In my previous post, I shared a design for a 120VAC to 50VDC unregulated power supply that can be used to power the solenoids in a pinball game. There was a flaw in the design though that results in the tripping of circuit breakers. Inrush.
Depending on where you turn on the power supply in the AC power cycle you can get an extremely large current load from two sources. The first is the toroidal transformer, and the second is from the bulk capacitors charging up.
There are a few ways to handle inrush, ranging from cheap and simple to expensive and complex. The simplest and cheapest is to use a negative temperature coefficient (NTC) thermistor. These devices have a small resistance (10 ohms for our case) at room temperature, but as current flows through it will heat up and the resistance will drop to a very low value (few milliohms).
An NTC thermistor is a decent choice for lower power supplies where you don’t need to worry about the power loss those few milliohms give you. In higher current supplies, however, even those few milliohms will start to add up in high losses. Another large negative with them is they take awhile to cooldown after you power off (minutes) during which if the supply is turned back on there will not be any inrush protection. A third problem that could present itself is in a pinball game, the load on the 50V supply is very small when the game is not actively being played and during that time the NTC thermistor could cool off causing a problem when a game is started.
An alternative that I chose to go with involves using a fixed resistor, 10 ohms, put into parallel with a relay. The relay’s coil is connected to the 50V output. How this works is while the capacitors are charging, the current will be going through the 10 ohm resistor which will limit the amount of current. Then, once the capacitors charge up to about 36V the relay will switch on bypassing the resistor from the circuit. While the relay is rated for 48V on the coil, if you read the datasheet for these types of relays you’ll see their guaranteed closure voltage is in the 33-36V range. They’ll then stay closed until around 5V when they’ll open.
Adding this to my design as shown in the schematic solved my inrush problem and it no longer trips breakers. There is still a risk I’ll have to address in the future. If the relay fails, that 10 ohm resistor is likely going to burn. It’s only rated for 16W and with 120VAC RMS going through it that’s far higher than 16W. I will probably look for a double through relay in the final design so that I can have the microprocessor monitor the relay and prevent the game from being played if the relay is not closed.
Inrush Limiting Schematic
In the schematic you’ll see two transformers, in truth it’s the same transformer I just couldn’t find a dual primary/dual secondary symbol readily available in KiCAD. The toroidal transformer I have has two primary windings and actually four secondary windings. Two of the secondary windings are for 35V and the other two are for 18V and 12V. When connected in parallel, as shown, it is like connecting a battery in parallel. You can add together their current capacity. On the primary side, if connected in series the transformer can be used in 220V countries while keeping the 35V secondaries in parallel.