0

I want to implement an interesting current source topology that I came across recently (picture below). It shows very good performance in simulation and the only thing I am concerned of is its temperature stability. As far as I understand the key is to choose the zener diode with the right voltage, that will cancel out the temperature drift of the bjt vbe. How can it be calculated? Is it dependant of the target current of the circuit(12.5ma in my case)? what parts should be in thermal contact for best performance?

enter image description here

Not sure about the exact purpose of R3, but in simulation adding this resistor significantly improves circuits performance. Its value is in the range or hundreds kohms and must be adjusted for different circuit current. R2 is 1Mohm and according to description is needed for circuit start-up.

Tolik4
  • 121
  • 3
  • What's the purpose of R3? There was a bias circuit popular in the 1980s that used LEDs rather than zeners in that position on the basis of matching tempcos. Low voltage zeners tend to have atrocious regulation - which is less of a problem with that bootstrap configuration. – Neil_UK Aug 02 '23 at 14:45
  • Not sure about the exact purpose of R3, but in simulation adding this resistor significantly improves circuits performance. Its value is in the range or hundreds kohms and must be adjusted for different circuit current. – Tolik4 Aug 02 '23 at 15:09
  • so R3 is a 'shim' – Neil_UK Aug 02 '23 at 15:18
  • When I was working in automotive industry I and our R&D consultant did numerous tests on Zener diodes *(BZX85C and BZV55C series, from 2V7 to 47V, and from two different manufacturers)*. And we found out that 5V1 and 5V6 were the best amongst the entire range/series in terms of Zener voltage variation vs temperature. Personally, I'd go for a TL431 instead as its ref voltage varies by 30 mV only across the whole temperature range (2.495V for normal or 1.250V for LV version). – Rohat Kılıç Aug 02 '23 at 15:29
  • @Tolik4 I've not used the circuit (nor observed it directly.) But there are a lot of temperature-dependent bits. Even copper has a tempco that likely should also be accounted for (as the Wyatt peaking current source takes into account.) Buried zeners (of Robert Bobkin legend) offered the first zeners with 0.3 ppm/K tempco and a noise spec of 7 uVrms. These were 6.95 V devices, back in the day. What analysis have you already attempted? Or are you expecting us to knuckle down and do a thorough sensitivity analysis of all design parameters for *best performance*? – periblepsis Aug 02 '23 at 18:01

1 Answers1

0

Not sure about the exact purpose of R3, but in simulation adding this resistor significantly improves circuits performance. Its value is in the range or hundreds kohms and must be adjusted for different circuit current. R2 is 1Mohm and according to description is needed for circuit start-up.

Not found what is cited here (?).

Here is what I have found ... for T = -25, 0, 25 °C.

My R2 & R3 (same value = ~ 1 kOhm) choose the current.
My R1 & R4 (quasi same value = 5 kOhm) choose the "internal" resistance of the "current" generator ( from ~ 13 V -> 30 V).

enter image description here

enter image description here

Antonio51
  • 11,004
  • 1
  • 7
  • 20