[Coulomb]

Elcon/TC Charger Firmware: Discussion

This thread is intended to discuss Elcon/TC charger firmware.

Please post all facts to the Elcon/TC Charger Firmware: Facts thread.

I know it's probably a hopeless task, but I'd like to keep most of the discussion in this thread, and keep the facts thread relatively free of "can I have this feature?" or other discussion, so it's easier to find the "meat" in the facts thread. We'll see how this goes; I can only ask for your cooperation.

 

[Yukon_Shane]

Firstly thanks to you (and others) for taking this on and publically documenting your findings. I think this charger is probably the best option for the DIY EV market at the moment with the exception of the (extremely frustrating!) inability to change the charging profile without sending it back to the manufacturer. I hope this turns into something that can be completed by the average person.

With respect to the legal question you raise in the other post, I can't speak to the legality of publically providing this information; however, my understanding is that this design was "borrowed" from delta-Q anyway so I don't think that there's much of a moral argument to be made for not sharing any and all of your findings regarding the charger design and operation.

just my two cents...

 

[Coulomb]

Firstly thanks to you (and others) for taking this on and publically documenting your findings.

Thanks for the kind words, and thanks for starting off discussion on the discussion page!

I hope this turns into something that can be completed by the average person.

Alas, I think it won't be accessible to everyone. But hopefully, they will be able to find someone at their own EV club or at least in their own city that can do it for them. Freight of complete chargers is a problem. Worse is the way that Chinese import restrictions effectively prevent sending anything *into* China. A colleague had his charger fail, sent it back, and it got stuck in customs, and wouldn't get through despite being marked "returned for repair". They eventually sent a complete new charger. They obviously can't afford to do that very often. So even if the manufacturer was more willing to take on repairs and/or firmware updates, they simply can't.

however, my understanding is that this design was "borrowed" from delta-Q anyway so ...

I've heard that, and I have some Delta-Qs (or maybe clones of those) from people that actually mistook them for Elcons. ("Oh, you repair and upgrade Elcons? Here, take a look at this for me"). I see a lot of similarities, perhaps even a similar basic layout with some exceptions. But the processor is totally different, so from the firmware point of view, nothing we've learned from the Elcons will apply (or if it does, only in a rather general sense).

Maybe Delta-Q's will be next...

 

[PStechPaul]

I'm not familiar with the Elcon chargers, and my only experience has been with the EMW DIY chargers, which are open-source. I am trying to come up with ways to improve that design for those who have had problems with it, and I am curious about the overall design of the Elcon. I will be following your "facts" thread to see if you will have some block diagrams, schematics, and theory of operation, and I might use parts of that for my purposes. I don't intend to compete in any way with either manufacturer, but perhaps I might be able to provide alternative monitor and control modules that could be used with the basic hardware of either.

I just noticed that this is about the firmware, and I am mostly interested in the hardware, but it is interesting to see some of the assembly code and peculiarities of the 8051 type device. It seems the Elcon website does not have very much detailed information.
http://www.elconchargers.com/index.html

 

[Coulomb]

I am curious about the overall design of the Elcon. I will be following your "facts" thread to see if you will have some block diagrams, schematics, and theory of operation...

Well, there is already the Elcon/TC charger 1.5 kW schematics and Troubleshooting and Repair threads. I was thinking about separate Elcon/TC charger FAQ page as well.

It seems the Elcon website does not have very much detailed information.
http://www.elconchargers.com/index.html

That's the Elcon distributor in California; these are re-branded TC chargers from China. Though I note that Elcon have some models (the 2.5 kW models) that TC are not advertising. I suspect that the 2.5 kW models are no longer made, but they are still in stock in Sacramento. Or perhaps their web site is out of date. There is a little more information on the TC charger site:
http://www.tccharger.com/english/ .

 

[kennybobby]

It seems to me that the difficulty in making changes is compounded by the fact that each model of the charger has specific voltage, current and power ratings, calibrations and limits that are unique for that model in both the hardware and the controller.

And the unique factors are stored in the chip's eeprom which is separate from the flash memory of the firmware.

It can be done, but it is not a one-click and done process as yet.

 

[Coulomb]

And the unique factors are stored in the chip's eeprom which is separate from the flash memory of the firmware.

This is true; this is notionally a good separation.

Unfortunately, in way too many of the firmwares that I've examined, they don't read the unique factor from the EEPROM as they could and should, but use "constants" that end up in the flash. When I say constants, I mean values that won't change day to day but will change if the voltage or current limit is changed. That makes our job a little harder.

But once we have found all these, we can actually fix them, and our canonical firmware(s) can be nicely independent of the unique factors.

 

[Caps18]

It seems to me that the difficulty in making changes is compounded by the fact that each model of the charger has specific voltage, current and power ratings, calibrations and limits that are unique for that model in both the hardware and the controller.

