Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Help needed, Aluminium anodising power supply revisited is dead!

Status
Not open for further replies.

ceefna

New Member
Hi all, I have a serious problem with the power supply. It has worked perfectly since march but now it is setting fire to the LM723's

I am on my forth now and that has just died. I have replaced all 4 transistors to see if it fixed the problem. When I switched back on the current adjust was working but it would not go over 3V before it set fire to the 723 again. I have tested the other components as best I can with my limited knowledge. The LM723's are still burning out without the supply connected to anything so it's not an external fault.
I run the circuit from a 12.5 amp 24v ac transformer through a bridge rectifier. I have got 33.5V at pin 11 and 12 and earth is ok( no shorts ) The 723's burn within a few seconds so its hard to get other voltage readings. The last one that flames shot out of melted its self to the dip socket on pins 11 and 12 but I don't know wether this was the chip internally shorting out.

Please can you give me some guidence as to how to start testing/ what voltages at pins etc.
I am really stuck with this problem and am running out of 723's!!!!
I have attached the circuit I am using it has worked fine for 8 months so I think the circuit is ok.

Please let me know what other info you need.
Thanks for your time.

Ceefna
 
The first thing I would do is replace the 5k6, 2k2 and 1k2 resistors with metal film types (existing are mostlikely carbon composite or film) also replace the 1000uF/40V
Than remove and test the 2 pots under load, rotating them to make sure there are no discontinuities
Than I would remove the 3 pcs. 3N3055 and replace the 390/5W with a 1k/1W and connect the emitter of the BD243 through a 47 Ohm resistor to the output (same as pin 3 of the IC)
Reassemble without the 2N3055, set output to lowest voltage and pray
Slowly increse the voltage (wait inbetween 10% steps for 10 minutes), no load on output, allways check for heat on the IC
Once alls OK add a heatsink to the LM723 and reinstall the 390/5W and use new 2N3055, repeat previous line (and the pray part!). E
 
Last edited:
@ceefna,
I suspect some dry joint around the 680 pF capacitor across &23 pins 2 and 13.
I had troubles when this cap developed a dry connection.

Later, depending on maximum supply voltage output needed, manage to reduce the input. I t appears , you are working at the maximum possible input to the chip !!!
 
According to my DataBook, 40 Volts is the ABSOLUTE MAXIMUM Input.
So your Really Pushing it Hard.

I would recommend you put a 8 Volt Zener in Series with Pins 10 and 11, Plus a 10uF Bypass Cap from those pins to Ground.
Now your 623 will be operating at about 32 Volts, but your Transistors should still function OK at the 40 Volts.
 
Hi guys thanks for the replies. I am actually running it at 33.5V which I know is at the top end of the 723 but equates to 83% capacity. Plus it has been working fine for 8 months at that voltage so I would be surprised if that was the problem.
I will fit a new 680pf cap so that will rule out the dry joint.
I have fitted 2 new 1K/20K 10 turn pots but not sure if they are wired correctly. Should the resistance at pins 2 and 4 increase with higher voltage/current or decrease?

Thanks
Ceefna
 
Last edited:
I feel that we need to analyse, concentrate on, how , a working power supply gets frequently damaged killing the control chip.
he could go for a new design , is a different aspect.
 
It's adjustable voltage and current? Current adjustment is what is important to me and as the last post said it was working so what has changed?
Really don't want to use a different circuit as this one anodised great
Please check my potentiometer wiring
Current adjust- pin2 to pot2(wiper) pin10 to pot1 pin3 to pot3
Voltage adjust- pin4 to pot2(wiper) pin3 to pot1 then pot2 to pot3 through resistor to ground.

Ceefna
 
Last edited:
higher the voltage at pin4 of chip(inverting input) lower the output. It means that, to increase the output you have to lower the voltage at Pin4. pin 5 the the non inverting input is referred to a fixed reference of around 2.5V. Thus you have to wire the pot in a way that as you turn clockwise, the voltage at pin4 should reduce.
 
