Linear Voltage Regulator Performing Strangely at High Voltages

Thread Starter

koya-47

Joined Oct 15, 2021
20
Hi,

I have a simple circuit (attached) running an L7812CV linear voltage regulator that provides 12V output, 1.5A max output current and is rated at a maximum input voltage of 35V (datasheet attached). It powers a GPS device that can run between 8V and 33V. The point of this circuit is to prevent the GPS device from overheating when in 24V trucks as the device gets quite hot at voltages above 12V and sometimes sustains thermal damage if the truck cab has been sitting in the sun for a prolonged period of time.

I'm currently bench testing the attached design. Something strange happens if I feed the regulator more than 26V. The GPS device still draws its usual current (about 100mA) but then suddenly drops to 40mA as if the device has gone into some kind of sleep mode then reboots itself over and over again. I've checked the voltage and current being fed to the device and it is consistent. All data is shown in the attached image.Screenshot 2021-12-23 at 10.20.41.png

When the regulator is fed 26V or less the GPS device works fine, when fed any voltage between 8V and 33V directly from the voltage source without the regulator it works fine. I've checked the voltage drop across the GPS device and the current to the device and it all seems consistent (i.e. all current drawn from the power supply is going to the device and the voltage across it is 11.7V even when it seems to go into sleep/low power mode).

Since it works fine below 26V and works fine without the regulator (ignoring the heat issue), I tried testing another circuit that I've attached below.Screenshot 2021-12-23 at 10.31.33.png

With this design, the device does not go into any kind of sleep mode even when Vin >26V. This leads me to assume the problem is caused by the regulator being fed a high voltage compared to its output, however I can't wrap my head around why it is happening despite the consistent voltage and current fed to the device.

If anyone could offer their two cents on why this is happening so I may understand and rectify the problem I would be hugely appreciative. My only other suggestion is that perhaps I've damaged something in the regulator during solder and fitting heat shrink sleeves?
 

Attachments

DickCappels

Joined Aug 21, 2008
8,674
From the datasheet:
Each type employs internal current limiting, thermal shut-down and safe area protection, making it essentially indestructible.

The symptoms you describe are consistent with thermal shut-down. Possible solutions include decreasing the input voltage, lowering the current out of the regulator, and improving the heatsink.
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
From the datasheet:
Each type employs internal current limiting, thermal shut-down and safe area protection, making it essentially indestructible.

The symptoms you describe are consistent with thermal shut-down. Possible solutions include decreasing the input voltage, lowering the current out of the regulator, and improving the heatsink.
Thank you for your response. I have attempted to supply adequate heat removal through a 3 deg C/W heat sink. I measured the temperatures on the heat sink and regulator and found that they have yet to even reach 45 degrees C. The datasheet would suggest that the temperatures could go significantly higher before the thermal shutdown is engaged. I also confirmed this by running the regulator with Vin = 26V for a period of time and the regulator ran for hours without any issues.

However, even from a cold start up, if I input a voltage greater than 26V the issue occurs before the temperatures even reach 35 degrees. Is there something else that may be rapidly heating up internally I am unaware of that is triggering the thermal shut-down? I do not have the knowledge of the inner workings of a regulator so it is something I've not looked into. Additionally, when thermal shut-down occurs, do regulators still provide the same output voltage? What within the regulator may be triggering the thermal shut-down specifically above 26V even when on for only a few seconds but doesn't trigger when fed 26V for hours?

It is worth noting in the second picture in my original post, the 7824 regulator has no heat sink, is above 50 degrees C and continues to function fine. Its datasheet seems to suggest it has very similar properties to the 7812.

I am attaching the datasheet for the heat sink I am currently using (FA-T220-64E) and the 7824 regulator (MC7824BTG).
 

Attachments

Ian0

Joined Aug 7, 2020
4,824
Have you looked at the ripple on the output? Whilst 330nF is more than the absolute minimum decoupling recommend by the datasheet, the 7805 is generally used with 10uF Aluminium electrolytic in parallel with 100nF ceramic on the output.
 

DickCappels

Joined Aug 21, 2008
8,674
When these regulators go into thermal shut down, they shut down (turn the output off) until things cool off.

I agree with Ian0, a larger capacitor on the output might be needed in your instance. Maybe your load is reactive and the regulator is oscillating.


Have you looked at the ripple on the output? (Some text removed for clarity)
While you are at it: Have you looked at the ripple on the input?

What happens at the input voltage while the output is acting up?

