET in the test socket, it was sprayed with a cooling spray and the operator noted which curve of constant V_gs did not shift with temperature.

One gets low TC for  V_gs at about 0.8 V, or with a current to be about 0.8 V from the threshold. So it depends on the current / threshold if one gets a temperature effect.

The 2nd italian circuit is more conventional: the battery as voltage reference and than FETs as souce followers. The FETs choosen are rather large with a treshold where the TC is small and thus work as low noise amplifiers.  The OP-amp part aims for a fixed DS voltatage to suppress the drain voltage effect and also to keep the power at the fets constant and thus allow a stable temperature. With the battery and likely not that great TC this only a short time stable current source, not at all long time stable.

[Overspeed]

The 2 italian circuits are using quite different ideas:
The first version uses the JFETs as constant current referenes. At a part specific current and with fixed DS voltage the current is reasonable temperature stable (the linear TC crosses zero , but there is a 2nd oder TC), especially in combination with a somewhat stable temperature.

The 2nd italian circuit is more conventional: the battery as voltage reference and than FETs as souce followers. The FETs choosen are rather large with a treshold where the TC is small and thus work as low noise amplifiers.  The OP-amp part aims for a fixed DS voltatage to suppress the drain voltage effect and also to keep the power at the fets constant and thus allow a stable temperature. With the battery and likely not that great TC this only a short time stable current source, not at all long time stable.

Hello
Thanks
I do not need long time stable solution as that to make some experiments asap  and allow me to find a '' rock solid '' 10 mA / 20 mA precision current source .

Regards
OS