ECN Forum

Linear Power Supply Quits Under Load

Posted By: Potseal

Linear Power Supply Quits Under Load - 04/17/17 08:02 AM

I salvaged an old power supply from a piece of abandoned equipment. It is a linear power supply that is rated for 24VDC at 4.8A (http://www.mouser.com/ds/2/373/sl%20power_linears_ds-329280.pdf). I wanted to use it at work at my bench for occasionally testing various items. After adding a switch, indicator light (power "on"), alligator clip leads, and a fuse holder on the secondary of the 120VAC-24VDC transformer I tried powering on a 24VDC servo motor. Did not work. I disconnected the motor and measured the output voltage - 7VDC. Then I began the process of trying to figure out went wrong. This power supply uses a LM723 voltage regulator. I replaced the regulator, powered it "on", and then measured the output voltage - 24VDC. With the proper voltage I tried a load using a 20 ohm resistor. With a voltage meter attached I powered it "on" and the voltage immediately dropped from 24 to nearly 0 volts. Checked the 2A fuse and it had blown. Tried this again with a 150 Ohm resistor and it worked, fuse held. I checked the circuit board and I noticed a variable resistor connected to terminal #2 (current limit) on the LM723. Reading the specs it states:

"The LM723/LM723C is a voltage regulator designed primarily for series regulator applications. By itself, it will supply output currents up to 150 mA but external transistors can be added to provide any desired load current."

There are three 2N3055 power transistors in this circuit which I assume gives it the capability to output the 4.8A it is rated for. So is it possible that this adjustable current limit resistor is set to limit output current well below that? For some reason that doesn't make sense but everything else I have done to troubleshoot this issue has led nowhere. Any help is greatly appreciated. Thanks.
Posted By: Potseal

Re: Linear Power Supply Quits Under Load - 04/17/17 08:05 AM

[Linked Image]

The LM723 is at the upper left corner of the board with the adjustable current limit resistor to it's immediate right. The adjustment wheel was sealed with a dab of adhesive. I measured the resistance at 290 ohms then gently pried the adhesive off the resistor. It varies from 0 - 420 ohm. I wonder if it's adjusted at the factory to fine tune the circuit, sealed, and then shipped? Maybe my issue is the power transistors?
Posted By: SteveFehr

Re: Linear Power Supply Quits Under Load - 04/17/17 09:55 AM

Could be. A lot of old low-end power supplies are designed with specific loads in mind and will not put out the right voltage if you test them open circuit or use them for something else with a higher load.
Posted By: Potseal

Re: Linear Power Supply Quits Under Load - 04/17/17 04:39 PM

Possibly.
The link above shows this item was once available through Mouser. The only thing it states in the specs is:
"Short Circuit Protection
Automatic current limit/foldback"

It looks like it was intended for general purpose use. IMO, it acts as though the power transistors are not in the circuit and current is limited by the LM723 alone which is 150mA.
Posted By: gfretwell

Re: Linear Power Supply Quits Under Load - 04/17/17 07:59 PM

If the transistors are there, they are going to be using them but you might have bad one. Where is the 2a fuse, in the supply or the load? Do you have a scope? There may be a problem with the rectifier or filter that is putting AC or a bad ripple on the bulk DC to the regulator.
These things are really pretty simple devices. It looks like a center tapped transformer feeding 2 diodes, some filtering and the pass regulator.
Posted By: Potseal

Re: Linear Power Supply Quits Under Load - 04/17/17 10:55 PM

Originally Posted by gfretwell
If the transistors are there, they are going to be using them but you might have bad one. Where is the 2a fuse, in the supply or the load? Do you have a scope? There may be a problem with the rectifier or filter that is putting AC or a bad ripple on the bulk DC to the regulator.
These things are really pretty simple devices. It looks like a center tapped transformer feeding 2 diodes, some filtering and the pass regulator.


I removed the transistors and tested each one. Using a DMM set to ohms, positive lead to collector, negative lead to emitter which gave me O/L. Then I connected the base to the collector and it showed a high resistance reading.

Right now I'm testing all the caps...

The 2A fuse is on the secondary of the transformer and yes, I have an older analog scope.
Posted By: Potseal

Re: Linear Power Supply Quits Under Load - 04/17/17 11:00 PM

[Linked Image]
Posted By: emolatur

Re: Linear Power Supply Quits Under Load - 06/02/17 03:06 PM

You need more testing of those transistors. You've shown that they are capable of being "off" and that there is a BC junction, but have not tested the BE junction nor confirmed that they are capable of being "on."

The most likely layout (I can only guess, as I don't have a schematic) is that the output of the 723 goes to both 3055 bases, the 3055 collectors are connected to the "bulk dc" line, and the 3055 emitters each go through a low-ohm high-watt resistor (likely R13 and R15), the other ends of those resistors being connected to the output.

My guess is that either the 3055s are bad (failed open-circuit, ie, they never turn "on") OR they are not seeing sufficient DC on their collectors. I'm imagining that the 24V you DO see is being realized by the output of the 723 passing through the base/emitter junctions and resistors to reach the output. This means you're overloading the 723 (throwing it into current limit). Also, if the 723 were able to provide enough power, you'd end up further destroying the 3055s anyway.

