| Electronics > Projects, Designs, and Technical Stuff |
| bench/lab psu design |
| (1/2) > >> |
| m3vuv:
lab type psu,Hi folks can you give me your thoughts on the attached schematic,cheers m3vuv |
| not1xor1:
--- Quote from: m3vuv on December 30, 2018, 05:26:49 am ---lab type psu,Hi folks can you give me your thoughts on the attached schematic,cheers m3vuv --- End quote --- you should first buy me a new pair of glasses ... ;D can't you post a larger picture? |
| David Hess:
There is a more legible copy of the schematic and a description here. The concept is good. The current control error amplifier clamps the output of the voltage control amplifier which should provide good performance. Unfortunately it does this through a common emitter transistor which is not good for stability but if it works, then it works. Adding a low resistance between the CS pin of U2 and common to stabilize the transistor's gain would help with this. Better would be to reverse the control of the current sense amplifier and clamp through a an emitter follower which has fixed unity gain. There is no output capacitance which would normally be required for stability except for RF filtering so I wonder about this. |
| Kleinstein:
The floating type regulator usually needs the output capacitor for stability. The compensation for the voltage loop also looks simpler than normal, which would suggest it should be slow to be stable at all - thus is would even need a rather large capacitor at the output. The current regulation could suffer from low precision: one point is that the base current also flows through the sense resistor. The other is that the 723 is not very low drift for the low low voltage at the shunt. The circuit looks like an old crude design idea (but not fully worked out) from the 1970s. I would not start from there. |
| David Hess:
--- Quote from: Kleinstein on December 31, 2018, 11:40:52 am ---The floating type regulator usually needs the output capacitor for stability. The compensation for the voltage loop also looks simpler than normal, which would suggest it should be slow to be stable at all - thus is would even need a rather large capacitor at the output. --- End quote --- Most power supplies either require output capacitance for stability or perform better with it. I suspect the designer either got lucky or left a 100uF output capacitor off of the schematic. I do not suggest doing this design unless one is prepared to do some transient response testing and tuning which just takes a function generator and oscilloscope. --- Quote ---The current regulation could suffer from low precision: one point is that the base current also flows through the sense resistor. The other is that the 723 is not very low drift for the low low voltage at the shunt. --- End quote --- I disagree insofar as the 723 differential input stage is a monolithic differential pair. Drift should be better than a 741 or 324 which have two Vbe junctions in series so say 10uV/C or 20uA/C with a 1 ohm sense resistor. That is several times better than the current shunt resistor just by itself so the 723 is not limiting performance. An input bias current of 1 microamp (from the 723 schematic) could yield an input offset current drift of 10nA/C or 30uA/C. I do wish that a jelly bean precision operational transconductance amplifier existed for these sorts of applications. --- Quote ---The circuit looks like an old crude design idea (but not fully worked out) from the 1970s. I would not start from there. --- End quote --- It looks like something copied from a much higher voltage design. With only a 20 volt maximum output, the floating configuration is not required. |
| Navigation |
| Message Index |
| Next page |