Voltage Regulator questions

[jerryk]

I am working on a automotive style voltage regulator that I have.  To further my understanding of the regulator I made a schematic of it from the circuit board with the intent of fixing it as a spare since they cost $180US.  I am to the point of reasonable certainty that the schematic is correct however there is a chance that it still may contain an error.  My main goal is to verify the accuracy of my schematic drawing so if you see anything that does not make sense please feel free to comment.

Some specific questions:

 1. I'm having trouble figuring the voltage of the input pins at the opamp U1 (lm201).

 2. The lm336 is connected in a way that is different for the datasheet.  Which side of it has the 2.5 V ref in the drawing? I assume the anode side.

 3. The 555 timer circuit for the low warning blinking light looks incorrect but I have gone over it many times and that's what I come up with.

 4.  I do not understand how Q2 and Q3 interact to provide current to the gate of the S6065K.  They seem to cancel each other out since when the base of the 2N3906 is connected to the collector of 2N3904.

Any help is appreciated.

Jerry

[jerryk]

As I progressed on the repair of my voltage regulator I answered the previous questions I asked above and realized how difficult it was to read the convoluted schematic originally posted.  I have redone the schematic and hope that you guys will give this a second chance since there is still some confusion that I'm encountering in my attempt to understand this voltage regulator.

I am now trying to test this voltage regulator.  When I vary the load at the field terminal this causes the voltage at D14 (2.5 reference diode LM258 Z-2.5) to vary. 

Example 1) .25A computer fan attached to field terminal = 2.53 V at pin 2 and 5 of LM2903

Example 2) 1.7A light bulb attached to field terminal = 2.96 V at pin 2 and 5 of LM2903.

I'm not sure what the best way to bench test a voltage regulator and set it's output,  but in the case of example 2, a reference voltage of near 3 volts causes the over voltage protection circuit to kick in at over 19V rather than 16.23 at the design reference 2.5V.

So what I'm trying to understand is why does the reference voltage at D14 vary with the load of the field terminal?  And what is an appropriate load to attach to the field terminal of this regulator to simulate a realistic load of the field windings in the alternator?

Thanks - Jerry

[floobydust]

The circuit has many multi-purpose sections and a bit fun to figure out.

Q2/R1 are over-current protection for the field winding at around 5A. Ford Motor Co. had problems with brushes or field windings shorting so they added  that in the 1970's to their voltage regulators.

A BUS or SENSE over-voltage, or field-overcurrent is detected by U2A, which triggers the protection crowbar. Once triggered, would stay on until engine is stopped and restarted. To make U2A latch, the (safety) reference voltage from D14 gets pulled down by R12 once the circuit trips.

Grounding test point 0V_TEST should disable the field-current monitoring. D14 must keep the U2 pin2+5 reference at 2.5V and you can get almost 3V and field current should not affect that. I would say there is a poor connection or open ground, or D14 is seeing too much current somehow, possibly a shorted U2. It's not making a lot of sense as there are resistors to limit current everywhere so D14 should not vary above 2.5V

[jerryk]

The circuit has many multi-purpose sections and a bit fun to figure out.

Q2/R1 are over-current protection for the field winding at around 5A. Ford Motor Co. had problems with brushes or field windings shorting so they added  that in the 1970's to their voltage regulators.

A BUS or SENSE over-voltage, or field-overcurrent is detected by U2A, which triggers the protection crowbar. Once triggered, would stay on until engine is stopped and restarted. To make U2A latch, the (safety) reference voltage from D14 gets pulled down by R12 once the circuit trips.

Grounding test point 0V_TEST should disable the field-current monitoring. D14 must keep the U2 pin2+5 reference at 2.5V and you can get almost 3V and field current should not affect that. I would say there is a poor connection or open ground, or D14 is seeing too much current somehow, possibly a shorted U2. It's not making a lot of sense as there are resistors to limit current everywhere so D14 should not vary above 2.5V

Thanks for taking the time to look at the schematic - You were right.  I'm not sure exactly where the problem was but I touched the positive lead to the ground in the circuit and still had .5V which explains why I was reading 3V on a 2.5V reference diode.  I was using breadboard jumpers to connect the regulator to the power supply and 2A load so I changed that to 22 gage wiring but also noticed that the temporary jumper I had connecting that Darlington transistor collector to the mounting nuts may have been making some contact with a resistor lead on the back of the board.  All is good now and the D14 is acting like a 2.5V reference. 

I am wanting to test the SCR(crowbar latch) D1.  I'm a bit uncomfortable just hooking a 2A fuse or breaker between BUS and GND and testing it that way.  I know the SCR is fast but it still depends on the speed of the fuse/breaker and don't want to stress the IC's on the board.  Are there any modern gentle ways to accomplish the same thing.

Another part of the circuit that I'm trying to understand is how the opamp is configured.  I assume it's being used as a comparator with the adjustable voltage divider setting the crossing point of the inverted and non inverting pins.

As for the 33v zener diodes D3 and D4, I do not understand their role in this circuit.

Thanks and I appreciate the help - Jerry

[floobydust]

To test the voltage regulator's functions, you can do the SCR crowbar test without having to trip a large fuse or breaker.
The SCR crowbar section can be powered up and tested separate from the regulating/field-winding portion. Crowbar gets power from either BUS or SENSE (via D8,D9) but monitors BUS voltage via R15.

