Facebook
Google
GitHub
Linkedin
The problem:
I have a switching power supply (current source) used to drive LEDs that I'm having trouble with. I have a 5Amp rated relay on the output to turn on and off the LEDs by breaking the circuit. Under steady state operation, the output current is user variable between 100mA and 3.5Amps. Under some circumstances, when the relay is closed, the contact will weld. I'm fairly confident my problem is the inrush current on the output (outrush current? haha).
What I know so far:
When the relay is opened to disconnect the output, the power supply will rise to 36V, where it switches to voltage regulation mode as a safety feature. The supply has an output capacitance of 85uF. In the worst case, The user connects a single LED to the output, selects 3.5Amps, and then the output relay is closed. If the LED drops approximately 3V. That means i have 33Vx85uF=2.8mC of charge shorted in the time period of of the relay set time. I found it hard to track down set times for a relay, but lets ballpark 100microseconds for contact bounce to stop and connection stabilize, this gives an inrush of 28Amps.
It takes about 2miliseconds for the output to fully stabilize regulation after the relay is closed, but it is in a safe region within 300us.
I tried to measure the inrush current in the worst case scenario using a scope across a 1ohm resistor in series with the output (which would of course lower the inrush current) and got peak values of ~40Amps for the first 10ns decreasing exponentially down to safe values in ~300uS. Obviously this is not great for the LED's nor is it acceptable for the Relay's 5Amp max.
How to solve the issue:
I think, when operating the relay in the future revisions, I'll run a signal to the shutdown pin on the power supply IC so it goes into shutdown mode when the relay is opened (The circuit must be physically broken as a design parameter). When coming out of shutdown mode the IC's built in soft start will engage (this takes 20mS, longer than the relay turn on time). I cannot reprint the boards for a few months. Until then, is there another way of doing this? I know there are specifically designed thermistors you put in series with the load (LEDs) for this reason, but they would produce a lot of heat and I'm not all that familiar with their specifics. Are there other tips/tricks i can try to reduce the inrush below 5Amps. I've simulated the problem in LT spice and the resulting output waveform is attached.
Any help would be greatly appreciated. I can provide more details if requested!
I have a switching power supply (current source) used to drive LEDs that I'm having trouble with. I have a 5Amp rated relay on the output to turn on and off the LEDs by breaking the circuit. Under steady state operation, the output current is user variable between 100mA and 3.5Amps. Under some circumstances, when the relay is closed, the contact will weld. I'm fairly confident my problem is the inrush current on the output (outrush current? haha).
What I know so far:
When the relay is opened to disconnect the output, the power supply will rise to 36V, where it switches to voltage regulation mode as a safety feature. The supply has an output capacitance of 85uF. In the worst case, The user connects a single LED to the output, selects 3.5Amps, and then the output relay is closed. If the LED drops approximately 3V. That means i have 33Vx85uF=2.8mC of charge shorted in the time period of of the relay set time. I found it hard to track down set times for a relay, but lets ballpark 100microseconds for contact bounce to stop and connection stabilize, this gives an inrush of 28Amps.
It takes about 2miliseconds for the output to fully stabilize regulation after the relay is closed, but it is in a safe region within 300us.
I tried to measure the inrush current in the worst case scenario using a scope across a 1ohm resistor in series with the output (which would of course lower the inrush current) and got peak values of ~40Amps for the first 10ns decreasing exponentially down to safe values in ~300uS. Obviously this is not great for the LED's nor is it acceptable for the Relay's 5Amp max.
How to solve the issue:
I think, when operating the relay in the future revisions, I'll run a signal to the shutdown pin on the power supply IC so it goes into shutdown mode when the relay is opened (The circuit must be physically broken as a design parameter). When coming out of shutdown mode the IC's built in soft start will engage (this takes 20mS, longer than the relay turn on time). I cannot reprint the boards for a few months. Until then, is there another way of doing this? I know there are specifically designed thermistors you put in series with the load (LEDs) for this reason, but they would produce a lot of heat and I'm not all that familiar with their specifics. Are there other tips/tricks i can try to reduce the inrush below 5Amps. I've simulated the problem in LT spice and the resulting output waveform is attached.
Any help would be greatly appreciated. I can provide more details if requested!
