Inverter Log

From Open Source Ecology
Jump to: navigation, search

Contacts

Yoonseo Kang - ykang404@gmail.com

Aaron Makaruk - aaronmakaruk@gmail.com

Tim Boyd - tboydlaser@aol.com

Matthew McDonald - msm235@gmail.com

Marcin Jakubowski - opensourceecology@gmail.com

Sergio Andermann - sergio.andermann@gmail.com

Resources

Design goals, design paths, schematics, simulation results, layout, ... condensed here: Inverter

Log

October 12, 2012

I'll take on the Arduino analysis to determine how much function we can compress into the arduino. Feedback is my primary concern along these lines- I'll check both out.

I guess we're still indetermined on whether to go buck/boost converter or transformer. I prefer buck/boost but we need more research about sourcing and fabrication along both lines at this point.

As for the power handling circuit, I assume we can establish H-bridge because we are working with high power specifications.

Then there is the filter. Start off with LC lowpass I assume?

And of course, the DC-DC step-down and regulator circuits for the logic (ex. 5V, 3.3V).

And that should be all of the systems we need to work with.

Best,

Yoonseo


Hi Aaron,

I’ll be happy to help

Thanks

Sergio Andermann LEED® AP Engineer | Electrical

I remember finding that exact same webpage some months ago. With the signal that code produces, couldn't we use it with one or two modifications to produce the MOSFET drive signal? Compare the upper waveform of this http://interface.khm.de/wp-content/uploads/2009/12/ddd_title.jpg to what's shown on page 26 of this: http://www.wpi.edu/Pubs/E-project/Available/E-project-042507-092653/unrestricted/MQP_D_1_2.pdf I may be over simplifying, but couldn't we integrate the code for 50 / 60hz operation and some of the feedback into the micro-controller as well? Device protection and HMI would need to be done through a separate controller but I think it should work.

I'm beginning to see your point, a solid state DC/DC buck/boost regulator would provide greater versatility and ease of fabrication compared to a custom transformer. My concern is how complicated the design will be with the power level we're dealing with, it will almost need to be a project unto itself. Here's an example of what someone else went through with a similar converter and power level. http://www.stevehv.4hv.org/boost_converter.htm

"I assume the feedback circuit would involve an output error that goes through a PID loop to affect the amplitude factor of the sine wave amplifier circuit?" - I agree.

~Matt

Regarding the digital sine wave generator, I am open to using either an Arduino or an IC. At the relatively low frequencies we need (~60Hz) I think distortion will be a non-issue. For modularity (microcontroller board) and versatility (we can readily change the frequency) , I am leaning toward the Arduino using direct digital synthesis as explained here [ http://interface.khm.de/index.php/lab/experiments/arduino-dds-sinewave-generator/ ] with available code and <1% harmonic distortion for our frequency range. Alternatively, we could use an IC but I was unable to find ones with low enough frequency ranges, so I have sourcing concerns. Thoughts?

Regarding the pulse width modulator IC, the immediate concerns that come to mind are integration with the feedback circuit. I checked the SG3524 datasheet and it appears we would have difficulty working within the constraints of the IC- ex. discrete circuit necessary for the oscillator Rt and Ct pins in order to produce a PWM sine wave. I'm thinking alone the same lines as you- too high level and limiting.

Buck/Boost converter vs Transformer - For versatility, I started leaning towards the buck/boost converter, and given that a DIY transformer now sounds relatively difficult to fabricate, I'm leaning more towards the buck-boost converter design pathway (albeit with higher filter complexity). Additional research/information might clarify things.

I assume the feedback circuit would involve an output error that goes through a PID loop to affect the amplitude factor of the sine wave amplifier circuit?

Cheers,

Yoonseo

October 10, 2012

Email Exchanges

Aaron - no problem.

Yoonseo - I think we both agree using a digital sine wave generator is the way to go. I've seen it done with an ardiuno and i'm sure something simpler exist's on an IC.

