LM723 PSU with 0V lowest voltage

[Les Jones]

To Les Jones: I am currently experimenting with the ACS712 but I get less than satisfactory results: the best stable resolution I currently get is worse then 100mA. I am experimenting with filter capacitors on the chip, with averaging readings etc. but I am not impressed so far.

I have not use these devices but remembered them as I thought they may be useful for future projects. I will avoid them as a result of your findings. One point. Have you made sure the device is not being effected by external magnetic fields as it works by detecting the small magnetic field produced by current flowing in the conductor. Being close to a transformer could effect it.

Les.

 

[earckens]

To Les: no external magnetic fields present as far as I know but considering I would use them in a PSU close to a transformer that would not do any good either.

 

[spec]

Hi Spec, included is a concept drawing for the current and voltage measurement. Since I hang the controller (Atmega Pro Mini 16M 5V) from the positive side, its ground hangs 5 V lower, yet 44Vdc (rectifier bridge output) - 5Vdc = 39Vdc above the PSU ground. The PSU regulates between 0Vdc and 30Vdc, therefor the voltage feedback to the controller is between -39Vdc + 0Vdc = -39Vdc and -39Vdc + 30Vdc = -9Vdc. So I need to convert voltages between -39Vdc and -9Vdc to 0Vdc and +5Vdc.

For the current measurement I insert the INA219 in the positive supply line. Powering the INA 219 is straight from the controller supply (over the Zener, 5Vdc).

Hy earckens,

Thanks for posting a schematic illustrating you proposed approach for current and voltage monitoring on your PSU.
I don't like to be negative, but here are a few comments.
(1) As the MCU precision 5V supply is referenced to the positive output from the bridge rectifier, the MCU will not actually be monitoring the PSU voltage output.
(2) The position of the current sense shunt resistor will be monitoring not only the current taken by the PSU load but also the current consumed by the PSU circuitry. This will impair current measurement accuracy.
(3) While the precision Zener supply will be adequate for the current amplifier and MCU, which only take a light current, you do not show the power supply for the display, which you imply consumes around 100 mA.

I have been rethinking your original plan of having low side current sensing and while what I say is correct: it will impair the voltage regulation, but it is a mater of degree. On closer inspection I now realise that the shunt resistor need only needs to be a few mili Ohms which will be no more than the resistance of a typical piece of wire.

As a result of this may I suggest that you reconsider a low side current monitor.

spec

(PS have you thought about getting EAGLE ECAD: it's free. )

 

[earckens]

Hi Spec: 1. point taken, so not feasible 2. same 3. total current incl LCD is less then 50mA
Conclusion, back to drawing board and indeed reference to ground. This will make voltage measurement very much simpler.

But current measurement? Can I insert the current shunt for the INA219 in the ground wire? I do not use this chip for voltage measurement (for the time being).

 

[earckens]

(PS have you thought about getting EAGLE ECAD: it's free. )

Not yet, but I will. Except that my wife already thinks I am nuts trying to learn C++, restart with electronics projects after gazzilions of years in the freezer, learning how to maintain a 54 year old Triumph TR4, and our building projects in Southern France

 

[spec]

Not yet, but I will. Except that my wife already thinks I am nuts trying to learn C++, restart with electronics projects after gazzilions of years in the freezer, learning how to maintain a 54 year old Triumph TR4, and our building projects in Southern France

Join the club. There are many people on ETO who love a challenge. You will find EAGLE relatively simple, compared to learning C++, although I only use EAGLE for circuit capture.

spec

 

[spec]

Hi Spec: 1. point taken, so not feasible 2. same 3. total current incl LCD is less then 50mA
Conclusion, back to drawing board and indeed reference to ground. This will make voltage measurement very much simpler.

But current measurement? Can I insert the current shunt for the INA219 in the ground wire? I do not use this chip for voltage measurement (for the time being).

Hy earckens,

In view of the very low value of shunt resistor required for the INA219, the shunt resistor should introduce an insignificant error in the voltage stabilization of the PSU so, yes, I would suggest that placing the shunt resistor in the 0V line of the PSU should be OK. This means that you would only be measuring the current taken by the load on the PSU, so current measurement accuracy will be high, and, as you say, 0V current sensing will simplify voltage measurement.

Sorry to say this, but I still think that even with only 50mA consumption for the display, a 5V shunt supply provided by an LM4040 5V precision Zener diode will be problematic because of the low current capability (16uA to 15mA) of the LM4040 5V. https://www.ti.com/lit/ds/symlink/lm4040-n.pdf. The problem is that you cannot guarantee that the display will take a constant 50mA.

You could have a precision 5V supply provided by an LM4040 5V for the current amp and the MCU and a separate 5V supply for the display.

spec

 

[earckens]

That is a good idea! Thank you.

