Will 50V into an LM317/LM337 destroy it? Time is of the essence

Hi guys.

Just working on a preamp I am building and I blew two OPA4134s (ouch) because the power supply dumped +-30V onto the power rails. The rectified DC reaches +-50V. I need +-15V. I do not have another transformer so I am in need of some help. This problem needs to be overcome today and I have very limited resources.

The datasheet for the LM337 says 'input to output differential voltage = 40V' and since the differential I was asking of the device was 35V, I assumed this would be OK. In hindsight I do not think my assumptions were correct lol. Is it possible, with two devices in series, I can step down the voltage once, and then once again? My idea is to use a resistor divider network to put the ground potential of the first pair of devices to like 20V, so the regulator itself only sees a max potential difference of 30V anywhere. I could then knock off say 20V with these devices. Then the second pair of devices with the normal ground potential (0V) to regulate the lower voltage down to +-15V.

I know having a lower voltage xformer would fix this straight away but I simply do not have one. I also only have two LM317s and two LM337s left, I cannot afford to blow these.

Any help is greatly appreciated.

Dan

 

It is best you post some kind of schematics I think. The max input voltage to LM317 is 38 volt

 

This is what I am envisioning, this circuit runs fine and drops the voltage to 33.8 after the first regulator, then 15.1 after the second. However using a resistor divider network (10k and 10k) for the 25V ground ref did not work, I had to add in a second power source. Any ideas why?

 

This does not work - no matter what I set R4 to, I get 47.3V out from the first regulator. Some kind of isolation is needed?

 

I would do mate but I can't get hold of them today, I need something that works now. I have discovered that if you keep the bottom resistor in the resistor divider network below about 1.2K, the circuit seems to work:

 

Page 18, first drawing.

 

Thank you, I do not know of this kind of circuit, will it allow for a higher input voltage than 37V?

 

It look like the the H type is rated for 40 volt as Absolute Maximum Rating. So you will fry it. If you have an extra small transformer use that.

 

I would have if I had one =[ Any thoughts on the second circuit I posed? Is there any reason why this would not work?

 

I posted that circuit because you have 317H chips and it will work. The limitation of the chip is about the voltage difference across the chip, not the absolute voltage of the supply. In that circuit, the first 317H chip will only see part of the voltage. Calculate the resistors to drop the desired amount of voltage in the first chip.

 

brillant, thankyou very much. i will get working on it

 

I'm guessing The same can be done with the 337 too?

 

I believe it will work with a 337 chip.

The whole purpose of your question is, "how to get a 40 volt chip to control 50 volts". The obvious answer is to get them to share the voltage. Another way to skin this cat is to use a resistor or a zener to waste some of the voltage but you did not say how much current is needed or whether it is a steady current. That makes the resistor/zener approach hard to calculate.

Your circuit in post#6 might work but I find it easier to hand you a simpler circuit than to teach you about the bias network. I'm just lazy.

 

Your circuit is Post #6 has a strange configuration. Why are you not just using the normal two resistors in the bias network to get your desired voltage? For 35V that would be 240Ω for R3, 6490Ω for R4 (to ground), and remove R1 and R2.

 

You could build a small supply for the LM317, and use a pass transistor.

Usually this would be done by a Z-Diode, or a chain of Z-Diodes.

And you can try this in LTSpice, the LM317 model is around on the net.

 

Why not use a darlington transistor as an emitter follower regulator, with a resistor and zener diode on it's base. It's a very proven design used in a lot of old commercial equipment to drop >40v to a regulator friendly voltage. Something like a TIP122/121 is commonly available and good for >100v.

 

Read post #1. Needs to be done today with very limited resources. No time to go buy zeners and Darlington transistors.

 

See the attached. It's basically the same as the circuit in the datasheet; just drawn a bit differently.

I've added caps to ground from the ADJ terminals; this will help to keep the output more stable.
If the input will be more than a couple inches from the filter caps, use a .33uF cap from the input terminal to ground.

You will need heat sinks on the regulators, even with a light load.

 

Sgt I have pretty much done what you have outlined. Thanks a ton for putting the effort in and making me a schematic!! #12 your suggestion was bang on.

I have successfully built the circuit, and it works f***ing fantastically!!! Thankyou all for your suggestions, I knew this board was the place to come in a pinch.

This is what I have made:



I will post the fully completed preamp in the projects section, for anyone who is interested in having a peek.

Thankyou all again,

Dan