Where to go from here

[G0HZU]

Not at all Richard. I cut my teeth on octal base valves.
I respect your detailed circuit analysis, however the OP and the thread name suggests limited knowledge and experience so IMO it is best to apply the KISS principle.

With a 317/338 it is possible to build a "rats nest" PSU to get him started, try that with a 723.

No offense meant, tautech. 
I agree that 723 is pretty long in the tooth and there are more modern (easier/faster/better/cheaper) solutions to that problem.
That is, indeed, a decades-old ham-radio solution.

@Mery128, you would probably be better off with a more modern circuit design with modern components.

For example, here is a more modern version of that circuit...
http://www.circuitdiagram.org/12v-10a-power-supply.html

That 'modern' 12V PSU circuit looks pretty hideous to me (for several reasons) and I don't think it will even work properly because the component values are obviously wrong anyway.

The overall design is so bad it isn't even worth considering and this is one to avoid IMO

[Richard Crowley]

That 'modern' 12V PSU circuit looks pretty hideous to me (for several reasons) and I don't think it will even work properly because the component values are obviously wrong anyway.

One to avoid IMO

What do YOU think is wrong with it?  I have seen manufacturers' application notes with virtually identical circuits. That arrangement is quite common. Perhaps you are only seeing it for the first time.

[G0HZU]

Are you serious? It obviously won't work (as a 10A PSU) with the 0.1R resistor at the input to the regulator. I can see how it's 'supposed' to work but even if you correct the obvious component errors it's an awful design. The 470uF cap is too big.... there is no soft start... the fuse protection is shite...

It's complete crap.

[Richard Crowley]

It is a minor variation of the circuit right out of the LM78xx series application notes.  Perhaps you don't understand how it works.

http://www.ti.com/lit/an/snoa660a/snoa660a.pdf

[G0HZU]

It is a minor variation of the circuit right out of the LM78xx series application notes.  Perhaps you don't understand how it works.

Perhaps you don't realise the impact of the 'minor variations' ?

[Richard Crowley]

If you want to publish a "correction" or "improvement" feel free to do so.
ANY circuit can use "improvement" for some factor that you consider more important than the original designer.
Meanwhile many people have used THAT circuit quite successfully.
You may not like it, but that doesn't prevent it from working.

[G0HZU]

Meanwhile many people have used THAT circuit quite successfully.

No they haven't.

How much current does the regulator have to pull through the 0.1R resistor to turn on the TIP2955s? If the 7812 pulls an AMP then there's only 0.1V drop across the 0.1R resistor.
How can it possibly work as a 10A PSU with a 0.1R resistor at the input to the regulator?

[Richard Crowley]

If you have actually shown that it doesn't work, suggest that you notify the website where it was published.
Did you actually read the datasheet and the app notes?

[G0HZU]

Did you actually read the datasheet and the app notes?

There's no need really... variations of this circuit (developed from the Fairchild 7812 datasheet) are all over the internet and there's no shortage of people who think their variation works properly. There's another popular 'variation' below and it is also riddled with basic design flaws.

http://www.zen22142.zen.co.uk/Circuits/Power/1230psu.htm

[theatrus]

While we may have gone off track, this does represent a learning opportunity.

Mery: build some of the variants, and discover why they do and don't work.

[Richard Crowley]

Go directly to the manufacturer's application notes. Even G0HZU can't argue with them.

[Mery128]

Well its like Dave said in that one video. "I hope your project doesn't work!" Best way to learn.

At this point I've got alot of reading and some simulating to do. Which is a perfect way to kill time while parts I ordered take their sweet time to get here.

[G0HZU]

Go directly to the manufacturer's application notes. Even G0HZU can't argue with them.

Why don't you practise what you preach?  Maybe if you read the manufacturer's application notes and analysed the circuit below then you wouldn't find yourself posting up a circuit that doesn't work as you claim 

The arrowed resistor is 0.1R.

This resistor is a key part of the PSU design. It's value depends on the max design current for the PSU, the max design current of the linear regulator and the current gain of the pass transistors and the Vbe drop. There is a simple design equation for this.

It isn't going to work as a 10A PSU if you fit a 0.1R resistor here. Your circuit below is WRONG.

It doesn't work as a 10A PSU. It will presumably work up to the limit of the LM7812 which will be typically much less than 2A. Also, once the 0.1R resistor value is corrected, the 470uF cap next to this resistor is about a thousand times bigger than it should be and this will screw up the response time of the feedback loop. This will cause all manner of voltage spikes on the output.

Even if you fix both of these issues the circuit is still crap. The spread in current gain of the pass transistors, the lack of overvoltage protection, the lack of soft starting mean that this PSU is a disaster for a beginner to build.

It isn't a 'modern' design either. The application note these designs are based on in the Fairchild datasheet dates back many, many years.

[G0HZU]

Well its like Dave said in that one video. "I hope your project doesn't work!" Best way to learn.

At this point I've got alot of reading and some simulating to do. Which is a perfect way to kill time while parts I ordered take their sweet time to get here.

If you want to analyse a bad PSU circuit then analyse Richard Crowley's suggested circuit (that he still seems to be defending?). The first thing you will find is that is doesn't work properly because the 0.1R resistor is too small in value.
So it will only perform about the same as a standard 7812 regulator on its own.

The next step is to download the latest Fairchild datasheet for the LM7812 (eg the 2013 version) and look up the design equations for the regulator when used with PNP pass transistors.

This will help you design a more realistic value for the 0.1R resistor feeding the regulator although you also have to factor in the other 0.1R resistors used is series with the TIP2955s. However, the circuit suggested by Richard still has a serious design flaw because of the large 470uF capacitor across the input to the LM7812. This huge value capacitor is NOT there in the Fairchild app note and for very good reason...

Try simulating the circuit with a step change (down) in load from several amps to a light load of a few mA. You will then see the circuit develop very significant overshoot spikes on the 12V output. These could easily kill the device the PSU is powering. The large 470uF cap causes this problem because it is a about a thousand times bigger than it should be.

It's an awful 'variation' of the original app circuit from Fairchild. Even if you correct the 0.1R resistor and the 470uF capacitor the circuit is still not recommended. I can tell you why this is so if you like but my best advice is to run away from this circuit and design something better. eg something with decent short circuit protection, decent overvoltage protection and a decent/adjustable current limit feature

What is really sad is that if you google TIP2955 then the first images you see are of various PSU circuits showing hobbyist developed variations of the Fairchild app note. Many of them have the 470uF capacitor across the LM7812 regulator and I guess all these circuits are derived from the original 'variation designer' (idiot?) who put the huge 470uF cap there in the first place