Switch Mode PSU Voltage Spikes

No this is not normal,

Halogen lights are resistive load and little inductive load is fed back to the suply.

I suspect the regulation system of the PSU, it is slow to respond to a sudden variation of the load,

try this simple test, when you switch on the lights, does the output drop? if this is the case than you have a slow regulator, refer to the manufacturer....

if not, try a free wheel diode on the PSU output Good luck

Some switching power supplies require a minimum load current to attain stable operation (and low noise). When lightly loaded or with no load (open circuit), the output voltage may vary out of the expected tolerance range and large switching spikes may be present.

Please consult the power supply manufacturers specs regarding the proper loading values. If your minimum load (the controller?) falls into the recommended range given by the manufacturer, then the spike noise indicates a power supply malfunction (or a really poor design).

I liked your post, I am sure you either hit the bullseye or very close.

Maybe the user should add a resristor across the output to take say 1 amp at all times. Suitably dimensioned/cooled of course......

A lower current value may do it as well, so some experimentation is really needed.....

Maybe the resistor could be switched in a split second before the lamps are switched off......

But the simplest method would be to switch off the power supply at the 415 volt supply area, and not just the lamps in the 24 volt area of course!!!

Andy, I'm just a knuckle dragging mechanical grunt (ok, I did take some electronics courses in my misspent youth) but wouldn't a fairly large capacitor across the 24VDC supply in parallel with a very high value bleed resistor go a long way towards stabilizing the supply voltage?

You could go one better by putting a zener (say a 1N5360) or TVS in parallel with the cap too.

Of course.

That is a better technical approach as well as long as the Zener can take the power pulse without damage.....though I have to say that not all SMPS like or can handle, and extra cap across the output.....but that could be easily tested for!!

In the case in question, I was just trying to achieve the simplest possible method that more "viewers" would understand.....

Actually upon further review a 1N5361 might be a better choice. And a TVS in parallel, say one of these, might be a good idea as well in order to help protect the zener from overload.

Great little things aren't they!!!

Adding a capacitor may help, but it could also make the problem worse. More often than not, adding capacitance does help with linear supplies.

Feedback control for switching supplies is a bit more complicated because the system parameters (poles and zeros) can change when the inductor current goes from continuous mode (normal loading) to discontinuous mode (light loading or open circuit). This alters the behaviour of the feedback loop, potentially making supply operation unstable. Adding capacitance doesn't eliminate the change in inductor current modes.

Although it somewhat defeats the purpose of using a high efficiency supply, the suggestion of adding a fixed resistor load that draws just enough current to stabilize the output is one (easy) solution.

Simple... Use a seperate PS for the controler.

Since the spike only appears on switching OFF the lamps, it will appear that Le Noble has given the right answer: PSU not responding fast enough---> design and also I will add maybe too big for the controller alone.The design may be relying on fast depletion of excessive stored energy at the output capacitor within the PSU (switching type). Since the remaining load might be very low, the spike will occur just long enough to decay to the regulated value.

The best way is either not to switch off the lamps in one go but in steps to reduce the stored energy, OR to keep them on as long as controller is running, OR get a separate PSU for the controller to suit it. Anything else will either keep the spike coming on or interfere with the PSU performance.

Isolate the controller supply.

Play around with the values depending on the current; the minimum voltage required by the controller, and, the width of the spike.

That is a good way to go in parallel with what some others (Rorschach, myself, mj1962853and at least one other) have mentioned to do.....

Many thanks for all your comments.

I have been looking at connecting a DVS 24 surge arrestor module from Roxburgh to the PSU output. The electrical drawing for the module appears to contain a couple of varistors which can be connected across the output and from +ve to ground.

The data sheet shows they are continually rated at 38V D.C. but capable of absorbing transients at the rate of 8.8joules /ms.

Will let you know if this works.

If not then I will be going down the track of fitting a separate PSU's for the lights.

I very much appreciate those who contributed to this thread.

Best regards Allan