# Mathematically challenged help for lm 317-337

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

#### semi932

I know this is basic for the pros but I'm getting confused as to the calculations.

To keep it short I won't go into how I figured the resistor values. Tried to go by the data sheet but I guess I missed the point!!!!!!!!

Trying to drop from 33v dc to 29-30vdc.

Apologize for numbness but can anyone help me?

So far my calculations have me dropping way too much voltage in my tests and I'm spinning back and forth.

Thanks

#### aardvarkash10

three volt drop is probably better done with a chain of diodes or a zener. Its right on the limit of the stable operation of an LM3x7 regulator.

#### mightydub

aardvarkash10 is correct, but in case you still want to do this:

Vout = 1.25*(1 + R2/R1) + Iadj*R2

neglecting Iadj*R2 for the moment, R2/R1 = (Vout/1.25) -1 = 23. So pick values to get the ratio close enough. Since Iadj = 50uA typical, a value of R2 in the range of 2K only changes the output by 100mV.

#### SirNickity

But if you want to try, assuming your R1 is 240, I calculate 5k + 470 for R2.

BTW, check out Paul Falstad's circuit simulator applet (Java). It's great for roughing out small circuits like this.

#### semi932

aardvarkash10 is correct, but in case you still want to do this:

Vout = 1.25*(1 + R2/R1) + Iadj*R2

neglecting Iadj*R2 for the moment, R2/R1 = (Vout/1.25) -1 = 23. So pick values to get the ratio close enough. Since Iadj = 50uA typical, a value of R2 in the range of 2K only changes the output by 100mV.

I seem to be getting confused between R1 R2.

I'm thinking to use 220R for R1 as I don't have 240 on hand.

I guess just adjust R2 until desired results.
I can drop 5v with no problem.

Just trying to regulate Gainclone 3886 supply to see if it makes improvement or unregulated.

Thanks

#### semi932

Thanks for all the leads as I'll research and hopefully learn something.

#### semi932

Great info for a learner folks.

Thanks for all the good tips. I'll try to soak it up.

Got the job done with dropping 5v and running GC at 28v.
Does sound a little tighter but maybe not quite as quick.

Quick Q if anyone looks at this thread.
Can the 317 337 be input with more than 37v?
I couldn't figure it out from the data sheet. DOH

I've got a transformer that will put out about 45v after rectification and would like to be able to trim it to about 34-35v for another application.

Thanks again.

#### mightydub

I think you're nuts to be using a regulated supply with the 3886, you have the 1.5A output current limit of the regulator in series with the current demands of the output stage. Any sustained high power draw will drain the reservoir capacitance and then your rails will collapse and the regulator will overheat.

It might make sense to do this if there were separate supply pins on the 3886 for the input stage and output stage, but the V+/V- pins are shared. See any of the books on amplifier design - Doug Self, Bob Cordell, etc. for a discussion on this.

On to your question: the limit on the 317/337 is the input to output voltage difference, not the voltage referenced to ground. See the link above.

#### semi932

I think you're nuts to be using a regulated supply with the 3886, you have the 1.5A output current limit of the regulator in series with the current demands of the output stage. Any sustained high power draw will drain the reservoir capacitance and then your rails will collapse and the regulator will overheat.

It might make sense to do this if there were separate supply pins on the 3886 for the input stage and output stage, but the V+/V- pins are shared. See any of the books on amplifier design - Doug Self, Bob Cordell, etc. for a discussion on this.

On to your question: the limit on the 317/337 is the input to output voltage difference, not the voltage referenced to ground. See the link above.

Your point is well taken and I will do some more research.
I know I've seen Cordell links and I'll follow some and try to understand.

Thanks

#### woody

You might use a cap multiplier instead of a true regulator. For an example of one look at the Pass Zen 9 article.

#### aardvarkash10

Great info for a learner folks.

Quick Q if anyone looks at this thread.
Can the 317 337 be input with more than 37v?
I couldn't figure it out from the data sheet. DOH

I've got a transformer that will put out about 45v after rectification and would like to be able to trim it to about 34-35v for another application.

Thanks again.

Yes it can - look up Maida regulator.

In essence it floats the regulator between the supply voltage and the regulated output.

