And add resistor R6 from the collector to the output of the operational amplifier. R6/R2 then sets the gain of the local feedback loop and R3/R4 controls the overall gain. The advantage of this configuration is higher operational amplifier performance and simplicity although understanding it is more complex.
A dual operational amplifier is not suitable for this because the output comes from the supply lead but the other amplifier in the package could be disabled.
Figure 2 on page 2 shows the general idea. It also shows how to add simple current limiting with another transistor:
http://cds.linear.com/docs/en/application-note/an18f.pdf
i tested NPN/PNP combo setup earlier, its oscillated like bad even using gain control Suggested by Hero (see first attached) picture. and then i made another sketching... and tested it beautifully (see 2nd and 3rd picture below), it turned out to be similar NPN/NPN combo that Hero suggested earlier, except its using TL431 as reference, my setup is from an opamp fed by resistive divider with potentiometer to control Vout...
but then, i found out, it has not great impedance when it goes negative voltage (see 2nd picture below), i'm thinking i will use the Vout into the LM317 ADJ pin and looking at LM317 schematics, ADJ pin needs to withdraw current, the above setup cant sink/drain/suck current satisfactory, so i spent the whole evening to simulate push pull type. below is what i find out to be the best, but i need comment, in case i miss some resistors that will blow up this circuit.
please bear in mind the objective is to control LM317 on high voltage rail (60V) from low voltage opamp rail ±15V. opamp output (10V) will be multiplied by 5 to get 50V Vout and fed to LM317 ADJ pin. high voltage opamp is not available. and please note, the below circuit is using mixed rail setup, where the push pull collector at the bottom is fed to -15V rail, because i want LM317 to output to 0V, so i hope this is safe. dont worry about LM317 input, it will not exceed 37V (Vin - Vout) because we have tracking pre regulator beforehand. if no comment i will assume and hope this wont smoke the low voltage components when i assemble them on the breadboard later, i killed 1 uA741 already because i shorted something on the breadboard, i guess the uA741 got fed with 60V and now its output is broken
any comments? thanks.
yay i've succesfully drive LM317 from 60V input to output 40 - 50V, using the earlier push "no pull" LV to HV amplifier. no smoking LM317. not that bad regulation, but that just me the suck arse guy... maybe soon i'll try the push pull amplifier, i hope better performance...
Be careful that a short to the output does not allow the maximum input to output voltage difference on the regulator to be exceeded.