How to eliminate early effect in constant current source

When designing an differential amplifier it is important to have a precise constant current source (which I will abbreviate with CCS). It helps to keep out signal injections from the supply rails. A precise CCS has the main property of firstly the current (of course) and secondly the impedance, which ideally is infinite. The thing about this is now that the impedance strongly depends on the Early voltage, which unfortunately is highly unknown for various transistor types and not mentioned in datasheets (nor is it easily derived from given specs). The impedance can be improved by adding a emitter resistor to the BJT but it still depends on Early voltage (Vaf). Is there a way to circumvent this dependency of the CCS from Vaf? We're talking about a very simple kind of current source as described here (first picture).

Furthermore, when looking into the spice model of the MPSA42/92 you find Vaf of 10V which looks way to little. Is that a true value? If it is, the MPSA would be quite useless as a transistor, since it would in essence resemble a normal resistor. See here. Therefore I believe the value in the model is just wrong.

How come that a very vital property of a transistor is so hard to find and so neglected? Furthermore Vaf is important for many parameters of a circuit, e.g. the gain of a transistor stage. Am I on a wrong track here and Vaf is not as important as I believe, or can it easily be ruled out (as is oftentimes possible for beta).

To make long story short: Is there a way to calculate the impedance of a CCS without knowing Vaf?

EDIT: As jippie propsed, I put my questions in a clearer manner here:

1. Is there a way to circumvent this dependency of the CCS from Vaf?

2. [...]10V which looks way to little. Is that a true value?

3. How come that a very vital property of a transistor is so hard to find and so neglected?

4. Is there a way to calculate the impedance of a CCS without knowing Vaf?

• Look for 'cascode current source' and/or 'Wilson current source' Eg.: ami.ac.uk/courses/ami4409_amsicd/u01 Aug 16, 2014 at 6:32
• Yes, I thought of that too, probably is this the way to go. But it doesn't explain if Early effect is an underestimated phenomenon or not. My questions remain the same. Aug 16, 2014 at 21:33
• Actually I think you are asking at least 3 and maybe even 4 different questions here, that is (in my opinion) why you are getting so little response. It is better to focus on a single question, perhaps open a separate one for the others. 1. Is there a way to circumvent this dependency of the CCS from Vaf? 2. [...]10V which looks way to little. Is that a true value? 3. How come that a very vital property of a transistor is so hard to find and so neglected? 4. Is there a way to calculate the impedance of a CCS without knowing Vaf? Aug 17, 2014 at 6:06
• Okay, the first two SPICE model parameters for an MPS42 I looked at had Vaf = 100V, so 10V seems too low. ro = Vaf/Ic, so a simple 0.1mA current source would have an equivalent output resistance of about 1M$\Omega$, not too terrible. Aug 17, 2014 at 17:13
• @SpehroPefhany Thanks as I searched again, and I found a link that said (centralsemi.com/docs/CSM/MPSA42.LIB) 9,9209V which does not look like a ballpark estimate. But I also found the value you said in another model: network.wenzel.com/misc/spice.htm Aug 17, 2014 at 23:27

1 Answer

1. The best way to circumvent the dependance on VA is through cascode mirroring. These have the effect of somewhat pinning the collector voltage on the output transistor so that most of the dependence on VA is removed because Vce won't be changing much on the bottom two transistors in the mirror.
2. It's an awfully low value for an early voltage, but likely not a problem for uses like a switch or a simple amplifier. The reason the early voltage is so low is because that transistor is designed for 200-300 Volts. Designing transistors to handle such extreme conditions requires design trade-offs. In this case, the gain and early voltage aren't that great.
3. Unless you're designing an integrated circuit most designers don't really have a use/need for it. You generally need matched transistors to perform what you're designing and discrete parts are hardly ever matched so adding in VA isn't going to help anything. This seems more like an issue of the right tool for the right job.
4. The impedance of the output side of a current mirror is completely dependent upon the Early Voltage as you can see here: http://users.ece.gatech.edu/phasler/Courses/ECE4430/Unit2/Current_Mirrors.pdf so there's no way around determining the output resistance if you don't know the early voltage. If you're using a simulator, you can plot the Vce vs Ic graphs and then determine the early voltage from a couple regression lines as seen here: http://en.wikipedia.org/wiki/Early_effect

In summary, I'd use different transistors to build a differential amp and go with a cascode current mirror.

• Thanks for the answers. The current mirror paper is something I have to dig in (though I hate this non-latex style... eye-cancer...). Relating to 4, I'm reading a book from Douglas Self on amplifier design (which contains way to little math in my opinion) where he describes one of the current sources with emitter degeneration. He calculates it to have 4.6Mohm but never states Vaf of the used device which made the calculation for me rather useless. So following your answer I'm probably right, that it is incomprehensible. Thanks, that helped a lot, but I may come up with more on that topic Aug 18, 2014 at 5:28