Last edited:
There's no current feedback so it's not what I would call a "constant current" supply. At best you could say it has a crude current limiting.

Anyway if you are happy with that then it's a matter of working out the failure mode, and adding protection.

For a start if pins 11 and 12 are the power pins to the IC there is no decoupling from the main PSU. You could add a resistor between PSU+ and the IC power pins, then a cap(s) from those pins to IC gnd.

Secondly there is no decoupling of pin3 from any spikes etc that occur on the main power output. I'm not that familiar with the 723 but I expect it might not like a high energy spark of +/- X volts on it's pin 3 which is probably a sensitive opamp input pin.

Seriously if it was my anodising supply I'd dump the 723, keep the 4 power transistors and driver and use something like an LM317 as the constant current controller to drive them. I would also keep the anodising tank (the load) entirely above the power transistors (between PSU+ and their collectors). With the right design the only parts exposed to full PSU+ voltage and any high energy output spikes would be the power transistor collectors. You don;t need voltage regulation at all, anodising is a current process so what you need is good current regulation and high ruggedness.
 
@CEEFNA,
We need to know the authenticity of the sch so that we can discuss.
suppose we draw one on reverse engineering, mistakes are likely to be there and we waste time.
Instead the diagram is extracted from the original equipment, it would be better.
please comment on the case here.
 
Just checking the obvious.....you do have the chip 'right way round' in its socket?
 
Thanks for all the replies, The schematic was from the internet so I had no idea how good it was or bad it seems. The 723 was in the correct way up :D
Roman, you have helped me out a lot in the past( the anodising supply where I linked a lot of LM338's together and the linistepper astro tracking mount that you had considerable input in)
I respect your judgement so the 723 has gone and I will be using a LM338 instead, I have got lots spare as the last power supply was super sensitive and didn't like charging my car battery one morning when I was late for work! That was the end of that lot of 338's.

I will connect the 338 as the "precision current control" circuit on the data sheet using my new 1K pot, I just need to know how to connect that to the existing transistors? and add protection of course

Thanks again for your time.

Ceefna
 
Hi Ceefna, looking at it again the LM317 or LM338 might not be the ideal controller as normally you would use PNP pass transistors with one, and you already have 2N3055 NPN pass transistors. Instead I would use a single NPN near ground as the current regulator, turned on when it gets >0.6v on the current sense resistor.

That should be extremely rugged and stable enough, but it has the drawback that current is set by the choice of a power resistor. Or if you want to use a pot the current adjustment is limited to about a 3 to 1 range via pot adjust. Of course you can have "ranges" selected by switching in different power resistors, so with say a single switch and two power resistors you could have two ranges;
low range = 0.5 to 1.5A
high range = 1.5 to 4.5A

If you are happy with that type of limitation then I'll rough up a circuit. Basically the main improvement you get with this is removing anything sensitive like an op amp, no need for regulated supplies to drive chips, and the entire electronics is in series with the load and on the low side, so only the power transistor collectors are exposed to the anodising tank and any shorts/spikes etc.
 
Hi Roman, that sounds perfect and a rough circuit would really help. I could do with it to cover up to 8A though as I have been doing a few large parts lately.

Thanks very much

Ceefna
 
Last edited:
Here's a very rough circuit, basically the power transistors are all turned on full, until the voltage on the bottom NPN transistor base gets to >0.6v when it reduces drive to the power transistors.

The 500 ohm pot and 220 ohm resistor give a range of adjustment on the current sense resistors of about 0.6 to 2v, (a current range of 3.3 to 1). The a switch can select high current range by adding another current sense resistor in parallel.

It's very crude and "old school" but there should be very little that can get blown up, unless you do something silly like reverse the PSU leads. ;)
 
Thanks Mr RB, that looks perfect. I will put it together this weekend and let you know how it performs.

Thanks again for your time.

Ceefna
 
One more thing, that bottom transistor needs to be reasonably cool and under some conditions may dissipate about a watt. I would put it on its own heatsink, something at least matchbox sized. :)
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top