Negative high voltage reg using LM337

[knudsenuk]

I wondered if someone could kindly comment on the circuit below. I built it and it works a treat. I wanted to build a complimentary negative/PNP version but can't get it to work the same. I must be doing something wrong. I have substituted Q1 and Q2 for PNP devices and have the collectors on the negative rail. I have reversed the direction of the diodes and the LM329 and swapped the LM317 for a 337. Am I missing something??

Any help much appreciated.

[mvs]

LM337 is not 1:1 implementation of LM317 for negative supply. It might be less stable in this circuit. Try to get it working with more basic circuit (Figue 1, same app. note LB-47).

PS I hope you do not use LowESR caps (MLCC, Mylar, etc) without series resistor at the output.

[iMo]

1. double-check your wiring
2. I would try with an 8.2V D1.

[knudsenuk]

A bit of a breakthrough: reversing the potential divider........

1. double-check your wiring
2. I would try with an 8.2V D1.

Please elaborate on why you recommend this higher value diode?

[spec]

+ knudsenuk

Your approach of changing NPN transistors to PNP transistors and reversing diodes is correct.

But as mvs says in reply #1, the LM337 is not an exact compliment of the LM317.

One important area is frequency stability. The LM317 does not need an input or output capacitor.  The LM337 does not require an input capacitor either but it does require an output capacitor. Although both regulators require an input capacitor if they have more than 12 inches of wire (6inches feed and 6 inches return) between their input and the raw power supply reservoir capacitor. But, it is good practice to fit input and output capacitors to both devices, and this is the approach used in most applications.

The LM317 high voltage circuit, described in National LB47, can't use an input capacitor because that would cause excess voltage across the LM317 if the LM317 output were shorted. An input capacitor may cause other problems, I suspect.

So the bottom line is that the more tolerant LM317 is stable in the HV circuit, whereas the less tolerant LM337 is not. As a result the LM337 is oscillating, which would give all sorts of weird behaviors.

Of course, this is just a theory and the problem may lie elsewhere, but you could try connecting a 1uf low ESR capacitor from the output of the LM337 to the 0V line.

You could also try connecting a capacitor with the same characteristics from the input of the LM337 to the 0V line.

CORRECTION 2018_11_01 (see following posts about suitable capacitors)

Suitable capacitors would be 2uF, or higher, X7R dielectric ceramic types or 1uF, or higher, metal film polypropylene types.

Be interesting to hear what happens 

Datasheets:

LM317
http://pdf.datasheetcatalog.com/datasheet/nationalsemiconductor/DS009063.PDF

LM337
http://www.ti.com/lit/ds/symlink/lm137.pdf

[David Hess]

Of course, this is just a theory and the problem may lie elsewhere, but you could try connecting a 1uf low ESR capacitor from the output of the LM337 to the 0V line.

You could also try connecting a capacitor with the same characteristics from the input of the LM337 to the 0V line.

Suitable capacitors would be 2uF, or higher, X7R dielectric ceramic types or 1uF, or higher, metal film polypropylene types.

The low ESR of a film or ceramic capacitor is exactly the wrong thing to use with a 337.  A capacitor on the output is required for stability but it must have a relatively high ESR.


If I was doing this, I would make some network analysis measurements on the output to quantify the stability.  This is a good idea for any regulator but critical for voltage and current boosted designs.

Do not forget that the LM317 and LM337 have different pinouts.  This is because the substrate (middle pin and tab of the TO-220 package) of the IC must be maintained at or below the most negative voltage which is the input for the LM337 and the adjust pin for the LM317.

[knudsenuk]

Thanks, I had read about this too. People do seem to think low ESR is a panacea. I also understand that a series resistor (low value) can work with an output capacitor.
 

[spec]

The low ESR of a film or ceramic capacitor is exactly the wrong thing to use with a 337.  A capacitor on the output is required for stability but it must have a relatively high ESR.

Where do you get this information from?

The front page of the January 2016 TI datasheet for the LM337 says:

†C1 = 1μF solid tantalum or 10μF aluminum electrolytic required for stability
*C2 = 1μF solid tantalum is required only if regulator is more than 4′′ from power-supply filter capacitor
Output capacitors in the range of 1μF to 1000μF of aluminum or tantalum electrolytic are commonly used to provide improved output impedance and rejection of transients

Solid tantalum capacitors are low ESR and if you fit an aluminum electrolytic capacitor the data sheet says to fit 10x the capacitance value than for a tantalum type. This is clearly in order to get a low ESR.

[spec]

Thanks, I had read about this too. People do seem to think low ESR is a panacea. I also understand that a series resistor (low value) can work with an output capacitor.

Good decoupling is very important to get many circuits to function correctly, voltage regulators being a case. Time and again problems with circuits are corrected by using low ESR capacitors. I don't know about panacea, but all too often people do not bother with decoupling at all.

I would also be interested to learn of your source recommending the use of high ESR capacitors with the LM337.

[David Hess]

The low ESR of a film or ceramic capacitor is exactly the wrong thing to use with a 337.  A capacitor on the output is required for stability but it must have a relatively high ESR.

Where do you get this information from?

The front page of the January 2016 TI datasheet for the LM337 says:

†C1 = 1μF solid tantalum or 10μF aluminum electrolytic required for stability
*C2 = 1μF solid tantalum is required only if regulator is more than 4′′ from power-supply filter capacitor
Output capacitors in the range of 1μF to 1000μF of aluminum or tantalum electrolytic are commonly used to provide improved output impedance and rejection of transients

Solid tantalum capacitors are low ESR and if you fit an aluminum electrolytic capacitor the data sheet says to fit 10x the capacitance value than for a tantalum type. This is clearly in order to get a low ESR.

I got it from the same source you did, the datasheets.

Solid tantalum capacitors are *not* low ESR compared to film, ceramic, or polymer electrolytic capacitors which is what makes them so useful at the output of an NPN negative regulator or any low dropout regulator.

[spec]

You are right 

I got this from the ON LM337 datasheet, page 7:

Using  the  classical  tantalum  or aluminum  electrolytic  capacitor  types  with  non−reduced ESR  (Equivalent  Series Resistance)  value  is  necessary. Low−ESR or  similar  capacitor  types  with  reduced  ESR value  and  ceramic  capacitors  can  cause  instability  or continuous oscillations in the application.

https://www.onsemi.com/pub/Collateral/LM337-D.PDF

[eV1Te]

I just ran into problems with the LM337 and thought I would add a comment here for future engineers.

The LM337 was oscillating like crazy, but not causing too much ripple on the output, instead it caused most ripple on the input rail. The board was filled by many OpAmps hence I of course assumed it was one of those first, then the input power rail, lastly after desoldering all op-amps I discovered the source was the LM337.

Initially I had ceramic caps, so I added an electrolytic without change, I replaced the ceramic caps with tantalums with no change!
When I saw the ON datasheet listed by spec I started checking the ESR on my tantalum caps and they were low ESR types (100 mOhm)

I added a 1 ohm resistor in series with the largest cap and that solved the issue! (I still have small 0.1 uF caps close to ICs without series resistor, they don't seem to cause any issues).

In the future I will try to stay away from LM337 if possible, too little documentation about its ESR requirements to be safe.
Can anyone recommend a "better" negative linear regulator that do not require special ESR caps?

[David Hess]

Can anyone recommend a "better" negative linear regulator that do not require special ESR caps?

Almost all if not all integrated regulators use NPN output devices so negative integrated regulators will always suffer from greater instability because the common emitter configuration has voltage gain which depends on load impedance much more than the common collector configuration of a positive regulator.