Tony Williams wrote...
A programmed hard-clamp of the voltage into the integrator
resistor is an alternative way. Not so elegant as the OTA,
but possibly good enough for your application.
R1 R2 R3 R4
Vin +-/\/\-+-/\/\--------+--/\/\--+--/\/\---+-->Vout
| | _ | |
| | / -|-+ |
| +--<OP1| |
| \_+|--+0v |
| |
| (+Vc-Vd) |
| _|_ |
| _ D1/_\ |
+--|- \ R5 | R6 C1 |
|OP2>--/\/\-+-/\/\--+---||----+
0v+--|+_/ _|_ | _ |
D2/_\ +--|- \ |
| |OP3>--+
(-Vc+Vd) 0v+-|+_/
(+Vc-Vd)/|\ /|\(-Vc+Vd)
| |
R7 | | R8
-15v -/\/\--+ +--/\/\- +15v
| |
Vc +----|>|---+-/\/\--+---/\/\---+--|>|--+
D3 R9 | R10 D4 |
| _ |
+--|- \ |
|OP4>-----------+ -Vc
0v+--|+_/
Vc= slewing rate control voltage, 0-10v.
Vout slewing rate = Vc/R6*C1, Volts/Sec.
R1=R2, R3=R4, R9=R10, say 10k each.
R5=R7=R8, value not less than 15v/5mA.
Haven't decided on the ratio of R5:R6 yet.
The problem with this approach is the diodes, of course, and
their lack of a sharp cutoff. Whereas the transconductance
amplifier circuit could be programmed accurately over nearly
a four-decade range, the diodes would restrict one to about
two decades, or even less depending on the accuracy spec.
This issue can be solved by adding more parts, in this case
a set of nearly-perfect active rectifiers.
.. R1 R2 R3 R4
.. Vin o-/\/\-+-/\/\--------+--/\/\--+--/\/\---+--> Vout
.. | | __ | |
.. | | / -|-' |
.. | '--< | |
.. | \_+|-- 0v |
.. | __ |
.. '--|- \ R5 R6 C1 |
.. | >--/\/\-+--/\/\-+---||----+
.. 0v --|+_/ | | __ |
.. | '--|- \ |
.. Vclamp __ | | >--'
.. o-+-------|+ \ D1 | 0v -|+_/
.. | | >--+--|<|--+ __
.. | ,--|-_/ | +-----|+ \
.. | | D2 | | | >--, Vnode
.. | +---|<|---' R7 | ,--|-_/ |
.. | '--------/\/\---- |--+---------+
.. | __ | |
.. | ,----|+ \ D3 | |
.. | | | >--+--|>|--' |
.. | | ,--|-_/ | |
.. | | | D4 | |
.. | | +---|>|---' R8 |
.. | | '--------/\/\------------------'
.. | '----------------,
.. | R9 R10 |
.. '---/\/\--+--/\/\---+
.. | __ |
.. '--|- \ |
.. | >--' -Vclamp
.. 0v --|+_/
I warned notstop it was a lot more complex this way.