Just in case it gives clues, what is your "voltage source" that is feeding this thing?
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
Have you looked at the ripple on the output? Whilst 330nF is more than the absolute minimum decoupling recommend by the datasheet, the 7805 is generally used with 10uF Aluminium electrolytic in parallel with 100nF ceramic on the output.
I am using a UT33D multimeter, am I correct in assuming to measure the ripple voltage I put the multimeter in AC voltage mode and measure across the various components? If so, the multimeter is showing 25.1 across the GPS device with both an input source of 26V and 28V to the 7812 regulator and output of 11.7V to the device.

I have some 10uF capacitors, I shall try adding it in parallel with the output capacitor soon.

When these regulators go into thermal shut down, they shut down (turn the output off) until things cool off.

While you are at it: Have you looked at the ripple on the input?

What happens at the input voltage while the output is acting up?

Just in case it gives clues, what is your "voltage source" that is feeding this thing?
Then I can confirm it is not thermal shut down, the output voltage remains a steady 11.7V even when the GPS device starts rebooting.

I am using an RS-6005D DC power supply.

Feeding the regulator with 28V (where the GPS restarts happen very frequently), the input voltage remains the same (28V) and the output voltage remains the same (11.7V) even during the GPS restarts.

Using the same method with the multimeter stated above (Vin = 28V, GPS device performing restarts), the reading with the multimeter on AC voltage reads 60.5 between input and ground pin and 35.2 between input and output pins.

For Vin = 26V, GPS on and working fine, the multimeter readings are 56 between input and ground pins, 30.8 between input and output pins.

Using the same method again, the AC voltage across input and ground of the power supply is 61 when it is providing 28V and 56.6 when it is supplying 26V.

I'm not entirely sure if I'm doing this correctly and I'm unsure what these numbers mean given my multimeter (UT33D) is fairly old and the AC supply options are 200V and 500V with 100mV resolution. The voltage readings its supplying are in volts but I'm unsure what a reading of 60.5V AC means on a DC supply. Additionally, from what I can find online shouldn't my readings be in millivolts? Here is a link to the manual for the multimeter if it helps: https://www.manua.ls/uni-t/ut33d/manual?p=1
 

DickCappels

Joined Aug 21, 2008
8,674
That's a nice meter - I have several from that family (for power supply work). I have never tried measuring DC with them, but the AC scale should tell you the RMS value of the voltage. To measure AC voltage on a DC power supply with a DVM put a capacitor (0.1 uf should be enough) in series with your voltage probe. That way you will block the DC part of the signal and still be able to measure the AC part if there is an AC part.

The drop in output voltage may be your power supply dropping out of regulation, the AC measurement might show that. Your supply is rated at 5 amps -but maybe not 5A while the output is 60 volts.
 

AlbertHall

Joined Jun 4, 2014
11,952
Many cheap multimeters will not read AC voktages correctly on the AC ranges if there is also a DC voltage.
Try the meter on an AC voltage range across a simple AA battery. The meter should read zero but it probably won't.
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
That's a nice meter - I have several from that family (for power supply work). I have never tried measuring DC with them, but the AC scale should tell you the RMS value of the voltage. To measure AC voltage on a DC power supply with a DVM put a capacitor (0.1 uf should be enough) in series with your voltage probe. That way you will block the DC part of the signal and still be able to measure the AC part if there is an AC part.

The drop in output voltage may be your power supply dropping out of regulation, the AC measurement might show that. Your supply is rated at 5 amps -but maybe not 5A while the output is 60 volts.
I tried this out with another 0.33uF capacitor and the voltage reading is now 0.0V AC. Does this mean there is no measurable ripple voltage (i.e. less than 100mV)?

If so, is my voltage regulator just defective perhaps? I am so stumped on this, I may just stick with the two regulator design seeing as it allows the GPS device to work with a DC supply above 28V.
 

DickCappels

Joined Aug 21, 2008
8,674
Yes, that means that there is not enough noise or ripple on the power supply for your meter to see.

What were the conditions when you read 0V RMS ripple?

@AlbertHall These are inexpensive DVMs but I have five from this family and they compare well with my old fluke.
 

old_beggar

Joined Jan 29, 2021
39
Koya, do you happen to have an oscilloscope? If not, can you borrow one?
Also, what is the "variable dc power source"? Surely on a truck it would only vary between about 22 and 30 volts?
On 78xx series, I think the middle pin (2) is usually connected to the tab. Did you isolate the 7812?
Is the 7812 definitely genuine?

