Trouble with a LM317HVT

So, I have designed a circuit to drive an ATMEGA 328P on a custom board.

I am using a 24 to 48V DC Power source and attempting to use a LM317HVT to reduce the voltage down to 5V.

So values are as follows :

Vin : 48V

R1 : 330 Ohm 1/4 Watt
R2 : 1000 Ohm 1/4 Watt
Vref : 1.25V

I have a sizable heatsink mounted and have high grade Thermal Paste attached.

I will be drawing lees than 300mA from the regulated side once I have the circuit stable.

The problem I have is that the R2 resistor is getting smoking hot without any load what so ever and the output remains at 48V

Any help would be appreciated.

Thanks

You are aware that tab is OUT, it's not GND ?

Looks like I have found the problem.....

Seems Fritzing was acting up and gave me the incorrect pin layout :o

Had the LM317HVT wired incorrectly.....

That'll teach me to trust 3rd party software.... Re-designing the board and then to print, etch clean, solder.......

Good OLD LM317's ...

I happen to have about 500 NOS (New Old Stock) LM317T (The larger package like TO220).

If anyone will (even eventually) use any of these to teach something to somebody, I'll send you 6 of them if you send me a SASE (Self Addressed Stamped Envelope)..

Terry King
152 Colby Road
West Topsham, VT 05086

LOVE these acronyms...

OK?

I am using a 24 to 48V DC Power source and attempting to use a LM317HVT to reduce the voltage down to 5V.

Do you think this is a good idea? 300mA through a linear regulator from 48 to 5V
dissipates 13W, if you use a DC-DC convert you will dissipate perhaps 0.3W instead and draw
perhaps only 50mA from the supply.

The LM317HVT can't necessarily handle the power anyway - if you look at the second graph
in this datasheet you'll see the max output current for the HVT with 43V difference is about 300mA
http://www.farnell.com/datasheets/22656.pdf,
I wouldn't want to run that close to the shutdown limit!

Go for an 18--72V input DC-DC converter, they are pretty standard devices and cover your input
voltage range with a plenty of headroom.

Hi,
If you continue to use the LM317, have you got bypass capacitors suggested in the datasheet also fitted?

Tom... :slight_smile:

MarkT:
Do you think this is a good idea? 300mA through a linear regulator from 48 to 5V
dissipates 13W, if you use a DC-DC convert you will dissipate perhaps 0.3W instead and draw
perhaps only 50mA from the supply.

The LM317HVT can't necessarily handle the power anyway - if you look at the second graph
in this datasheet you'll see the max output current for the HVT with 43V difference is about 300mA
http://www.farnell.com/datasheets/22656.pdf,
I wouldn't want to run that close to the shutdown limit!

Go for an 18--72V input DC-DC converter, they are pretty standard devices and cover your input
voltage range with a plenty of headroom.

Hi MarkT

I thought about using a DC-DC converter but was unable to find anything locally that would suite the conditions it could be exposed to. (Living in a 3rd World Country where LM317HVT's and DC -DC drivers are difficult to come by - South Africa. ) The input could drop as low as 12V....

The board at it's peak would be drawing a total of way less than 100mA. One ATMEGA 328P and 3 X 3mm LED's at 20mA each. I used the reference of 300mA as absolute worst case.

My calculations of the LM317HVT dissipation was approx 4.3 W which is based on 100mA Iout.

I'll keep an eye on the Heatsink and see how it handles.

Thanks

Hi,
Please fit a 500mA fuse in the 48V line, an LED indicator to show power is present, and a series 1Amp diode for reverse polarity protection.

Tom..... :slight_smile:

Xproximity:
Hi MarkT

I thought about using a DC-DC converter but was unable to find anything locally that would suite the conditions it could be exposed to. (Living in a 3rd World Country where LM317HVT's and DC -DC drivers are difficult to come by - South Africa. ) The input could drop as low as 12V....

The board at it's peak would be drawing a total of way less than 100mA. One ATMEGA 328P and 3 X 3mm LED's at 20mA each. I used the reference of 300mA as absolute worst case.

My calculations of the LM317HVT dissipation was approx 4.3 W which is based on 100mA Iout.

I'll keep an eye on the Heatsink and see how it handles.

Thanks

So the drawing "way less than 100mA" would mean the absolute worst case was also less then 100mA, not
300mA.

Ha ha!! Sounds like you speak from experience! LM317 is not LM7805. :slight_smile:

Xproximity:
I'll keep an eye on the Heatsink and see how it handles.

Thanks

Keep a finger on the heatsink. You eye won't see anything until it's red hot and slumping into a blob of molten aluminum... at which time it will be way too late.

(actually, the LM-317 will shut itself off long before it gets that hot, but that kinda ruins the point of what I was trying to say).

MarkT:
So the drawing “way less than 100mA” would mean the absolute worst case was also less then 100mA, not
300mA.

Apologies for the incorrect data of 300mA. That was an assumption based on the use of high output LED’s which I have now replaced with lower power ones…

I ran the board for a couple of hours last night and it seems to be holding.

I’m going to be checking for another heat sink since the one I have is getting warm to hot. I was able to keep my hand on the heatsink for minutes without it burning me.

What is the safest temp I can let it run at? I’m guessing that it’s currently at 50 to 60 *C

Thanks

In general cooler is better, but semiconductors can take pretty high temperatures, like 100C on the die,
but its everything around it that suffers (PCBs can degrade over long periods if cooked, for instance).
Reliability reduces as running temperatures go up too.

If you can put your hand on it for minutes it's probably less than 50...

Xproximity:
Apologies for the incorrect data of 300mA. That was an assumption based on the use of high output LED’s which I have now replaced with lower power ones…

I ran the board for a couple of hours last night and it seems to be holding.

I’m going to be checking for another heat sink since the one I have is getting warm to hot. I was able to keep my hand on the heatsink for minutes without it burning me.

What is the safest temp I can let it run at? I’m guessing that it’s currently at 50 to 60 *C

Thanks

You know, there’s a simple way to boost the 317 (or 7805 or any other 3 terminal regulator).

(click pic for full size)
vreg.jpg

MarkT:
In general cooler is better, but semiconductors can take pretty high temperatures, like 100C on the die,
but its everything around it that suffers (PCBs can degrade over long periods if cooked, for instance).
Reliability reduces as running temperatures go up too.

If you can put your hand on it for minutes it's probably less than 50...

Data sheets usually state 100 C or 105 C for a max temperature. That's crazy. My rule of thumb is "Too hot to keep my finger on == too hot".