Hey everyone,
Quite often I use cap multiplier regulators with an NPN or NMOS with the standard output side current limit. However, I'd like to move the current limit to the input side while keeping the pass element an Nch device, and make the current limiter a foldback type. It just seems to me that one of the benefits of the regulator is the low output impedance, which the current limit spoils via the series resistance. I've found a few where the limiter is on the input, but the pass element is Pch device. Anyone know of an elegant way to do this?

 

Here's an image to illustrate the question:

 

Low Output Impedance is handled by a big fat multiple Capacitor-Bank with very low-ESR Caps.

Do You want a Circuit-Breaker, or a Crowbar-Circuit ?,
and why is over-Current an issue ?,
what needs to be protected ?
.
.
.

 

I’d say that low output impedance is created by an error amplifier with very high open loop gain.

 

Good for ~3-Amps ..........

.
.
.

 

In very simple terms:- using a pnp device with b-e junction (emitter towards power input) across a suitable sense resistor placed before the pass device and after the input filter will give you a control element.
An inverting stage will be required, driven by the collector of this pnp device to act on the drive to the output.

 

Low Output Impedance is handled by a big fat multiple Capacitor-Bank with very low-ESR Caps.

Thank you, LowQCab!
Correct, but also the effective output impedance of the pass element is the impedance of C3 x β, or C3 x the MOSFET gate current gain (if I remember correctly), which will be magnitudes lower than the current limit resistor. Of course, this will be in parallel with said "big fat multi-cap-bank" so the difference won't be as significant. But I figure, if the cost is just a few simple parts, is it worth it?

Do You want a Circuit-Breaker, or a Crowbar-Circuit ?, and why is over-Current an issue ?, what needs to be protected ?

Preferably I'd like foldback option, where the current drops with voltage. Just playing with some ideas, it seems that a breaker/crowbar is pretty easy to accomplish with an SCR.

I like using the current limits just to protect the supplies *in case* something goes wrong. No other reason than that.

Currently I'm working on a solid-state power amplifier, but I'm thinking in the general sense here as something I could use in most/any projects. That's always bugged me a bit about the standard limit circuits.

 

I’d say that low output impedance is created by an error amplifier with very high open loop gain.

Ian0!! I'm still dreaming about running my own pub after you told me about your local place. I hope you're doing well.

It's interesting you mention this, as for whatever reason I have always avoided closing the loop on a simple regulator like this...but I have no reason why I haven't. That technically solves the problem, doesn't it?

 

In very simple terms:- using a pnp device with b-e junction (emitter towards power input) across a suitable sense resistor placed before the pass device and after the input filter will give you a control element.
An inverting stage will be required, driven by the collector of this pnp device to act on the drive to the output.

This has been how I've been approaching it. I've managed to make 2 simple circuits that work as a constant current limit, but the foldback action has been eluding me. Here they are:

 

I think that you may have some confusion regarding the different behaviors of ........
Current-Limiters,
Fold-back-Current-Limiters,
Circuit-Breakers, and,
Crowbar-Circuits.

Fold-back-Current-Limiters, Circuit-Breakers, and Crowbar-Circuits,
must be reset after activation, ( either manually, or Electronically ).

A Current-Limiter simply lowers the Voltage as much as necessary
to prevent the Current from exceeding a preset level,
and has immediate Voltage recovery when the Current demand is reduced to a safe level.

A big consideration with a Current-Limiter is that
it must turn the over-current-Power into HEAT,
unless You are using a Switching-Topology, ( which is probably the ideal plan ).

Many Switching-Regulator designs incorporate built-in Current-Limiting.
.
.
.

 

I think that you may have some confusion regarding the different behaviors of ........

Interesting, I had always thought foldback was dynamic like the constant current type, just that as the supply voltage was brought down, so too was the max current allowed to flow.

A big consideration with a Current-Limiter is that
it must turn the over-current-Power into HEAT,

This brings up the problem with what I'm seeing in the sims...with a 35V 2.5A supply, 110W in the pass element is like the best I could do in a short-circuit situation with a limiter. That's not going to protect the BJT/MOS at all, lol.

So when designing a supply to be as robust and safe as possible, what approach should be taken? I can see a couple options:
1. Just let the pass element and or fuse blow.
2. Size a current limit such that the supply's fuse should blow, but if I doesn't limit the current. (Seems a little unsafe)
3. Use a crowbar or electronic breaker, such that the supply should shutdown in overcurrent situations, with a fuse sized to back up the shutdown if it doesn't work.
4. Some combination of the above.

Switch regulators also aren't a bad idea, but I like to keep them away from audio if at all possible.

 

Why do You need a Regulated-Supply for an Audio-Amp ?
Are You building a "Class-A" Amp ?

Some of the best Amps that You can buy utilize a SMPS, and are "Class-D",
with a ~120+db down Noise-Floor.
.
.
.

 

No Class-B, I just like using them when I do audio stuff. Perhaps its overkill with a decent PSRR but I tend to over-design everything I do. It's a good way to learn (for me at least). But for this application, I don't really want to have to deal with dissipating 6W+ of waste heat in the regulators.

Perhaps I need to revisit the SMPS for audio. I use them sometimes and it's always amazing to me seeing a little SMD part putting out a regulated 50W and it's barely warm to the touch. Years ago when they became more commonplace in audio, I remember not liking the sound of gear that had them. I always felt like they sounded flat and lifeless, as if the transients were absent. Same thing with early Class-D stuff (not liking the sound...haven't spent enough time with them to figure out why), which has no doubt improved since then too. Any specific amps you like that I should look at?

 

Low Output Impedance is handled by a big fat multiple Capacitor-Bank with very low-ESR Caps.

Not all the way down to DC. AND - that approach defeats the purpose of a current limiter. The lower the capacitor ESR, the larger the transient current peak value.

ak

 

Back in the day, ( early '80's ) when I was maintaining ~70+ Discos,
the BGW brand Commercial-Power-Amps that we used had a simple
Rocker-Switch / Circuit-Breaker on the front panel, plus, a Crowbar-Circuit.
They were heavily abused, and extremely reliable.

( out of about ~200 Amps in service,
Amplifier-Module failures were somewhere around 1 per month,
with absolutely zero Power-Supply problems for around ~3-years )
( BGW replaced the Amp-Modules at no cost )

The Power-Supply consisted of a ~50-Pound-Transformer, and 2 massive Capacitors.
Less than 0.01 THD from 20Hz to 20kHz, at roughly ~1000-Watts.
.
.
.

 

50 pounds for a 1000 W transformer sounds a bit high for just the xfmr.

I used Crown DC-300's, the ones with an opamp as the core of the amplifier circuit. Once I hammered into the staffs that they had to take 1 minute per week to vacuum off the fan intakes of everything in the racks, the failure rate dropped to about 1 per forever.

ak

 

Q3 in your two high side current sensing circuits requires a resistor in series with it’s base and another from the base to ground to keep Q3 in its linear mode and to set the ratio of maximum current to fold back current ( short circuit current).

 

Q3 in your two high side current sensing circuits requires a resistor in series with it’s base and another from the base to ground to keep Q3 in its linear mode and to set the ratio of maximum current to fold back current ( short circuit current).

Boom! That worked. I was trying to get the wrong transistor to foldback, lol.

I was able to get the short circuit power dissipation in the pass element down from 110W to 40W. Still too high for this particular example but it should work well on smaller supplies. Thank you!