Another possibility is to use a pulse width modulator IC. Texas Instruments designs a few, the SG3524 being one example. It's cheap and would simplify the design, however, I think something like this might be too high level and pre-packaged. The same argument could be made for the digital reference signal but I think the increase in build quality is worth it. Thoughts?

Buck/Boost versus Transformer: Component sourcing should be the same level of difficulty for both, I've managed to find sources for the electrical steel and magnet wire as well as the buck/boost converter main components. Although I'm curious about the cost's of each, I'll have to look into that more. I think fabrication is going to prove more difficult with a DIY Transformer, this would require either manually winding the bobbins or constructing a coil winder that can handle #4 to #10 wire and maybe some tooling to accurately punch out the laminations.

Have a good one, Matt

Hello gentleman

Since I am the Johnny come lately to this project I am just learning what is going on. All of these ideas seem excellent to me. I will be observing for now.


Regards

Tim Boyd tboydlaser@aol.com

Sergio,

We'd like to invite you to join this email discussion about building the open source power inverter with Open Source Ecology? If you reply to this thread, please reply to all.

Everyone, this was our last correspondence with Sergio:

Hi Aaron, Vann,


Thanks so much for your email, sounds like a fun project.


I have limited experience in this type power electronics; while I’m familiar with UPS systems which, in principal work similar to aninverter like the one you describe I wouldn’t be able to build one myself without spending some time doing research. The information in the wiki page may not be sufficient.

This is an ambitious project and may require more than just one person to complete. The reality though is that, these types of devices require several circuits within, some of which I can design faster than others.


I would like to help, but I wouldn’t be able to do this on my own. Also, I wouldn’t be able to start before September.

Please let me know if this is useful.

Kind regards

Sergio Andermann LEED® AP

Engineer | Electrical

Arup

560 Mission Street Suite 700 San Francisco CA 94105

t +1 415 957 9445 d +1 415 946 1693

f +1 415 957 9096 c +1 720 883 2114

www.arup.com

Regarding the digital sine wave generator, I am open to using either an Arduino or an IC. At the relatively low frequencies we need (~60Hz) I think distortion will be a non-issue. For modularity (microcontroller board) and versatility (we can readily change the frequency) , I am leaning toward the Arduino using direct digital synthesis as explained here [ http://interface.khm.de/index.php/lab/experiments/arduino-dds-sinewave-generator/ ] with available code and <1% harmonic distortion for our frequency range. Alternatively, we could use an IC but I was unable to find ones with low enough frequency ranges, so I have sourcing concerns. Thoughts?

Regarding the pulse width modulator IC, the immediate concerns that come to mind are integration with the feedback circuit. I checked the SG3524 datasheet and it appears we would have difficulty working within the constraints of the IC- ex. discrete circuit necessary for the oscillator Rt and Ct pins in order to produce a PWM sine wave. I'm thinking alone the same lines as you- too high level and limiting.

Buck/Boost converter vs Transformer - For versatility, I started leaning towards the buck/boost converter, and given that a DIY transformer now sounds relatively difficult to fabricate, I'm leaning more towards the buck-boost converter design pathway (albeit with higher filter complexity). Additional research/information might clarify things.

I assume the feedback circuit would involve an output error that goes through a PID loop to affect the amplitude factor of the sine wave amplifier circuit?

Cheers,

Yoonseo


I remember finding that exact same webpage some months ago. With the signal that code produces, couldn't we use it with one or two modifications to produce the MOSFET drive signal? Compare the upper waveform of this http://interface.khm.de/wp-content/uploads/2009/12/ddd_title.jpg to what's shown on page 26 of this: http://www.wpi.edu/Pubs/E-project/Available/E-project-042507-092653/unrestricted/MQP_D_1_2.pdf I may be over simplifying, but couldn't we integrate the code for 50 / 60hz operation and some of the feedback into the micro-controller as well? Device protection and HMI would need to be done through a separate controller but I think it should work.

