Improved LM317 ?

[buchignani]

No matter what I have tried, I have never been able to ully read the values in the schematic image in this article.

(The figure is at: http://archives.e-insite.net/archive.../di_03fig1.htm ).

Does anyone have a more clear version available?

Thanks

Norman

[MBK]

Hi,

I have built a regulator based on BB OPA548, a power amp chip. The OPA548 has an internal voltage reference that can be used to generate the first voltage, this then gives the reference for the second chip, to generate the negative voltage. I implemented this very straightforwardly using the datasheet (another version of OPA548's used like that can be found on headwize.com). It is driving my preamp prototype, still on breadboard :-) This should be somehow similar to the Walt Jung concept, except for the power devices which are not needed, the OPA548 can deliver 3A cont.

Now my question: cranking up to full gain (15dB) I can still hear some faint line ripple from my woofers and some noise from the mids. How would I go about improving such an OPA548 based concept? Is "bootstrapping" possible w/o adding power transistors (that would defeat the purpose of using a power op amp in the first place) and w/o using another regulator, say 78xx/79xx, in front (the headwize.com project uses LM317 before OPA548 - I find that is a bit much of brute force, elegance is always appreciated ).

MBK

[jarthel]

Did you build the Gilmore-designed amp?

Jayel

[ALW]

There's not much difference between the original TAA schematic and the EDN update, but hopefully this will help.

A.

[jarthel]

Thanks Andy.

Jayel

[MBK]

Jayel,

s'ppose that question was addressed to me.

No, I did not build the Gilmore amp. I just happened to have a lot of OPA548 from times when BB/TI sent out up to 10 samples per item :-) ... I use them as power amps actually, in a bi-amped system with active x-o. I will use those as headphone drivers soon as well, I am now at a stage where I finally start to have a halfways decent understanding of op-amps so I'll stick with chips for the time being... I am very happy with OPA548, they have similar specs to LM3886 which is highly regarded in some circles; for some reason the OPA548 does not get attention.

Anyway, it can be used as low noise regulator as well and I needed that for my op-amp based DIY preamp. Gilmore uses it as second stage after a 317 and with a precision external voltage reference, me I used the built-in voltage reference as per datasheet, the rest of the circuit is similar to Gilmore's. I just wanted to know if one can improve on that.

MBK

[carlmart]

Hi!


Here's a clean drawing of the latest WJ's version for that supply.

How can Jan's suggestion be implemented on this version?



Carlos

[jarthel]

thanks

Jayel

[jarthel]

Some clarification

Are Term3/6/20/10/17 connected to +ground? Are Term11/13/19/18/15 connected to -ground?

Is -ground connected to + ground? or There are 2 circuit grounds?

What's the purpose or +/-load and +/- sense? In the schematics, these are Term2/9 for positive voltage and Term14/17 for negative voltage. Where are these terms connected to in the circuit?

Thanks

Jayel

[ALW]

These reg's have remote sensing options, i.e. the ability to regulate at the load.

In order to do this Term 9 is connected to the +ve rail at the load and Term 10 is connected to 0V at the load. If using this option though the 10R / 0.01u shown in my diagram may need to be implemented.

If this option is wanted connect Term9 to Term2 and Term 10 to Term3.

Term 2 is the +ve feed to the load, Term 3 is the 0V feed to the load.

Additionally Term6 needs returning to the raw supply 0V (e.g. smoothing caps).

Term 7 connects the -ve sense point to a ground plane on the PCB.

P.S. change the values in the tracking pre-reg options to R101=R102=R103=R104=249R.

A.

[carlmart]

Well, I think the best person to answer or comment those questions would be Jan himself.

You should also read the original articles, as all aspects are commented on and explained on them.

Grounds (+ and -) are all separated and joined only on one point on each board. There's also a ground plane.

The sense points serve to include the feedback from the powered chip entering point, that is there will be a resistor connected just at the pin point. And you should only power one chip with each +/- pair.

It is a very tricky circuit, and I'm not sure it's very DIY.


Carlos

[jan.didden]

Ref ALW's circuit above: you can implement the CRD by deleting everything around the Q2 (I think it's Q2, can't read it very well) current source: delete the LED, the 10k, R7, the transistor. Connect the CRD between Vraw and the junction of the zener and Base of Q1 (the output transistor).

Note: I have NOT discussed this with Walt Jung. Since ALW's circuit is basically Walt's further development of our common baseline article, he may or may not agree to this mod. I only can tell you that it works for me with a Vraw to Vout reserve of at least 6 volts at max load.

On the question of the physical implementation of the AA articles, on the board basically everything is returned to the star ground. This is quite important. The star point is the only zero volts point on the board, because when I laid out the board I decided that it should be the zero voltage point. (See what you make of that!). The consequence is that everything has to be referred, and thus returned, to it to make it work. Ideally, no two component leads should join a track or wire, because there will be interaction. Read the original article for more info.

Cheers, Jan Didden

[Kevinbd]

Hello,

Just wondering if anyone can give output noise specs for the jung circuits using the AD797.

Is the noise in uV's or below ? I think noise can be specified in db also, but not sure how to convert this.

Thanks

Kevin

[fmak]

[QUOTE]Originally posted by carlmart
[B]Hi!


Here's a clean drawing of the latest WJ's version for that supply.

------------------------------------------
Carl

Do you know of a source of PCB for this; and who makes the D4xH11 transistors?

Thanks

Fred

[ALW]

The circuit shown can work with noise levels of the order of 3nV/rt Hz @ 1k.

Basically the noise of the opamp (1.2nV / rt Hz peak) + the filtered reference.

In order to acheive this careful layout is needed, as Jan's post implies. My own PCB also incorporates star earthing and star- supply for all connections.

Of more importance than the noise is the output impedance, which starts at 40 uOhms at LF rising to only 1mOhm or so at several 100 kHz.

Line rejection is awesome too.

Jan's update is not something Walt would implement, since he wanted low dropout for the super-reg's and the bipolar current source is much better than a FET in this instance.

A better solution to the dynamic stability and slightly worse performance of the bipolar source is to use a pre-reg. This is Walt's latest incarnation and it uses a tracking LM317 pre-regulator, set to maintain the dropout votage of the super-reg, referenced to the super reg output.

See carlmart's post above - these reg's are totally bloody awesome!

A.