Constant Current Test Circuit

[Anonymous]

Here is the circuit I used for my experiments. The screen display is quite good, however, for printing you should download a high resolution PDF from here: www.tb-electronic.de/pi_tech/pulse_cc_circuit.pdf
Everybody who is interested in PI basics should build this circuit. I don

 

[Anonymous]

Hi Thomas
Thanks for you good description on this circuit and I believe you when you say its difficult to get to work the right way.
The failure in the schematic is that you turn on the circuit by pulsing the plus supply for the opamp. This is a very bad way to do this because you switching on and off the supply to the circuit that should do the regulation.
How should it work proberly? it takes time for it to turn on and when you "pulse +" signal goes low it loses its supply, and its capability to switch off the Q1 the right way.
There is several other ways to do this much more effient, first thing to do is connecting +12 volt permanent to pin 7 of the opamp. Then the simplest solution as i see it is to place a NPN transistor like BC547 across R4, Collector to input 3 of the opamp and emitter to ground side.
Then connect a basis resistor of 10K to the transistor and use this as you pulse input. What you now have is a current on/off switch that rise from zero to CC adjust settings. The on/off of the Q1 will be very controlled and I guess nearly every opamp will work very well. Maybe some experiment with the C1 capacitor, will be needed for risetime.
Thanks a lot
Mark

 

[Anonymous]

Hi Mark,
thanks for your response.
The little trick with switching the OP supply is intended and causes no problems. It just simplifies the design. When the opamp "wakes up", the coil current is zero and starts to rise. The mentioned instabilities arise during the crossover from the linear current ramp to the constant current phase, and this is not directly related to the switching of the power supply. But due to this "switch mode" we can not add a decoupling cap directly across pin 4 and 7 -- maybe this causes problems. I will check this tomorrow. Anyway, the "ringing" I mentioned is not really a problem.
The turn off of the IRF640 is fast enough even with the positive rail missing.
We already used "switched opamps" in other designs, for example in our low power PI design: the opamp is only turned on while needed, i.e. during the receive phase. The rest of the time it's not needed and energy can be saved. A real disable input would be better, of course, but most opamps don't have one.
Thanks again for your input!
Thomas

 

[Anonymous]

Hi Thomas
Thanks for the explanation, it just seemed a little strange to me, I always design with a stabile power supply as a major issue. I am also sure that to make this circuit work stabile you most have a well defined rise and fall time
for the pulse you use as power on/off or else it will change everything.
Thanks for the design idea, I like it and will build it to see how it react.
Mark

 

[Anonymous]

Hi Mark,
just a few remarks for you and anyone else who likes to build the curcuit:
- I tested the behaviour with a continuous and decoupled supply for the CA3130. No difference.
- The drive pulse is not that critical. In fact, on my test board it is driven directly from a 4538 cmos mono.
- I probably did test all this three years ago, just don't remember all details. The OP is very critical, I did not find any replacement for the CA3130 (note that most other OPs require an additional negative power supply, whereas the CA3130 common-mode input-voltage range includes the negative supply rail).
One point I forgot in my first post concerning the energy balance:
If a compensated coil is used and the constant current is really smooth so that there is no noise in the RX coil during this period (I doubt that this is possible), then we could take two receive samples per pulse: the usual one after the flyback pulse plus another one after the TX pulse, i.e. during the CC phase. This would improve the efficiency of the 400 volts version.
Thomas

 

[Anonymous]

I have just updated the CC circuit. For details see my post in JC's thread of Jan 18, 2002.
Thomas