I'm beginning to see your point, a solid state DC/DC buck/boost regulator would provide greater versatility and ease of fabrication compared to a custom transformer. My concern is how complicated the design will be with the power level we're dealing with, it will almost need to be a project unto itself. Here's an example of what someone else went through with a similar converter and power level. http://www.stevehv.4hv.org/boost_converter.htm

"I assume the feedback circuit would involve an output error that goes through a PID loop to affect the amplitude factor of the sine wave amplifier circuit?" - I agree.


~Matt

October 8, 2012

Email Exchanges

Good Evening Everyone,

I have a few questions regarding the inverter oscillator design and high voltage DC regulation.

Oscillator - I'm looking through Yoonseo's design [ http://opensourceecology.org/wiki/Inverter/Yoonseo_Design ] is the use of the Bubba oscillator set in stone? I'm reading through the WPI project report and it appears they had a good deal of trouble with it's operation. Would it be possible to go with a fully digital reference sine wave? Has there been any more development on the reference wave feedback control circuit? http://opensourceecology.org/wiki/File:06_Feedback_Circuit.jpg

DC regulation - With the need to accept a wide range of source voltages and output a consistent 120 / 220 VAC signal, has there been any thought into the need for a high powered buck / boost converter to acquire the needed 170 / 311 VDC? Another possibility is to use a multi-tap output transformer, I would be in favor of this route for the following reasons: 

1. Reduces the complexity of the output filter
2. Can be designed to replace the high powered buck / boost converter and therefore improving reliability by simplifying the over all circuit.
3. In the case of a shorted MOSFET or some other fault, will eliminate the possibility of high voltage DC present on the output signal
4. Output transformer can be be manufactured given a well made sheet metal press, various hardware and proper research

See this instructables link for a reference on a transformer coupled design: http://www.instructables.com/id/250-to-5000-watts-PWM-DCAC-220V-Power-Inverter/

Also, in regards to the failing inverters already in place at FeF, has any effort been put into investigating the possibility of repair? If these are preventing collaborative production runs of the GVCS it may be faster to fix them in the mean time until this project is completed. Aaron, if you're open to the idea, would it be possible to send one of the failed units to Denver so I could have a look at one?

Have a good night, Matt

Hello Everyone


Does the inverter need to be grid tie now or in the future?

Tim Boyd tboydlaser@aol.com

The circuit for generating the reference sine wave is not set in stone. I prefer a fully digital reference sine wave in preparation for versatile operation. If we pursue the digital generation I assume we would just source the relevant integrated circuit?

No development on the feedback circuit to date.

Buck/boost converter considered, but not developed to date. Between the buck/boost converter and the multi-tap transformer, I am most concerned about ease of fabrication/sourcing and would like more information regarding that factor.

Best,

Yoonseo

Marcin asked to join this thread so he'll be in on all further discussion - Tim - being grid tied is not a priority right now.

Aaron

Introductory email between all project participants

  • To be done - 10/08/12
  • Introduce everyone
  • Let's discuss general inverter topologies to suit OSE Specifications
  • We will start an email chain discussing general inverter topologies (ex. full bridge, half bridge)
  • Let participants know that the email chain will be uploaded to the Inverter Log in a subsection.
  • Needs:
  • Input 72V DC
  • Output 120V AC at 60Hz pure sine wave with power rating of 5kW or 10kW [consider stackability].
  • OSE Specifications
  • These things are preferable but are less of a priority than function, reliability, and ease of fabrication.
  • Input should be flexible enough to accommodate a wide range of input voltages.
  • The output should be flexible enough to vary (ex. both 120 and 240 V AC)
  • Modular design - Example
  • Use of open source design software


  • Email was sent
Retrieved from "https://wiki.opensourceecology.org/index.php?title=Inverter_Log&oldid=82324"