### Author Topic: LM317 specifications question  (Read 12829 times)

0 Members and 1 Guest are viewing this topic.

• Regular Contributor
• Posts: 175
• Country:
##### LM317 specifications question
« on: February 22, 2024, 12:38:33 pm »
Hi

I have a 48V POE source that I need to step down to a stable as possible low noise 30V. The 30V is used for a bias voltage of a avalanche photo diode (APD). The APD requires at most 2mA continuous current when activated. I will have a small capacitor near it.

My question
The LM317  have a Absolute Maximum rating of 40 between Vin and Vo  (https://www.ti.com/lit/ds/symlink/lm317a.pdf)
I will have Vi - Vo = 48 - 30 = 18V over the device.

What about the 48V to ground on the input terminal? Will that be a problem?
I cannot find anything that indicate 48V on the input would be a problem.

P.S.
I need to build this circuit in a hurry. So no time to source the LM317HV with abs. max. of Vi-Vo = 60V
I will also ensure the Minimum load current requirement of 10mA max is satisfied by adding a load resistor on the output.
« Last Edit: February 22, 2024, 12:44:10 pm by Cicada »

#### Ian.M

• Super Contributor
• Posts: 12892
##### Re: LM317 specifications question
« Reply #1 on: February 22, 2024, 12:49:33 pm »
Depending on the risetime of the 48V source at power on, it may be a problem as there can be >36V from Vin to Vout as the small capacitor initially charges.  The fix is a resistor in series with Vin and a Zener direct from Vin to Vout, chosen to limit the transient voltage to <36V but not conduct in normal operation (i.e. >>(Vin-Vout) ).  Try 1K and a 24V Zener.

The following users thanked this post: Cicada

• Regular Contributor
• Posts: 175
• Country:
##### Re: LM317 specifications question
« Reply #2 on: February 22, 2024, 01:06:02 pm »
Depending on the risetime of the 48V source at power on, it may be a problem as there can be >36V from Vin to Vout as the small capacitor initially charges.

Good observation. I will implement the protection as prescribed.

• Regular Contributor
• Posts: 175
• Country:
##### Re: LM317 specifications question
« Reply #3 on: February 22, 2024, 01:12:02 pm »

I did notice now in section 5 of the document they give a input voltage limits for the different LM317 devices.

So they do limit the input voltage to 40V on the non-HV option devices. HV option can have up to 60V on input.

Taking into account that my current draw is so low I can maybe build some sort of a pre-regulator circuit.

#### Andy Chee

• Frequent Contributor
• Posts: 711
• Country:
##### Re: LM317 specifications question
« Reply #4 on: February 22, 2024, 02:03:34 pm »
For reference, I propose using the circuit depicted in Figure 10 of the pdf

https://ixys.com/Documents/AppNotes/IXAN0063.pdf

The problem of course is, not everyone has a depletion-mode MOSFET in their parts bins

#### kimballa

• Contributor
• Posts: 32
• Country:
##### Re: LM317 specifications question
« Reply #5 on: February 22, 2024, 03:59:38 pm »
If you've got an LED handy, stick that in series with the load resistor to GND, now you've got an "ON" indicator for free.. As long as you're already using that 10mA current, may as well make it useful.

10mA is right down the center for what a typical LED can handle.

#### Ian.M

• Super Contributor
• Posts: 12892
##### Re: LM317 specifications question
« Reply #6 on: February 22, 2024, 04:42:45 pm »
If you've got a TL431 shunt regulator in your parts box, I'd go with that rather than messing with a LM317, protection and an extra load resistor.  Use a 4K7 dropper which gives 3.83mA to play with.  2mA for the load, and the rest for the Ref divider and to satisfy the TL431's min 1mA current to regulate to spec.  Unless you have a TL431C you may have to include a trimmer in the divider to fine adjust the output voltage and make sure your 'small capacitor' decoupling the 30V doesn't violate the TL431's Stability Boundary Conditions (see plot in datasheet), which isn't hard to do at 30V VKA - it simply mustn't be between 400nF and 900nF!

The following users thanked this post: Cicada

#### wraper

• Supporter
• Posts: 16943
• Country:
##### Re: LM317 specifications question
« Reply #7 on: February 22, 2024, 05:17:46 pm »
You could add a Zener diode in series to reduce input voltage. Do not forget that LM317 requires 5-10mA minimum load depending on what flavor of LM317 you use, voltage divider can act as such if resistor values are low enough.

