I saw this post by Liquibyte, and decided to take up the challenge of offering suggestions. First, I'd like to say that I've enjoyed looking through this thread, and am very impressed with the work that has been done here by Liquibyte, Audioguru and others. I have made some alterations to this new design, BUT I've not yet constructed it. However, there are a number of things that I discovered when simulating with LTspice.
In the attached, I've made a number of alterations - rather than using a diode to ensure that the voltage reference is pulled down only when U1 is active, I've used a transistor. The reason is that if the output were short circuited, then the inverting input of U3 would be zero, so U3 can never get the inverting and non-inverting outputs equal. Hence the current is only limited by the output transistor drive capabilities since the minimum of the non-inverting input is the diode voltage drop plus the minimum output of U1. Instead, I've used a transistor which can pull the non-inverting input of U3 very close to 0V (VIN-), which will be less than VOUT-, so the output of U3 will drop, and cut the voltage at VOUT+.
The second major difference is to remove the unnecessary output transistor Q4. A single 2N3055 can drive up to 15A, so doubling up is not necessary. (In fact, unless the transistors are well matched, it may have no effect whatsoever as at these "low" currents one transistor would carry the bulk of current if the matching were not perfect). I was puzzled by the need for R13 until I realised that this is supposed to be part of circuitry to limit the current output of Q3. I'm assuming that at some point in the past, the diodes that make this work were dropped - perhaps to keep the voltage up under high drive conditions? In the redesign I've added three diodes from VOUT+ to the base of Q2. If U3 tries to drive too much current through Q3, then the excess drive current will be routed through the diode chain, effectively limiting the drive current into Q3. If the voltage is dipping under heavy load currents (less than the limit of 3A), then the output resistor can be reduced in size. However, the consequence is that the maximum output current spike will be higher.
One other final structural modification is to separate the voltage sense and current sense parts of the circuit to allow the common Voltage/Ammeter module to be installed. As I've not built the circuit yet, I don't know what the internal resistance of the ammeter is, so I've put a nominal 0.15Ohm resistor in there. Obviously this should be changed to whatever is appropriate.
With a little reorganisation in the drawing, which includes some test circuits: a 555 astable which has power spikes as it is not decoupled, a resistor that is suddenly shorted out at 0.5s, and a resistor that varies between an almost infinite resistance to zero in a way that should make the current rise linearly up to 0.5s, and then the voltage drop linearly. Also, I've included a period after 1s when the supply is switched off. When I did this, I found it necessary to modify the power supply to U1 to ensure that it continued to limit current, otherwise I ended up with a large current spike after the power source had been removed. I've also added a few notes discussing things I've discovered when changing things around. I'm offering this revised (but note UNTESTED) design to the forum.
Just one note, I've seen pictures of some builds, and am slightly concerned that the casing of 2N3055 is not covered by everyone. The case of the 2N3055 is the collector, which will be at around 38V. If someone connects VOUT- to earth, then the full 38V will be present on the case, with whatever current the transformer can supply. Under the right (i.e. wrong!) conditions, this could be fatal. Please make sure the case is covered!
If anyone builds this, then I'd be really interested to hear how they get on. If you do, please try to understand how the design works first - working with high currents with an untested circuit is not something that should be taken on by those not very familiar with circuit design/construction. My plan is to build a lower current/voltage version of the supply (I really don't need 30V, nor 3A), but finding the time to do this is my biggest problem...