TOPswitch-GX soft-start time - is it adjustable?

Posted by: davjporter on 02/03/2015

I have a TOP249Y, universal input design providing 27-29VDC @2.2A from 100-265VAC.
It can startup under full load over the full input voltage range under normal conditions, but my application is not "normal". The DC output is battery backed-up.
If the batteries are not connected, the start-up under full load is fine over the entire AC-input voltage range. If the load is powered from the batteries and the batteries have discharged to, say 22V, the switcher fails to start if the AC-input voltage is below about 115V. (The switcher does start properly at higher AC-input voltages.)

I should mention that the battery has a blocking diode to prevent the switcher output from "charging" it.

Troubleshooting tells me that the feedback-loop does not close. I believe this is because the secondary has 22V from the start and there is no load until the output reaches this point, at which point the full load "turns-on" abruptly.
I am thinking that "softening" the start-up more than usual might resolve the issue. (So might increasing the current limit, but I would prefer not to do that.)

Can the soft-start be stretched with simple component changes?
If so, what is the safe range?

Thanks for your help.

Comments

Hi,
Thanks for sharing the information.
Looks like you found the root cause of the issue. (load kicks in when the output voltage reaches certain level). Do you using the opto-coupler + TL431 for your feedback in your design? If yes, you could add external soft start circuits with adding one diode, one resistor and one diode on the feedback circuit. Pleaser refer to the design for detail:
D4/C10/R7 are the external components for soft start:
http://ac-dc.power.com/design-support/reference-designs/design-examples…

Hope this is helpful

Hi Terry,
Thanks for your "feedback".
Yes, I am using an opto-coupler + TL431 for my feedback. My feedback circuit is not quite as simple as the one in the EP-34 schematic, but perhaps is close enough.
I did not try adding D4/C10/R7. I have some reservations about whether they will work in my case, as the output voltage is already very close to "normal output voltage". (As I indicated in the original post, there is a battery holding-up the output voltage to within a vew volts of the regulated value.)
I did, however, find another way to slow the soft-start process. I am sourcing a small current (about 700uA) to the control pin by adding a resistor across the opto-coupler output. I realize that this is a crude current source, since it tracks my bias voltage, but the bias voltage is very stable under normal operation, and stable enough during this start-up. As I am sure you know, the bias voltage typicially tracks the output voltage, so this "current source" remains fixed over the entire range of AC-in (100-265V). :-)

Hi,
If your power supply is designed to work at a certain voltage and you do not allow it to reach that voltage by applying a load such as a discharged battery which is a fairly stiff load, then the feedback loop will not close and the power supply will do something that is described in the datasheet as "auto-restart". Basically the feedback loop does not close as the power supply is "power limited or overloaded" due to a large load (discharged battery).

If you provide us a copy of your schematic, we may be able to recommend a work around but it appears what the controller is doing is what is expected of it, if the feedback loop fails to close.

You would need to modify the circuit to use a CVCC feedback circuit that closes the loop based on either the output voltage or the output load current feedback. Please see the link below for a design example:

http://ac-dc.power.com/design-support/reference-designs/design-examples…
Best Regards

As stated in the original post, the battery is DIODE-ORed with the switcher output. Once the switcher output voltage reaches the battery terminal voltage (give-or-take a diode-drop) it will begin to power the load, but the diode prevents it from supplying current to the battery.

I believe the problem is that the feedback loop is OPEN with NO LOAD too long, causing the switcher to shutdown and restart. Because this is the case, a CVCC feedback circuit won't fix the problem.

Everything else the same, if the AC-in voltage is increased, the supply starts fine. This is becasue the output voltage climbs faster, reaches the desired voltage and provides the required feedback in time. If only I could get the switcher to wait just a little longer before giving up...

No one has commented on the original questions...
1) Can the soft-start be stretched with simple component changes?
2) If so, what is the safe range?
I have (in my previous reply) covered one way to stretch this timing, but with no comments on what I did.
I would like to know what PI thinks of this method, and if they have any other recommendations regarding how this could (or should) be done.

It would be best for you to attach a schematic of your design showing the blocking diode and the battery so we will be able to understand better and also perhaps markup the schematic to suggest appropriate changes.

The circuit proposed by Terry(RCD circuit) is called the "soft-finish" circuit as it slows the rise of output by closing the loop artificially as the output rises until the point where the capacitor charges to the full output voltage and the diode decouples the capacitor and resistor from the optocoupler circuit.

As far as stretching is concerned, there is no specific issue with making the timing long except that the bias winding filtre capacitor must be able to charge such that the optocoupler collector is being fed from a voltage high enough so that there is current flow from the optocoupler to the TOPSwitch within a time that is shorted than the time it would take for the control pin to discharge below 4.8V.