Standard LM317, can get 50V?

[DragonMaster]

Hi,

I need a ±45V regulator, but I can't find any "simple" ways of doing it except by using LM317/LM337 HV parts. The problem is that I can't find these anywhere, even Mouser doesn't have them. Anyone knows how I could get 45V off the standard parts?

I found this:
http://www.national.com/ms/LB/LB-47.pdf

The regulator sees a 5V difference, but I don't know how which resistance I should give to R6 and if the circuit is possible to build for the LM337.

[martin clark]

It's no problem at all, the voltage limit for 317/337 parts is the difference between input and output. You can use a standard 317/337 to regulate valve supplies if you like.

The key to long life is to protect the regs 'seeing' more than the 37-39v difference they are rated for, which is a risk at turn-on and switch-off. The diagram you link to does this, but an easier way is to 'clamp' the difference voltage. Add a 24v, 3w zener reverse-biased between the input and output pins of each regulator - which will avalanche first and save the regulator. It has no effect when the supply is up and running. You can then calculate the voltage set resistors as you would normally.

Yo must also include both the protection diodes shown in the LM117/317 datasheet, too.

[leadbelly]

Originally posted by DragonMaster
I need a ±45V regulator, but I can't find any "simple" ways of doing it except by using LM317/LM337 HV parts. The problem is that I can't find these anywhere, even Mouser doesn't have them.

http://search.digikey.com/scripts/Dk...me=LM317HVT-ND

Originally posted by DragonMaster
I found this:
http://www.national.com/ms/LB/LB-47.pdf

To build the Maida regulator from LB-47 would be outrageous overkill for a 45V supply. Personally, I would just use plain LM317 & LM337 and not sweat anything.

[DragonMaster]

Thanks for the replies!

So this would work?

[leadbelly]

Originally posted by DragonMaster
Thanks for the replies!

So this would work?

A few issues:

This would only work if your circuit has a significant draw. If you are feeding a few mA with this, the zener will supply the circuit needs and your regulator will be bypassed.

240 ohms is a little high for the out/adj resistor. National always uses 120 on their datasheets.

Why bother to add the input resistor?

If I was building this, I would just throw everything out except 2 resistors and the capacitor.

[martin clark]

Yes, but using 240R cuts total dissipation in R2 and R3. The reg itself.really doesn't care what R1 is, it only sets a constant-current through the lower voltage-setting resistor(s).

Adding a smallish cap to ground after R0 will help no end. 0.1-10uF is enough, with R0 this will reduce noise feedthrough and keep input impedance low at the regulator - 317s like <10ohms source impedance.

BTW the ADJ pin averages 50uA leakage current, which x 8K4 means the reg a posted will deiver about [[(1.25/240) * (8K4)] + (50uA * 8k4) + 1.25v] , about 45.5V. Not way off target, but when using large-value voltage set resistors the ADJ pin current can be significant.

Adding 10uF across R2+R3 will significantly improve ripple rejection and lower the output impedance BTW.

[leadbelly]

Originally posted by martin clark
The reg itself.really doesn't care what R1 is, it only sets a constant-current through the lower voltage-setting resistor(s).

That's not true. Read what AN-181 says about 5 mA being a typical programming current and 10 mA for commercial applications.

[poobah]

Interesting discussion...

Might I suggest something?

TL783CKC Digikey 296-10229-5-ND

Double check the part number... that might be a surface mount part.


about 69 cents... will take you up to 125 Volts...



http://focus.ti.com/lit/ds/symlink/tl783.pdf

Oh, and BTW, don't buy them from Joe's Cryo-special Speaker Cable and More. There was a bad batch a few years back. DigiKey will have good ones. Joe will have the bad ones.

[DragonMaster]

240 ohms is a little high for the out/adj resistor. National always uses 120 on their datasheets.

I see 240 everywhere, they even recommend it IIRC ;-)

http://www.national.com/ds/LM/LM117.pdf

Why bother to add the input resistor?

I bought them already...

This would only work if your circuit has a significant draw. If you are feeding a few mA with this, the zener will supply the circuit needs and your regulator will be bypassed.

!

What I want to do is to regulate the tube's voltage supply in the JLTi DIY gainclone. I got the amp working over 8 months ago, but there are loads of hum. The only time where the speaker output is hum free is during the few seconds when the amp is running on the capacitors after I switch the power off. I have to run the amp off a single transformer, unlike what the design recommends : one transformer for the tubes and one for the amp.

Why bother to add the input resistor?

More filtering = less noise ? (I already bought the resistors, I have to use them somewhere...)

Oh, and BTW, don't buy them from Joe's Cryo-special Speaker Cable and More. There was a bad batch a few years back. DigiKey will have good ones. Joe will have the bad ones.

Most likely going to be a free sample if I use it anyways

Do you know if there's a negative voltage version?

[leadbelly]

Originally posted by DragonMaster
What I want to do is to regulate the tube's voltage supply in the JLTi DIY gainclone. I got the amp working over 8 months ago, but there are loads of hum. The only time where the speaker output is hum free is during the few seconds when the amp is running on the capacitors after I switch the power off. I have to run the amp off a single transformer, unlike what the design recommends : one transformer for the tubes and one for the amp.

Whoa there. If you are ready to move onto tube PS regulation, have you already exhausted fixing the hum by troubleshooting the heater wiring and grounding scheme? Those are much more likely sources of hum in a tube circuit.