 

[spec]

Hy earckens,

Here is the updated LM723 PSU schematic, now at version 3, as promised. I hope it it correctly incorporates the amendments that you outlined; perhaps you would like to check it.

spec

NOTES
(1) All non electrolytic decoupling capacitors are thru hole (not surface mount) ceramic, X7R dielectric, +- 10% or better.
(2) All frequency compensation capacitors are metal film type, +- 5% or better
(3) RV1 can be a wire would multi turn potentiometer to give fine adjustment
(4) RV3 can be a wire wound multi turn potentiometer to give fine adjustment

 

[earckens]

Chuck, this is great! Thank you very much.

Have a good evening,
Erik

 

[spec]

Chuck, this is great! Thank you very much.

Have a good evening,
Erik

And you too.

Chuck

 

[earckens]

Chuck, yesterday I downloaded and installed the Eagle edu 7.6 version of the software. It looks impressive, but also with a steep learning curve.
Question: I have some .brd file that needs conversion to gerber files; I looked it up on internet and did the conversion, however I am not sure I did it right: would you mind having a quick look at my files please? I will pm them. Thank you!
Erik

 

[spec]

Chuck, yesterday I downloaded and installed the Eagle edu 7.6 version of the software. It looks impressive, but also with a steep learning curve.
Question: I have some .brd file that needs conversion to gerber files; I looked it up on internet and did the conversion, however I am not sure I did it right: would you mind having a quick look at my files please? I will pm them. Thank you!
Erik

Hi Erik,

EAGLE 7.6 is easy to get the hang of. Although it is not in the same league as Altium, EAGLE is simple, solid, and friendly and there is a wide user base (EAGLE is also free ).

The first thing to do is to set the grid to 0.05 inches (don't try using metric: it causes all sorts of problems).

The next thing to do is to select the whole component library and 'Use' it. If you do not 'Use' the library you will have no components to use and that is what throws most first time users.

After that you will soon be able to do a schematic.

No it is not you: the standard EAGLE component library is a complete shambles.

Afraid I have only used EAGLE for circuit capture and know nothing about Gerber files- sorry.

But there are some EAGLE PCB experts on ETO: perhaps start a new thread.

Chuck

 

[earckens]

Hi Chuck,
I did check the conversion with Gerberview, and then a double check again with a conversion done in Kicad (yes, I downloaded that one to in the meantime) -steep learning curve, I can tell you!- and found no errors. So that is out of the way.
Can you maybe send the Eagle files you used to draw the circuit diagram for the LM723? So I can start making a pcb..
Thank you!
Erik

 

[spec]

Hi Chuck,
I did check the conversion with Gerberview, and then a double check again with a conversion done in Kicad (yes, I downloaded that one to in the meantime) -steep learning curve, I can tell you!- and found no errors. So that is out of the way.
Can you maybe send the Eagle files you used to draw the circuit diagram for the LM723? So I can start making a pcb..
Thank you!
Erik

Yeah sure, but bear in mind that I have only made a schematic and you will need to check the packages and net lists etc before attempting a PCB layout.

Update: see your messages inbox

spec

 

[earckens]

I need to come back on this subject: after several weeks I finished a prototype for this PSU but the problem I have is that the BC181 (Q1) does not appear to start conducting (no output voltage on PSU).

On power-up -and after settling- I get following voltages:

1. bridge rectifier output: 22.4Vdc (I reduced the transformer output for testing purposes; when applying full power -ie. 36Vdc I get similar results)
2. -comp (pin 4): between 0V and 0.87V (when pot B -voltage set- is turned to full right)
3. +comp (pin 5): 5923V
4. Vref (pin 6): 7.21V
5. Vc (pin 11): 21.7V

So there is about 0.6V Vbe on Q1, it should conduct, right? So the emitter voltage should continue to the collector, and hence be applied to Q2 and Q3. But Vcollector of Q1 measures 0V.

What is wrong in this circuit, and what is wrong in my reasoning?

PS: I am sorry spec, I had not yet had time to get this drawing done in Eagle.

Edit 08/12/2016: attached drawing corrected: -comp and +comp on pins 4 and 5 were swapped

 

[spec]

Hi earckens,

Nice to see you back on ETO,

I haven't analyzed your problem yet: what actual transistor do you have fitted to to the power supply PCB where the BC181 is shown?

spec

 

[earckens]

spec, BC181 is PNP: https://alltransistors.com/transistor.php?transistor=22448

At the moment I have a 2N2905 fitted (same as 2N2907 but with bigger can).

 

[earckens]

The BC181 has a hfe of min 60 while the 2N2905 has a hfe of 300. I do not know whether this DC current gain has an influence, but I think that Q1 should act as an amplifier and that it should operate in the amplification part of its curve, not the switching part? Or am I confusing things?

 

[spec]

spec, BC181 is PNP: https://alltransistors.com/transistor.php?transistor=22448

At the moment I have a 2N2905 fitted (same as 2N2907 but with bigger can).

Whoops- I just woke up.

I have revised post #77.

spec