The Solar Power Expert Blog

Manufacturability of the New Free Charge Controller Design

by Chris - September 11th, 2011.
Filed under: Charge Controller. Tagged as: , , .

My proposal for the DIY Dual Voltage Converter design for the Free Charge Controller project generated some good feedback. This post is a copy of the one of the replies and focuses on the manufacturability of the two designs and some of the different aspects of the two designs.

On Wed, Sep 7, 2011 at 10:35 PM, Raj Nambiar wrote:

Dear Chris,

I have been in the group for quite some time. Let me introduce myself. I own a a small company that sell technology in certain field to manufacturers. We work in UPS, charge controllers, displays, SNMP agent for UPS, real-time SOFTWARE etc. We are also into remote monitoring, factory automation etc. In PC software field, we sell around 10 packaged applications. My site is Not all my products are on my site.

I have found myself excited on seeing your effort, and have always been in admiration of your work ever since. It looks quite like a brilliant product, considering that you are doing it in your free time, and can compete with commercial products except on one or two features. I haven’t built it but am planning to build one just to see the performance.

I can build PCBs, get magentics for you and test them and send you details. Advise me whether it would involve conflict of interest since we are already involved in selling charger technology. We work with Silabs / AVR controllers, generally. Most of ouw works are not original. We take ideas from white papers, sample projects from chip manufacturers, and some times even reverse engineer existing products.
Some minor points that I wanted to mention are below. Some of them could be stupid, but it they could be discussed and corrected, it could be worth the time, I guess.

  1. In the first version, the except the diode losses, the product looks as good as any commercial one. I haven’t gone through the data on losses in inductor, but when I build one, I will.

  2. Using the same charger for voltages upto 120 or 160 Volts battery system would have been great. I had read the thread pertaining to that. It would have been great if I could set the current and voltage through a PC application. If I get time, I will try to work on it and send you the details.

  3. A specific board, in my opinion, is better than a arduino, since we can closely control the parameters and cost.

  4. Even though chasing the commercial products is not the best way forward, chasing the features available in commercial products should be given some consideration. New controllers in the market are cheaper and sport features like auto-boost mode. Many integrators tend to add extra panels in series to avoid dropping of output voltage. An auto-boost mode, where the controller goes into boost when the voltage dips does bring more ROI, when you want to squeeze out more watts from the same panel and the same sun.

  5. The newly proposed boost topology with two diodes and the transformer may have a marginally less efficiency than a buck boost topology. It may be worth considering whether keeping the buck part same and switching on the boost block when the voltage drops, might turn out to be cheaper and more efficient, even though, marginally.

Regards, Raj Nambiar

I responded on Sunday 9/11/11 with:


Thank you for the compliments, and congratulations on building such a successful business. There is no conflict if you’d like to continue development with the v3 board. The only legal requirement for using the design (for commercial sale or otherwise) is attribution. That just means that you need to reference in the source code and you need to put a ‘’ somewhere on the PCB in the silkscreen or copper.

Let me address your discussion points below:

  1. There is an extra diode in the design. This was intended as a bit of ‘dummy proofing’ so that people couldn’t backfeed current into the board. It’s up to you if you want to leave it in or take it out. It does absorb a bit of energy that could otherwise go to the load.

  2. Working with those higher voltages gets tricky. Typically any buck or boost circuit is going to put the full system voltage across the MOSFET when the circuit is ‘off’ and MOSFETs are not typically designed for voltages much above 20 volts. This is the really difficult part of designing charge controllers for high voltages. As I learned the hard way, you also have to be carefull about the voltage tolerances of any current sensors you put in too.

  3. You are correct. From a manufacturer’s standpoint, it is definitely better to integrate the microcontroller into the board. That was exactly my line of thought when designing the board.

    However, I personally consider my design to be a failure. Not because of any part of the design, but because no one in the community has built one despite releasing the design about a year ago. By making a header for people to plug-in an external microcontroller board – like the Arduino, Maple, etc – people can use their favorite microcontroller to control the board.

    This focus on customization and ease of use is the focal point of my new design proposal to the community.

  4. I couldn’t agree with you more on this point. In fact, I said nearly the same thing in the podcast with Open Source Hardware Junkies. A charge controller is a charge controller is a charge controller. Manufacturers should compete on features required by specific use-cases (such as needing an auto-boost).

    My hope with the new software-heavy design is that people can easily configure the hardware to enable features (such as auto-boost) and control the precision and function through software.

    I personally dislike the industry wide adoption of the requirement that solar panel voltages be higher than the system voltages. i.e. an 18 volt panel for a 12 volt system. This requires the charge controller to be a ‘buck’ converter. If panels used fewer cells in series – so that they had a low voltage, high current output – it would reduce detrimental shading effects and improve performance durring low-lighting conditions. However, this requires a ‘boost’ converter to pull power from that type of solar panel configuration.

    In the end though, I think the decision needs to be with the consumer and the market. I assume there are large forces (which I may not fully understand) which have forced the industry to adopt the standard that they have today.

  5. I think you’re speaking of the v3 design here and not the newly proposed design. I agree that a buck/boost topology would be more flexible and possibly more efficient. Like I said, the extra diode is for ‘dummy proofing’ to a certain extent.

    I had actually never built a buck (or boost) converter before I designed this board. I tried to follow the AVR app-note and keep things simple. That’s why I never went after a buck/boost topology in the first place and just went with buck instead.

    The new design is much more a ‘controlled transformer’ than a buck/boost design. The primary difference is that in a buck/boost circuit the inductor value is fixed and the inductor is driven at a high frequency. In this new design, the circuits ability to buck or boost voltage will come from how the user winds the torroid. Also, the transformer will be driven at closer to 60Hz as opposed to frequencies in the Khz for a buck or boost circuit. These severely relax the design and software requirements which I think is important for ‘hobby’ use.

Leave a Reply

Your email address will not be published. Required fields are marked *