So a current-limiting light bulb (i.e. 1157) in series with your power supply to BUS, no load on the field, SENSE I would run from the PSU output ahead of the light bulb.
Increasing BUS >16.2V should trip U2/Q4/Q3/SCR. There is also OV_TEST point to inject an over-voltage trip there.
Once the SCR trips, it crowbars BUS down to a volt or so, and SENSE would have to be up to keep the circuit powered, although that doesn't matter once the SCR is latched. It does not seem to activate the blinky UV warning lamp.


There are six zener diodes for over-voltage/reverse-voltage transient protection.
1N4752A 33V 1W
D5 protects power to U1, D12 protects power to U2.
D3, D4 protect U2's inputs (in theory) but 33V is too high to actually protect the IC, and the resistors help there.
D13 protects Q5 gate. For the field circuit 1N6282A 30V TVS D6 protects Q1.

The voltage regulating portion is your usual voltage reference LM336-5.0 and op-amp LM201 U1 as a comparator looking at the SENSE voltage. Realistically, none of this operates in a linear-mode, the whole thing oscillates pulsing the field-winding on and off at a few Hz. Thus many capacitors to keep things stable. You would see this running the regulator in a running engine+battery, not on a workbench.

If this is for aircraft, I would do a test plan that checks all functions of the LR3B. The design here is made to be solid but I see a few weaknesses, from my armchair.
What parts did you end up replacing? Sometimes understanding why it failed can make repairs easier.

[jerryk]

The failure of the regulator was the fault of Q1.  It's the big Darlington Transistor on the outside of the regulator case.  The case of the Q1 is it's collector and it had rubbed through the nylon insulator and made contact with ground through an oil filter housing that I installed on the engine.  So I have replaced Q1 however, it appears to be fine.  I have tested it with my DCA 75, FWIW, and it seems to test fine.  I also put it in circuit and it works but will go with the new replacement since I have one.  What started as a repair project has turned into a learning project, hence the effort to strip most of the board and derive a useable schematic.  I went as far as replacing almost everything (for no real reason) but the resistors and put a temporary socket in the opamps location so I can test various types since the LM201a is only available in smd.  The original Thyristor did not get detected by the DCA 75 so i replaced it also, however, in a test circuit it seemed to latch up just fine.

Thanks for clearing up the role of the 33V zeners.  As I see it the circuit could have been better served with lower zener voltages in the operating ranges of the devices they are to protect.  Apparently protection at these points is not a big deal anyway since this regulator is very popular and has a reasonable reputation.  A nagging curiosity still surrounds the crowbar latch concept.  I wonder if it is possible to just kill the power to the opamp through a transistor connected to U2A's output as a way to disable the field output.

I am looking for tht replacements for the LM201a and a lower temperature rated replacement for LM336 since it's range is 0-85C.  Currently I have an LM741 in the opamp socket and all runs good but it has the same 0-85C temp range which might be problematic here in Alaska.

This week I will be testing it at an aircraft component shop that has the full bench setup to make sure it functions in the complete charging circuit.  The final step will be to remove the opamp socket and secure all the caps with some anti-vibration goop of some sort.  Not sure what to use.  Neutral Cure Silicone gets mentioned on the internet but I would rather not use a silicone rubber.  The original install had a yellow foam of sorts securing the radial components.  It was reasonable remove which is the problem I have with silicone.

I have attached a photo of the board as it stands right now.

I wish there was a "Buy That Man a Beer" button.

Otherwise - Thanks  - Jerry
 

[SeanB]

Don't worry about the 741, it will work from the full -55 to 125C military range without issue, so long as it is a National, Ti or Fairchild part, as they were made on the same production line as the military parts, to the same spec, and were just not given the full Monty testing to pass the MIL qualification, which also included a zero device failure rate on the 1000 samples per batch test. As the military parts were just a random pull of the regular run, which went through the full qualification after manufacture QC, they made sure all devices passed to the MIL spec on all batches.

I used to use the standard commercial part and run them at the mil range with no issues, and a similar failure rate as well. They would generally only have slight offset drift ( within original spec though, just drifted away from the original setting by a few millivolts) after a long period of operation at high temperature and cycling.

[jerryk]

Don't worry about the 741, it will work from the full -55 to 125C military range without issue, so long as it is a National, Ti or Fairchild part, as they were made on the same production line as the military parts, to the same spec, and were just not given the full Monty testing to pass the MIL qualification, which also included a zero device failure rate on the 1000 samples per batch test. As the military parts were just a random pull of the regular run, which went through the full qualification after manufacture QC, they made sure all devices passed to the MIL spec on all batches.

I used to use the standard commercial part and run them at the mil range with no issues, and a similar failure rate as well. They would generally only have slight offset drift ( within original spec though, just drifted away from the original setting by a few millivolts) after a long period of operation at high temperature and cycling.

That's great.  Ti's Mil spec 741's are $11.10 and LM741CN are 73 cents.  Mousers product data on the CN version lists it as rated down to -40C however the datasheet and footnotes on Ti's site give it the 0-85C rating.

I was also looking into the CA3140 but knowing this I think I will install the 741 as part of the refurbishment of the regulator.

Thanks - Jerry