PS Well done on providing lots of info without having to be "grilled"!
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
Yes, that means that there is not enough noise or ripple on the power supply for your meter to see.

What were the conditions when you read 0V RMS ripple?

@AlbertHall These are inexpensive DVMs but I have five from this family and they compare well with my old fluke.
I ran it at both 26V and 28V input into the regulator and had the same results each time; no measurable ripple voltage.

Koya, do you happen to have an oscilloscope? If not, can you borrow one?
Also, what is the "variable dc power source"? Surely on a truck it would only vary between about 22 and 30 volts?
On 78xx series, I think the middle pin (2) is usually connected to the tab. Did you isolate the 7812?
Is the 7812 definitely genuine?

PS Well done on providing lots of info without having to be "grilled"!
The variable DC power source as stated in one of the replies is an RS-6005D power supply. Initially I wanted to use it to test the temperatures the regulator reaches to ensure it'll function in a truck that has been baking in the sun at a trucks highest voltages (as you stated, approx 30V) but found that voltages higher than 26V cause the connected GPS device to repeatedly power down and power up again. The GPS device works up to 33V when connected directly to the power supply.

I'm fairly certain the 7812 is genuine, it was bought from a reputable wholesaler (Mouser Electronics) and is manufactured by STMicroelectronics, though I suppose when purchasing things online anyone can impersonate anyone else.

The datasheet for the 7812 is attached in the first post, it is the TO-220 model, I followed the circuit diagram on that datasheet. When you say the middle pin is usually connected to the tab do you mean the back metal plate with the through hole? And please explain what you mean by isolate the 7812 and I can give it a try tomorrow.

Thank you, I learnt from my last post to provide as much information as possible!
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
11.7V sounds suspicious to me, it should be closer to 12V if you are really only drawing 100mA.

Bob
I was thinking the same thing, im concerned I may have damaged it slightly when soldering. I have a few of the 7812s, I'll try them out and see if any of them give me closer to 12V. If so, its probably safe to say this regulator's issues were caused by my careless assembly.
 

old_beggar

Joined Jan 29, 2021
39
I'm pretty sure that the centre pin (Gnd or 0v in your case) is electrically connected to the metal tab on the back of the TO220 device which you are using, which means that the heatsink will also be at 0v, unless you use an insulator - can be silicone, mica, "Kapton", etc. and an insulating washer to prevent the screw touching the tab.

I agree with BobTPH in that I would expect closer to 12v than 11.7, even though the data sheet specifies that this is (just) within limits. It is not unknown for reputable companies to be hoodwinked (or otherwise) into purchasing fake devices.
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
I'm pretty sure that the centre pin (Gnd or 0v in your case) is electrically connected to the metal tab on the back of the TO220 device which you are using, which means that the heatsink will also be at 0v, unless you use an insulator - can be silicone, mica, "Kapton", etc. and an insulating washer to prevent the screw touching the tab.

I agree with BobTPH in that I would expect closer to 12v than 11.7, even though the data sheet specifies that this is (just) within limits. It is not unknown for reputable companies to be hoodwinked (or otherwise) into purchasing fake devices.
Yes I gathered the same from the datasheet. Would that affect the performance of the regulator? The heat sink and regulator are isolated from any other connection or conductor (held in their own plastic casing).
If this information helps, I designed the same circuit with the 24V regulator I specified in the original post and ran it up to 40V (its absolute maximum rating according to the datasheet) without issue and the output voltage being within 100mV of 24V. It is the same TO-220 style casing but had no issues like this one does.

Tomorrow I shall attempt to redo the circuit using a breadboard to avoid any risk of damaging components, new regulators and capacitors will also be used. I will update you all on whether or not the same results occur.

Thank you for all your inputs. This has been a very educational experience.
 

drjohsmith

Joined Dec 13, 2021
377
My thought when I get this sort of thing,
is this is something real silly ,
at which point I'd take a step back,

can you test the 7812 without the GPS module and just a resistor load,

Start with say 12K ohm load,
that should give you exactly 12 V out,
 

Thread Starter

koya-47

Joined Oct 15, 2021
20
My thought when I get this sort of thing,
is this is something real silly ,
at which point I'd take a step back,

can you test the 7812 without the GPS module and just a resistor load,

Start with say 12K ohm load,
that should give you exactly 12 V out,
I tested it with a 12k and a 64k resistance and the voltage drop across them was 11.65V both times.

I'm retrying the circuit on the breadboard now with a new regulator. Hopefully that solves all the issues.
 
Top