Disable or trick pulse-skipping/skip-cycle/eco-mode

[b_force]

I have here a power supply with a NCP1239.

I am pretty happy with the overall design around this chip, except for the fact that it goes into some kind of "eco-mode" with light loads.
Unfortunately this is causing all kinds of noise issues in my design (which has a dynamic load).
Before I go for a different power supply, I was wondering if this mode can be tricked somehow?
I am fine trading in some efficiency for this obviously.

I have been searching and looking around for a bit, but can't really find any satisfying solution so far.
(for the record, I am skilled in working with main voltages and designing similar designs myself, so I am only just asking for a solution to disable this mode)

[Wimberleytech]

I have here a power supply with a NCP1239.

I am pretty happy with the overall design around this chip, except for the fact that it goes into some kind of "eco-mode" with light loads.
Unfortunately this is causing all kinds of noise issues in my design (which has a dynamic load).
Before I go for a different power supply, I was wondering if this mode can be tricked somehow?
I am fine trading in some efficiency for this obviously.

I have been searching and looking around for a bit, but can't really find any satisfying solution so far.
(for the record, I am skilled in working with main voltages and designing similar designs myself, so I am only just asking for a solution to disable this mode)
If you are willing to sacrifice some efficiency, just add a fixed load.  Not so pretty, but it will address the problem.  Most likely you have thought of this and have reason not to do it.

[b_force]

I have here a power supply with a NCP1239.

I am pretty happy with the overall design around this chip, except for the fact that it goes into some kind of "eco-mode" with light loads.
Unfortunately this is causing all kinds of noise issues in my design (which has a dynamic load).
Before I go for a different power supply, I was wondering if this mode can be tricked somehow?
I am fine trading in some efficiency for this obviously.

I have been searching and looking around for a bit, but can't really find any satisfying solution so far.
(for the record, I am skilled in working with main voltages and designing similar designs myself, so I am only just asking for a solution to disable this mode)

If you are willing to sacrifice some efficiency, just add a fixed load.  Not so pretty, but it will address the problem.  Most likely you have thought of this and have reason not to do it.

Obviously I have, the problem is that this fixed load has to be pretty big to get out of this eco mode.
Wasting around 1A around 24V.
A little bit to big to my liking, so I was wondering if there other ways to do this?

[Wimberleytech]

Going into skip mode below 1A doesn't feel right.  1A is not "light load"

[b_force]

Going into skip mode below 1A doesn't feel right.  1A is not "light load"

Well, around 1A it goes into a lower frequency, around 20kHz which is causing a lot of noise issues.
Going down to around 500mA and this frequency is even lower.
With no load it's completely unusable, go to around 4-5kHz or thereabouts.

This power supply can deliver around 150W, so I guess 24W is considered a light load?

[b_force]

No other ideas?

Isn't there a way to tick the feedback pin or something?

[mikerj]

If you try to prevent the voltage on the FB pin falling low enough to reach the pulse skipping mode, then the supply will become unregulated at low loads and the output voltage will increase considerably.

I suppose you could put an active load across the output that prevents the FB voltage dropping by applying additional load (instead of a a fixed load resistor), but this seems like a lot of additional complexity.  Starting with a fixed frequency controller would be a better plan.

[b_force]

Starting with a fixed frequency controller would be a better plan.

I know, it's only very difficult to find them these days, since most of these controllers have some kind of green/eco skipping mode built it

[David Hess]

Wouldn't increasing the inductance delay the point where pulse skipping mode is active?

[T3sl4co1l]

Keep FB above Vskip (0.8V).

Since FB has an internal pullup, a series resistor between the optotransistor (assuming the application circuit) and FB will do this.

The R_UP doesn't seem to be rated, so YMMV.

Tim

[mikerj]

Keep FB above Vskip (0.8V).

Since FB has an internal pullup, a series resistor between the optotransistor (assuming the application circuit) and FB will do this.

The R_UP doesn't seem to be rated, so YMMV.

Tim

But this will cause loss of regulation, unless performed indirectly by adding load.   Additionally the OP want's to avoid the drop in frequency prior to pulse skipping, so FB voltage would have to be limited somewhat higher than 0.8v, as much as 1.9v to avoid any change in switching frequency.

[Miyuki]

If it is not a consumer product and thus dont need to pass green/eco requirement

Just use simple old UC38xx or similar driver with steady regulation in all range

I dont see any reason to use modern eco controller in situations where it cause instability and other problems when you dont need that low standby power consumption
And that NCP1239 dont look like have any fancy resonant mode to offer benefits in efficiency at full load so it have no benefits for your use

[T3sl4co1l]

But this will cause loss of regulation, unless performed indirectly by adding load.   Additionally the OP want's to avoid the drop in frequency prior to pulse skipping, so FB voltage would have to be limited somewhat higher than 0.8v, as much as 1.9v to avoid any change in switching frequency.

They asked, that's what the block diagram says. ¯\_(ツ)_/¯ 

If that solution requires a bleeder resistor, then change the transformer ratio.  Or, check that the transformer's leakage is what it needs to be (which, if light-load regulation is the problem, more leakage might actually be helpful?).

If it is not a consumer product and thus dont need to pass green/eco requirement
...

All this ^^^

Tim

[b_force]

I dont see any reason to use modern eco controller in situations where it cause instability and other problems when you dont need that low standby power consumption
And that NCP1239 dont look like have any fancy resonant mode to offer benefits in efficiency at full load so it have no benefits for your use

To be very honest, I am investigating if I can take an off the shelf PSU.
The reason is that nowadays they are cheap as chips and perform pretty reasonable.
As far as I know, these ECO modes are only to make the people at the government happy that idle power isn't than a certain value.
In our case not a huge problem since the device is being powered of by a hard switch.

But like I said, the biggest problem it has that it goes down to much lower frequencies, which is causing all kings of noise problems.

Keep FB above Vskip (0.8V).

Since FB has an internal pullup, a series resistor between the optotransistor (assuming the application circuit) and FB will do this.

The R_UP doesn't seem to be rated, so YMMV.

Tim

The issue is that I need the voltage regulation, or at least that it's not going through the roof.