My prescription is to replace both 2N3055s. They're dirt-cheap NPN BJTs. If the power supply is built the way I think it is, you should actually be able to substitute in just about any transistor having similar (or higher) power and voltage ratings. Before anybody questions me on that, I know the gain could sorta matter, but since the 3055 sucks so bad, anything modern is going to be the same or higher. smile
Posted By: JoeTestingEngr

Re: Linear Power Supply Quits Under Load - 07/05/17 04:38 AM

If I recall, a 2N3055 can handle 15A on its own. It's the world's most common series pass transistor used in linear supplies. The beta, and in your use, the alpha, is relatively low. If your supply has 3, 1 is probably acting as a driver for the other 2, in a Darlington configuration. See if the bases of 2 are connected, and if the same 2 transistor's emitters, are connected through fractional ohm balancing resistors, to the output.

If you really had a DMM on ohms, instead of diode check, you would have read only read leakage. On diode check, you would get a slightly lower reading from base to collector, than base to emitter.

Frankly, if you were going to use a meter instead of a scope to troubleshoot, you started off on the wrong foot. Most of those units fail because their electrolytic capacitors dry out. Many of them derive a V++ supply with an extra diode and filter cap, to power the LM723. Don't go changing components until you've gone through with your DMM on AC Volts. More times than not, you'll find a substantial AC ripple voltage at 60 or 120Hz. If that isn't the problem, you could spot if the circuit is oscillating, not a good thing.

The current sensing will start coming into play when you get 0.6 or more volts between the output and current sense input, because you start biasing on the current limit transistor in the LM723. I don't have my datasheet open but the XX723, or sometimes 1723, has been around for decades, and is the most common regulator IC for linear supplies with external pass transistors.

The linear supply that I have to repair quite often is a +-15V, open frame, Lambda. It uses 2 LM723s, 2 2N3055s, and the separate V++ supply. The positive section is a straight forward design, with the emitter driving the output, and normal looking connections to the error amplifier. The negative output is not intuitive because the output is the negative supply rail, and the pass transistor connects to the common point. the error amplifier of the negative regulator is driven from a divider across both outputs, to make it "track" the positive regulator.
Joe
Posted By: Potseal

Re: Linear Power Supply Quits Under Load - 07/07/17 04:30 AM

Thank-you for the feedback Emolatur and Joe. This item has been sitting on my workbench too long. This weekend is already spoken for but holiday time looms ahead. Hopefully I can get to it sooner and next week I will try to make some progress. For something that looks so simple in design it's problem is not obvious to me.
Posted By: JoeTestingEngr

Re: Linear Power Supply Quits Under Load - 07/08/17 05:48 PM

That should be a good linear supply to learn on. I should have had the datasheet in front of me when I responded. The current is sensed between pins Current Limit (2) and Current Sense (3). Check that you have <<0.6V between pins 2 and 3 to verify that you aren't in current limit. Then, use your DMM on DC, then AC Volts, with black on Pin 7 (V-). Pins 11 and 12 have to be stable and several volts above your final output voltage. If that's OK, the (typical) Reference Voltage at Pin 6, should be 7.15VDC. As with any Op Amp, the 723's main purpose in life, is to keep its inputs, pins 4 and 5, equal. VRef, or another stable reference voltage, is fed into one input, while a divided down sample of the output, is fed to the other. On supplies <7V, you would see the reference voltage divided down. It's all of the inputs and outputs available on the 723, that make it useful for both positive and negative supplies. Output current is limited only by your filtered source, and room for driver and output transistors, and emitter balancing resistors.
Of course, you don't want the 723 to work too hard if it's driving other transistors. So you would keep the PDIP well under its 150mA and 660mW maximum ratings. That's where the Alphas, or emitter current gains, of the driver and output transistors come in. A 5 amp supply with its 723 sourcing <= 50mA, would have to use transistors with a combined gain of at least 100. That's why you will see a driver transistor driving the bases of multiple, emitter balanced, output transistors.
Now, looking at the datasheet for the On Semi 2N3055, that I just downloaded from Digi-Key, the features say it all. The DC current gain ranges from 20 to 70 at 4 amps of collector current. You can see that a transistor with a gain (Beta) of 20, (Alpha) of 21, would need a base current of almost 230mA, to get an emitter current of your supply rating of 4.8A. That's more than the 723 has to offer, and why your supply will be using an external driver transistor. The 2N3055 isn't a common driver for such a small supply so I wonder if that third one might be part of the current limit or a crowbar circuit.
Good luck,
Joe
Posted By: JoeTestingEngr

Re: Linear Power Supply Quits Under Load - 09/16/17 03:37 AM

Potseal,
Did you ever get back to troubleshooting the linear supply?
Joe
Posted By: JoeTestingEngr

Re: Linear Power Supply Quits Under Load - 10/11/18 12:47 AM

Dwayne,
Did you ever get your LM723 based PS figured out?
Joe
Posted By: sabrown

Re: Linear Power Supply Quits Under Load - 10/17/18 09:06 PM

It is a bummer when you take the time to create an excellent reply and all you hear is the echo of your voice in an empty room. I hope you hear something besides the lurker in the dark, me.

Shane
Posted By: JoeTestingEngr

Re: Linear Power Supply Quits Under Load - 10/19/18 12:47 AM

Thanks Shane, I help when I can. It is frustrating when you never hear back.
Joe
© 2020 ECN Electrical Forums