Hi all.
I vould like to know if anyone has a particular favorite transistor for an application i am considering.
I am doing current mirrors and have the intuitive idea that Vbe vs IC linearity is the most critical part, if it is linear in a region i can adjust the emitter resistors to acheive more gain and thus adjust for base currents, which also look more like a resistive load when the Vbe Ic caracteristic is linear.
The basic requirements are:
Not impossible to get a hold of.
PNP devices primarely.
IC linear region +-1.2mA
Max current small signal transistor <200mA
High beta (if possible)
Spice models available.
Are the 2SA1455k (ROHM i think) any good and are there any spice models of them.
I vould like to know if anyone has a particular favorite transistor for an application i am considering.
I am doing current mirrors and have the intuitive idea that Vbe vs IC linearity is the most critical part, if it is linear in a region i can adjust the emitter resistors to acheive more gain and thus adjust for base currents, which also look more like a resistive load when the Vbe Ic caracteristic is linear.
The basic requirements are:
Not impossible to get a hold of.
PNP devices primarely.
IC linear region +-1.2mA
Max current small signal transistor <200mA
High beta (if possible)
Spice models available.
Are the 2SA1455k (ROHM i think) any good and are there any spice models of them.
I think you mean logarithmic Vbe vs Ic, right?
The "ideal" transistor's collector current is given by
Ic = Is*e^(Vbe/VT)
where VT = kT/q, the thermal voltage, about 26 mV at room temp, and Is a scaling constant that depends on the emitter area.
Anyway, even in an ideal transistor, this is pretty far from linear.
On the other hand, in the ideal transistor,
Gm = d(Ic)/d(Vbe) = Ic/VT
Which is a linear relationship.
Or maybe you mean Ib vs Ic linearity, or constancy of beta vs Ic?
In that case, I'd probably just look for a high beta transsitor to minimize the beta nonlinearity, or use a beta-immune current mirror topology.
Or maybe the question is, how does one make the most linear current mirror. Which is probably open to all sorts of debate.
-- mirlo
The "ideal" transistor's collector current is given by
Ic = Is*e^(Vbe/VT)
where VT = kT/q, the thermal voltage, about 26 mV at room temp, and Is a scaling constant that depends on the emitter area.
Anyway, even in an ideal transistor, this is pretty far from linear.
On the other hand, in the ideal transistor,
Gm = d(Ic)/d(Vbe) = Ic/VT
Which is a linear relationship.
Or maybe you mean Ib vs Ic linearity, or constancy of beta vs Ic?
In that case, I'd probably just look for a high beta transsitor to minimize the beta nonlinearity, or use a beta-immune current mirror topology.
Or maybe the question is, how does one make the most linear current mirror. Which is probably open to all sorts of debate.
-- mirlo
Thanks,
I was trying to avoid the physics of the transistor but it comes up and bites you whenever you try.
The thing i am searching for is as you say the gm.
The gmneeds to be constant or as slow moving as possible over at least the +-1.2 mA delta ic.
I suppose linear current mirrors is the thing i am searching for but my gut feeling tells me that the requirements on the transistors differ depending on type of current mirror you use.
I was trying to avoid the physics of the transistor but it comes up and bites you whenever you try.
The thing i am searching for is as you say the gm.
The gmneeds to be constant or as slow moving as possible over at least the +-1.2 mA delta ic.
I suppose linear current mirrors is the thing i am searching for but my gut feeling tells me that the requirements on the transistors differ depending on type of current mirror you use.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.