Inverter

From Open Source Ecology
Jump to navigation Jump to search

Specs

Initial goal is to develop an inverter capable of taking up to 72V DC (standard forklift battery) and selectable AC output at 50/60 hz, 100v/120v/230v AC, up to perhaps 5000W. Eventually we’d like two and three phase power, perhaps expandable via modules. We intend to use a modern, solid state, high frequency design, to help keep component costs low.

Design Methodology

Outline of suggested design methodology (from Darren Vandervort, formerly of R&D of Power Electronics at Intel)

  • The first step is to come up with design goals, and define constraints. (time, budget, power, size for example.)
  • Pros and cons of different design paths are measured against goals and constraints, and a design path is chosen. (at this point the design goals are more or less locked. The design dictated which path was chosen, and if that changes the path will most likely change and any work performed after this point is wasted.)
  • Schematic capture and part sourcing. These are inherently linked and lots of compromises are made during this step to match the values you described. For example, there may be a pre-manufactured module that can save time, but cost more money in the final design.
  • Simulation and refinement. Simulation has become much more reliable in the last 6-7 years and this will save several iterations and the results are fed back into the schematic capture phase. After this phase is done the design should be considered closed.
  • Layout. This is the design of the actual PCB. For high frequency there is an extra simulation step here as well because the layout becomes a much larger part of how the circuit performs. Unless one of the constraints for this design is "as small as possible" the frequencies involved will not necessitate simulation here.
  • Prototype and testing. I would not consider the first design to be the first version. There are always things overlooked by the engineer that will have to be addressed.

Industry & Open Source Research

Inverter Papers from Universal Power Supply WIki Page

Yoonseo's Design

Inverter/Yoonseo Design

Scanned, not yet digitally schematized or otherwise analysed. Yoonseo referenced some of the above papers in his design.

Other DIY Inverters

  1. 10 kW induction heater + inverter circuit - [2]
  2. Microwave Transfomers - they propose to use modified transformers from old microwave ovens, you’d need about 10 broken micros’ and 20 N-type power MOSFET’s for a 9kw unit. [3]
  3. Open Source 100kW Electric Vehicle Controller/Inverter. This is a senior project for electrical vehicles. Its out is 3 phase. This could be of interest for the induction furnace. The design is very simple. Not sure if it was ever implemented. It is made mostly from IC's. Maybe these could be replaced to reduce some of the costs. [4]
  4. Solar Grid-Intertie Inverter - http://www.timnolan.com/index.php?page=solar-grid-intertie-inverter
  5. 500W Inverter - [5]

Scalable Commercial Inverter

The company MGE UPS seems to have merged with another one named APC, but the concept is good. This inverter is scalable up to 20kw. It looks like the inverters are designed to share some resources and are just connected in series.

(bad link) http://rep.mgeups.com/3ppdf/s4.pdf

From wikipedia on scalable 3 phase AC: To construct inverters with higher power ratings, two six-step three-phase inverters can be connected in parallel for a higher current rating or in series for a higher voltage rating. In either case, the output waveforms are phase shifted to obtain a 12-step waveform. If additional inverters are combined, an 18-step inverter is obtained with three inverters etc. Although inverters are usually combined for the purpose of achieving increased voltage or current ratings, the quality of the waveform is improved as well. 1

Arduino Based Inverters

Legacy Inverter Designs

I2K Power Inverter

This is a decently well documented inverter system, 2 kW, using transformers. See: http://192.197.62.35/staff/mcsele/i2k.htm

  • Disadvantages: about 70-80% efficient; appears that non-solid state inverters are less efficient than solid state. The author also confesses it is an older design, that uses a software generated PWM signal.
  • Commercial, solid-state inverters are 90-95% efficient (ex., [6])

Institute for Appropriate Technology Inverter

Smaller-power, lightly documented example, up to 1000W: http://www.i4at.org/lib2/inverter.htm

Retrieved from "https://wiki.opensourceecology.org/index.php?title=Inverter&oldid=67764"