#### cosmicray

• Frequent Contributor
• Posts: 309
• Country:
##### Re: LM317 specifications question
« Reply #8 on: February 22, 2024, 06:49:02 pm »
Since you mention POE as the voltage source, my experience is that various POE want a supply side voltage (and the voltage injection end of the POE circuit) and a minimum voltage at the far end of the POE circuit (to account for long runs of POE wire, and the resistance of the wire causing voltage drops). Do you have a good estimate of what the minimum voltage into the LM317 could actually be ?

n.b. I once set up something custom, to a device rated for maximum 24v POE, but from deep cycle FLA batteries, which were typically delivering 12.6-12.0v. It worked fine, because the POE device was designed for a minimum voltage of 10v.
it's only funny until someone gets hurt, then it's hilarious - R. Rabbit

• Regular Contributor
• Posts: 175
• Country:
##### Re: LM317 specifications question
« Reply #9 on: February 23, 2024, 12:55:06 pm »
If you've got a TL431 shunt regulator in your parts box, I'd go with that rather than messing with a LM317, protection and an extra load resistor.  Use a 4K7 dropper which gives 3.83mA to play with.  2mA for the load, and the rest for the Ref divider and to satisfy the TL431's min 1mA current to regulate to spec.  Unless you have a TL431C you may have to include a trimmer in the divider to fine adjust the output voltage and make sure your 'small capacitor' decoupling the 30V doesn't violate the TL431's Stability Boundary Conditions (see plot in datasheet), which isn't hard to do at 30V VKA - it simply mustn't be between 400nF and 900nF!
This will work. I will be able to get TL431C's in time to finish this project.

• Regular Contributor
• Posts: 175
• Country:
##### Re: LM317 specifications question
« Reply #10 on: February 23, 2024, 12:55:50 pm »
You could add a Zener diode in series to reduce input voltage. Do not forget that LM317 requires 5-10mA minimum load depending on what flavor of LM317 you use, voltage divider can act as such if resistor values are low enough.
Something sort of like this in attached file?

• Regular Contributor
• Posts: 175
• Country:
##### Re: LM317 specifications question
« Reply #11 on: February 23, 2024, 12:57:33 pm »
This will be a very short run cable.

#### Swake

• Frequent Contributor
• Posts: 562
• Country:
##### Re: LM317 specifications question
« Reply #12 on: February 24, 2024, 09:32:04 pm »
Careful, PoE is not that simple:
1/ The real spec for the 48V of PoE is a voltage at the device between 37V and 57V DC.
2/ At start the full PoE power is not yet present. The connected device must ask for it by placing a resistor over the power pair. Else the switch will not put the PoE power on the line.
3/ There is a minimum consumption requirement, else the power is switched off as the switch will consider the device to be disconnected. I don't remember how much, but definitively more that 2mA.
4/ Power can be supplied over the data cable pairs or over the spare cable pairs. It is the switch that decides what pair to use. Your device must/should support both.

The reality is more complex that the above, but it gives a good idea.
The simple solution consist in using a ready made PoE module. It is like 5 bucks from aliexpress. Depending on what your requirements are in terms of low noise and stability you might want to filter the output.
When it fits stop using the hammer

#### iMo

• Super Contributor
• Posts: 4817
• Country:
##### Re: LM317 specifications question
« Reply #13 on: February 24, 2024, 09:48:54 pm »
For example..
I got to the very edge of the abyss, but since then I have already taken a step forward..

#### wraper

• Supporter
• Posts: 16943
• Country:
##### Re: LM317 specifications question
« Reply #14 on: February 24, 2024, 10:58:05 pm »
If you've got a TL431 shunt regulator in your parts box, I'd go with that rather than messing with a LM317, protection and an extra load resistor.  Use a 4K7 dropper which gives 3.83mA to play with.  2mA for the load, and the rest for the Ref divider and to satisfy the TL431's min 1mA current to regulate to spec.  Unless you have a TL431C you may have to include a trimmer in the divider to fine adjust the output voltage and make sure your 'small capacitor' decoupling the 30V doesn't violate the TL431's Stability Boundary Conditions (see plot in datasheet), which isn't hard to do at 30V VKA - it simply mustn't be between 400nF and 900nF!
This will work. I will be able to get TL431C's in time to finish this project.
TL431 would be the right choice here because it allows for wider input voltage range and likely with less power waste. Just be aware that it can still heat quite a bit, so don't push more than a few mA through it or use SOT-89 package soldered to a copper polygon on PCB.

#### Zero999

• Super Contributor
• Posts: 19601
• Country:
• 0999
##### Re: LM317 specifications question
« Reply #15 on: February 24, 2024, 11:09:29 pm »
You could add a Zener diode in series to reduce input voltage. Do not forget that LM317 requires 5-10mA minimum load depending on what flavor of LM317 you use, voltage divider can act as such if resistor values are low enough.
Something sort of like this in attached file?
Just put a zener between the input and output. A series resistor can be added to limit the surge. The downside to this is there's no short circuit protection. You could increase the value of the series resistor or add a small, fast blow fuse.

The following users thanked this post: Cicada

• Regular Contributor
• Posts: 175
• Country:
##### Re: LM317 specifications question
« Reply #16 on: March 01, 2024, 03:01:29 pm »
This is what I built in the end. Runs quite well. No parts get hot. All within their power specs.
See diagram attached.

#### cosmicray

• Frequent Contributor
• Posts: 309
• Country:
##### Re: LM317 specifications question
« Reply #17 on: March 02, 2024, 12:28:33 pm »
Keep in mind that RL will be dissipating ~330 mw. At minimum use a 1/2w device for that.
it's only funny until someone gets hurt, then it's hilarious - R. Rabbit

Smf