Protecting circuit from power supply spikes

Thread Starter

Netwaves

Joined Feb 9, 2015
44
I have two power supplies that supply 3.3VDC at up to 2A, 5.0VDC at up to 3A, and 12VDC at up to 5A. Unfortunately, I have now lost two Raspberry Pi and a host of chips in my circuits before realizing that my expensive power supplies are spiking up to 12V when first turned on. Can I use Zener diodes to protect my circuits and equipment from these spikes? If so, what value/part# for each range? Yes, I can buy new power supplies but I really don't want to loose my investment in the ones I have. Also, any suggestions on in-line fuses? Any help would be greatly appreciated. TIA
 
Last edited:

MikeML

Joined Oct 2, 2009
5,444
So that the rest of us can avoid your mistake, post the model number, url, picture of the supplies is question.

Moding the supplies to eliminate the start-up spike may be possible. Can you find a schematic of the supplies?
 

Thread Starter

Netwaves

Joined Feb 9, 2015
44
So that the rest of us can avoid your mistake, post the model number, url, picture of the supplies is question.

Moding the supplies to eliminate the start-up spike may be possible. Can you find a schematic of the supplies?
I'm using three of these connected to a 15VDC 10A switching power supply. One is set for 3.3VDC, one for 5.0VDC, and one for 12VDC:

DROK® LM2596 Adjustable DC Voltage Regulator Digital Converters Module with Screw, 4.0-40V to 1.25-37V DC DC Converter 36V to 24V to 12V to 5V Volt Switching Stabilizers Car/Automotive Battery Step Down Buck Variable Volt Power Supply & Red LED Voltmeter Display

I could not find a reliable schematic for the unit.
 

Attachments

MikeML

Joined Oct 2, 2009
5,444
These are poorly-designed switchers with crappy transient response. I guess you get what you pay for.

The only way to fix them would be to trace the schematic, and then redesign the feed-back loop filter, which determines the overshoot on start-up and transient response if a sudden step load is applied to the supply. Big job that should have been done by the clueless designer that came up with that product.
 

DickCappels

Joined Aug 21, 2008
7,244
You might be able to stop the destruction with zener diodes, but I hope you don't plan to test the idea using your remaining Pi, but be careful to use a zener whose voltage during the surge is low enough to protect your circuit yet high enough when not experiencing the surge to not load down the power supply and/or melt itself. That might be a difficult nut to crack.

Tending to favor MikeLM's way of seeing this, I think that if I were in your shoes (and I am not) I would toss the bad supplies before they get a chance to kill something else.
 

AnalogKid

Joined Aug 1, 2013
9,151
The supplies are non-isolated buck regulators based on National Semiconductor's 7xxx series of "Simple Switcher" parts. These devices rarely deviate from the circuits in the data sheet and app notes for the controller. The 5-pin part top center is the culprit. Tell us the part number and you might get lucky - it might have a soft-start input pin.

I agree with DC (doesn't everyone?), low voltage zeners are difficult to size properly for this problem. They probably will work if you can get the right ones.

ak
 

Thread Starter

Netwaves

Joined Feb 9, 2015
44
You might be able to stop the destruction with zener diodes, but I hope you don't plan to test the idea using your remaining Pi, but be careful to use a zener whose voltage during the surge is low enough to protect your circuit yet high enough when not experiencing the surge to not load down the power supply and/or melt itself. That might be a difficult nut to crack.

Tending to favor MikeLM's way of seeing this, I think that if I were in your shoes (and I am not) I would toss the bad supplies before they get a chance to kill something else.
Looks like I'm in the market for some new supplies then. :)
 

crutschow

Joined Mar 14, 2008
26,961
An alternate way is to add a delay at each output consisting of a series P-MOSFET and delay circuit.
Do you know how long it spikes to 12V at startup?
 

Thread Starter

Netwaves

Joined Feb 9, 2015
44
An alternate way is to add a delay at each output consisting of a series P-MOSFET and delay circuit.
Do you know how long it spikes to 12V at startup?
The spike is around 1 second. Adding such a circuit is probably more than I'm willing to do at this point. Adding a single Zener for each would have been ideal and worth the effort. I will probably just return these and purchase something different.
 

bill nahd

Joined May 15, 2015
6
Putting a resistor in series with your power supply output does not improve the characteristics of the supply voltage.
power-on delay with fet or relay is not such a bad idea... Is it? It is cheap and does not affect the characteristics of the psu voltage.
 
Top