Vulcan Power Supply: Difference between revisions

From Adam Colton
Jump to navigation Jump to search
No edit summary
 
Line 50: Line 50:
* Opto-Isolators: these sit between the driver logic and the gate drivers. Because of the wide current range they each need a current regulator on the diode side.
* Opto-Isolators: these sit between the driver logic and the gate drivers. Because of the wide current range they each need a current regulator on the diode side.


== Rectifier ==
I can use mosfets in reverse to create a rectifier, but I'm not sure it's worth the complication. We want to handle 15A@110V. A diode will drop 1V, so 15W, but it always uses 2 diodes of the bridge, so 30W. That's effectively a resistance of 0.066Ω. A quick look is showing N Fets are still at least 0.2Ω.
And the situation only gets worse if we start looking at something like 30A@230V.


[[Category:Vulcan|Power Supply]]
[[Category:Vulcan|Power Supply]]

Latest revision as of 13:55, 2 July 2020

The power supply is one of the real core pieces of the lab. It needs to fulfill many purposes.

Goals

  • V/A limiting output
  • Multiple isolated outputs
  • Wave Generator
  • High power 3 phase output
  • High power ESC
  • Welder driver

Design

Controller

This will need to be a powerful microcontroller. It can be configured externally (usb, wifi, etc) but the microcontroller needs to guarantee no interrupts.

High Power

There will be 3 high power outputs with a common ground. These are used to drive a 3 phase motor. One of the outputs will be rated for welding.

Isolated Lab Outputs

These are much lower power but designed for current limited and much cleaner. I may make one of them analog regulated.

Stages

Input

Can take 220V, a single 110v or two 110v which it will check are 180° out of phase. Current limit can be set.

Rectification

Can be done with a diode rectifier though I think an active rectifier with an H bridge would actually be more efficient. Output is smoothed with a cap.

Oscillator

Converts the DC to a high frequency AC tuned to the powdered core.

Controller Power

The microcontroller needs it's own power to startup the supply.

I can create a cheap one by just using a caps reactance to drop the voltage most of the way, bridge rectifier then a transistor for smoothing. But that will be connected right to main and will be really noisy. So that option might be best just for startup.

I could also use a large cap or a small battery for startup then charge it up again once the power supply is running. If it's ever out of juice I'll need another way to start it.

Or I could just give it a dedicated supply board - it won't need much.

Transformer

Using the oscillator allows us to use a cheaper powdered core transformer, possibly even a DIY version. There will be several secondary windings. Enough to compose a few into voltages within 20% of any desired voltage, plus one for high side driver and probably a dedicated 5v and 12v.

1.5 Bridge

To be able to output 3-phase, we need 3 half bridges. And because we'll be supporting a wide range of voltages, this get to be a lot of fun.

  • Driver Logic: Takes a signal for what gates to turn on and off and makes sure the high and low side of one bridge are never both on. Each bridge can be in 3 states (high, low, off). So 9 total states rounds up to 16, so 4 bits of input.
  • 10V: just need a 10V reg for the low side
  • Opto-Isolators: these sit between the driver logic and the gate drivers. Because of the wide current range they each need a current regulator on the diode side.

Rectifier

I can use mosfets in reverse to create a rectifier, but I'm not sure it's worth the complication. We want to handle 15A@110V. A diode will drop 1V, so 15W, but it always uses 2 diodes of the bridge, so 30W. That's effectively a resistance of 0.066Ω. A quick look is showing N Fets are still at least 0.2Ω.

And the situation only gets worse if we start looking at something like 30A@230V.