#### qusp

yep, it has no ground pin, so only input output differential matters, could run at 400V as long as you didnt have more than 430v input. also, what sort of sink have you got it on, with those sorts of current draws i'm surprised it didnt fry within minutes.

but to continue, try to keep the resistance values reasonably low, but at the right ratio for the VOUT you want, keeps noise under control, also bypassing the ADJ pin works wonders. I would substitute LT1085/1033, or LT317/337; MUCH better regulators that are drop in replacements

#### semi932

yep, it has no ground pin, so only input output differential matters, could run at 400V as long as you didnt have more than 430v input. also, what sort of sink have you got it on, with those sorts of current draws i'm surprised it didnt fry within minutes.

but to continue, try to keep the resistance values reasonably low, but at the right ratio for the VOUT you want, keeps noise under control, also bypassing the ADJ pin works wonders. I would substitute LT1085/1033, or LT317/337; MUCH better regulators that are drop in replacements

Thanks.
Looked up 1085 and with my limited knowledge am a bit confused.
When the version says 1085xx(50), does that mean 50v input?

Also data sheet shows two schematics. One with adjustable output.
If I order the xx50 type can it be variable from 45 down to about 32v?
Assume I would use the value of R1 and 2 as normal?

Sounds like a good idea to me and I'll order some if it will work cause I have got another project I'm going to work on.

Actually I felt like the original sounded a lot better without regulation.
Seemed to lose some zip in the music when regulated to my ears.

I appreciate all this knowledge.

Thanks again

#### Andy5112405

Be careful. YES the LM317 can operate with a difference of 40V between its input and output.

HOWEVER, at switch on the output may see 0V due to uncharged capacitors in following stages. If the I/P is greater than 40V then the device will fail.

Care must be taken to ensure that I/P to O/P is not exceeded.

#### horizonrays

does the use of any type of regulator will benefit the LM3886 based amplifier ? If yes how ?

#### Andy5112405

Basically:

If you were to look at the ouput of the amplifier with an oscilloscope you would see a complex waveform that varies in both frequency and amplitude.

If you were to feed a single stable sine wave into the amplifier you would see a pure sine wave on the 'scope.

If the Power Supply is not stable, ie the voltage "sags" under load, then you could see the amplitude of the sine wave vary (distort). Generally the higher the output power the worse this sagging presents itself.

In order to prevent this happening you need to try to present the amplifier with supply rails that stay at the same voltage regardless of the load.

The second factor is the impedance of the Power Supply.

Consider the PSU as being in series with the loudspeaker. (It's not as simple as that but this is just for thought). If you put a large resistor in series with your speaker you will lose a considerable amount of fidelity.

A regulator will generally give you a stable voltage and a low output impedance. However, you must take into account the peak current requirements of the amplifier. This is significantly higher than the DC current that is being drawn.

A good transformer with good regulation, coupled with adequate good quality reservoir capacitors will generally provide a pretty stable supply. There are many threads talking about capacitors in power supplies, yes, it is possible to overdo things and spoil your good work.

10000uF should be sufficient for each rail, many would divide this up into numerous smaller values in order to improve the series impedance.

If the LM3886 is to be used at full power then the transformer needs to be around the (stereo) 250VA mark, the bigger the better as regulation will be improved.

Last edited:

#### AndrewT

If you were to feed a single stable sine wave into the amplifier you would see a pure sine wave on the 'scope.
if you now monitor the current in one supply rail you will see DC current when the amplifier is sending zero power to the load. As the Power to the load is increased you will see an AC current superimposed on the DC current. Actually AC voltage subtracted from the DC voltage, due to finite output impedance of the PSU.
When the amplifier sends more current than the quiescent DC supply value then the current waveform changes to rectified pulses varying from zero amps to output amps. The sudden switch on and switch off at the ends of the chopped sinewave are effectively many harmonics of the sinewave frequency and theoretically these harmonics extend to infinite frequency.
If the Power Supply is not stable, ie the voltage "sags" under load, then you could see the amplitude of the sine wave vary (distort). Generally the higher the output power the worse this sagging presents itself.
the varying current demand in the supply rail to feed the load pulses the PSU. If the PSU is regulated and it receives pulsed demand that extends to VHF then it may show signs of instability. It might ring and/or oscillate.
The design of a combined Regulator and Power Amplifier is not a simple exercise. It needs or matches the skills that a good Power Amplifier Designer has, i.e. if one can't design a Power Amplifier and stabilise it then you cannot be ready to design a combined Regulator and Power Amplifier.

Stay simple.

Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.