LM723 PSU with 0V lowest voltage

[spec]

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

No, while the DC current gain is important, a wide range of current gains will be suitable for your power supply. The 2N2905 is an excellent transistor, but unfortunately it is obsolete now.

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?

Afraid that is not right.

spec

 

[earckens]

So Q1 with 2N2905 or 2N2907 should be ok.. Then what is wrong with the circuit, Q1 does not want to start conducting, Vcollector remains 0V?

 

[spec]

So Q1 with 2N2905 or 2N2907 should be ok..

Yes, both transistors will work fine, but note my previous comment, that you will need a more powerful transistor on a heatsink when you start taking any significant current from the PSU.

Then what is wrong with the circuit, Q1 does not want to start conducting, Vcollector remains 0V?

I will have a look (domestic things going on at the moment).

spec

 

[spec]

Pin 8 of the LM723, should be left open circuit, not connected to 0V.

spec

 

[earckens]

Hi spec, I have a TIP42 with heatsink ready for use. Currently the 2N2905 is fitted with a heatsink.

 

[earckens]

Pin 8 of the LM723, should be left open circuit, not connected to 0V.

spec

Ok, I check this (will be this afternoon, I am on the road now) and report right back!
Thks, Erik

 

[spec]

Hi spec, I have a TIP42 with heatsink ready for use. Currently the 2N2905 is fitted with a heatsink.

That will be fine.

Ok, I check this (will be this afternoon, I am on the road now) and report right back!
Thks, Erik

No probs. There is a chance that there is an internal connection to pin 8. If that is the case connecting pin 8 to 0V may be upsetting the LM723. Unfortunately, there is also a chance that the LM723 will be blown.

spec

 

[Les Jones]

Hi spec,
If the recents posts are referring to the schematic in post #76 then I think there is an error in it. Pin 10 (VO) (The emitter of the internal pass transistor.) should be connected to a more negative point so the collector of the internal pass transistor (pin 11) can pull the base of Q1 negative to turn it on.

Les.

 

[spec]

Hi spec,
If the recents posts are referring to the schematic in post #76 then I think there is an error in it. Pin 10 (VO) (The emitter of the internal pass transistor.) should be connected to a more negative point so the collector of the internal pass transistor (pin 11) can pull the base of Q1 negative to turn it on.

Les.

Hi Les,

Thanks for the observation about the circuit of post #76.

​

Above is the LM723 functional diagram from post #9. A 56K resistor connects between 39V and pin 10 of the LM723 to bias the 6.2v Zener in the LM723 on. Vz (pin 9) is then connected to 0V. This is done because the base of the series pass transistor in the LM723 will not operate at 0.6V above 0V.

The collector of the series pass transistor is then connected to the base of the PNP output transistor driver.

So all is well (in that area), I think.

spec

 

[Les Jones]

Hi spec,
In the schematic in post #76 pin 9 is not connected. In the schematic in post #69 it is connected to zero volts.

Les.

 

[earckens]

Hi spec,
In the schematic in post #76 pin 9 is not connected. In the schematic in post #69 it is connected to zero volts.

Les.

Hi Les, pin 9 (Vzener) should not be connected. The original PSU I had made long time ago pin 9 was not connected, so I would keep it that way.

 

[earckens]

Spec, but I have no explanation why Q1 does not conduct even though Vbe is O,6V approx.?

 

[spec]

Hi spec,
In the schematic in post #76 pin 9 is not connected. In the schematic in post #69 it is connected to zero volts.

Les.

You are right Les. I am not sure how I managed it, but I was looking at the schematic of post #69 instead of the schematic of post #76. Anyway, I have rebooted and cleared the cache on my laptop and can now see the correct schematic that you refer to.

Spec, but I have no explanation why Q1 does not conduct even though Vbe is O,6V approx.?

earckens, you need to connect Vz (pin 9) to 0V, as Les says.

spec

 

[earckens]

You are right Les. I am not sure how I managed it, but I was looking at the schematic of post #69 instead of the schematic of post #76. Anyway, I have rebooted and cleared the cache on my laptop and can now see the correct schematic that you refer to.


earckens, you need to connect Vz (pin 9) to 0V, as Les says.

spec

Hi spec and Les, I am sorry for the confusion caused between posts 69 and 76. Currently it is the schematic of post 76 which is used in the prototype, and which also was in use in my first PSU made 30 years ago.

But ok, I will 1. disconnect pin 8 and 2. connect pin 9 (Vzener) and 3. buy a new LM723 for the prototype

This afternoon this will be executed and reported back to you.

Thks!
Erik

 

[spec]

Hi spec and Les, I am sorry for the confusion caused between posts 69 and 76. Currently it is the schematic of post 76 which is used in the prototype, and which also was in use in my first PSU made 30 years ago.

But ok, I will 1. disconnect pin 8 and 2. connect pin 9 (Vzener) and 3. buy a new LM723 for the prototype

This afternoon this will be executed and reported back to you.

Thks!
Erik

Hi earckens,

You do not need to disconnect pin 2.

Just make sure pin 8 is not connected and connect pin 9 (Vz) to 0V.

spec

 

[earckens]

Hi earckens,

You do not need to disconnect pin 2.

Just make sure pin 8 is not connected and connect pin 9 (Vz) to 0V.

spec

Hi spec, I will edit with more clarity in the future:
I wrote:
1. disconnect pin 8
2. connect pin 9
...etc...


Erik

 

[spec]

Hi spec, I will edit with more clarity in the future:
I wrote:
1. disconnect pin 8
2. connect pin 9
...etc

sorry Erik. It is not your fault. It is me. I need to reboot my brain.

spec

 

[earckens]

Current situation: pin 9 (Vzener) to ground, pin 8 disconnect (NC), a new LM723 inserted.

On start-up I could see the output voltage very briefly jump upward (voltage pot set halfway) and 0V

Measured:

-comp (pin 4): between 0.09V and 0.82V, depending on voltage pot setting (turn full right = 0.82V)
+comp (pin 5): 5.9V
Vref (pin 6): 7.2V
Vout (pin 10): 6.59V
Vc (pin 11) 21.7V
V+ (pin 12, supply): 22.34V
Vbase Q1: 21.7V
Vcollector Q1: 0.52V
Vemitter Q1: 22.4V

Is Vout (pin 10) at 6.59V normal (edit: yes it is!)?

Vbe Q1 is 0.7V, yet Vcollector still is 0.5V (edit: still not conducting, what could be going on)?

Erik

 

[Les Jones]

The voltage across the current sense resistor (0.12 ohm 5W)is not being used. I think the bottom of the current limit pot should go to the input end of that resistor.
Eric, Can you repeat the test with the current limit pot set to the 0 volts end of the track. I think the current limit may be comming into play to drop the output voltage.

Les.

 

[earckens]

Hi Les, I think you hit a major snag! I just checked my initial notes from when I reverse engineered the original pcb and I notice that I did overlook the different ground potentials to be used!

From my notes I think to deduce that following items should be connected to the "input" of the 0.12 ohm resistor:

CS (pin 3)
the 2.2nF capacitor on the CS side

I am not able to deduce whether the grounds for the current sense pot have to be connected before or after the 0.12 ohm resistor; same for voltage pot ground (although I think to remember they were connected at the output of this resistor).

Now what I will do, based on your comment, is to first connect CS and 2.2nF to the input of 0.12ohm. Subsequently also connect the pot grounds there.

I will report back tomorrow (my missus wants to spend evening time together )

Thanks a lot, I think you may have hit the nail.

Erik