And the unique factors are stored in the chip's eeprom which is separate from the flash memory of the firmware.

It can be done, but it is not a one-click and done process as yet.

But, could you clone the eeprom of a version of the charger with the voltage level you want then?

 

[pdove]

But, could you clone the eeprom of a version of the charger with the voltage level you want then?

Yes most definitely. We can extract the EEPROM values from one an put them in another.

However, the EEPROM is uniquely tuned for the resistors present in that unit so they won't be exactly the same.

We have replaced processors and created the EEPROM values based on the resistors dividers in the unit.

 

[piotrsko]

What I have seen is the max voltage is based on the transformer ratio and detector resistor network, or in the case of my 216 max unit, 216 vdc. Lower parameters to about 170 are possible (20%) with the eeprom settings but not higher. Custom fiddling with the divider could get you lower still at the expense of higher tops. This appears to be limited by the 0 -5 range of the comparator.

 

[Coulomb]

What I have seen is the max voltage is based on the transformer ratio and detector resistor network, or in the case of my 216 max unit, 216 vdc. Lower parameters to about 170 are possible (20%) with the eeprom settings but not higher.

I'm not sure what you are saying here. With the right firmware, you can go pretty much as low as you want. I have converted a 144 V nominal charger (170 V max) to a 48 V nominal lithium charger. It's still 30 A (it is a 5 kW charger, 30 A at ~170 V), and I'm still limited to 15 A. I got this charger free as part of an EV kit; I really only wanted the lithium batteries for a solar power system. So I don't mind throwing away about 66% of the charger power by running at one third voltage. So a lot more than 20% reduction in output voltage is possible.

Edit: I experimented with forcing more current out of it, assuming that the bottleneck was the front end. I did manage 18 A out of one side (normal maximum is 15 A per half; a 5 kW charger has two 2.5 kW chargers paralleled inside). But I figured that the output stage has limits too, so I put the limit back to 15 A.

 

[Coulomb]

I'm having trouble finding a way to post snippets of code.

There is the code tag, but that puts the code into a narrow window with a scrollbar, about a third of my desktop screen, and not adjustable. The scrollbar is a nuisance. I tried saving the result in HTML, which looks great in color, but the HTML tag shows the HTML code, and doesn't render it.

I tried attaching an image, and that works well. I can get the URL of the attachments, and insert them inline, so I can refer to them with next near the code. But it's a lot of work, and it leaves these large thumbnails at the bottom of the post. I think I've seem others get rid of these thumnails, but can't find a way to do this. I tried removing the attachments, and that seems to work, but I was horrified to find that it was only showing them from my browser's cache, and now a few days after posting them they've disappeared. They likely never appeared for anyone else.

I've tried just pasting text and using a fixed point font like Courier New, but multiple spaces are compressed to a single space, making the code unreadable.

Does anyone have a good solution for this?

Now I'm off to re-upload all my images. Joy.

 

[Coulomb]

My question is this, are they doing this on purpose to make it hard for someone else to program, or is it to using older and maybe cheaper processors etc.

I saw Jack showing how to program with the CAN unit, and it was easy to re program. I'm just amazed that the software the factory used has not been hacked, I'm sure that the hardware they use cant be anything special.

I see no evidence that they are making this hard for others to program, other than the security bits, which are pretty standard in the industry.

Yes, the CAN version is easy to program, and I'm a little surprised that more people don't go that route. But I guess many people want a one box set and forget solution. Perhaps if they put a real, isolated RS232 port on the charger instead of the CAN bus, more people might have chosen the CAN option. But I can see that CAN is more suitable for cars, which is surely their main market.

Or maybe make them all CAN, but with a plug-in, easily removable (say with one screw) module that controls the parameters for stand alone applications, and that module is friendly for people to program/reprogram. But I guess that would add a lot to the cost. Actually, I guess I'm describing Jack's programming box; I suppose I'm suggesting making that box either part of the charger, or readily available from the manufacturer (as a sort of ordering option, like the CAN box is now).

 

[PStechPaul]

The way I display images is to store them on my web space and then use the image button to wrap the [ img ] tags around the URL.

I thought there might be a BB code equivalent to the HTML < pre > tag but I could not find it.

Perhaps it would be useful to post a direct link to the code in TXT or ASM format which would be recognized by a smart text editor like Notepad++ and displayed as source code. I think you can attach text files and source code. I see that TXT is supported but not ASM or CPP. Maybe use DOC format?

The images may be helpful to show comments and highlight parts of the code, but having the actual text may be more helpful.

 

[dcb]

I thought there might be a BB code equivalent to the HTML < pre > tag but I could not find it.

use CODE (or the hashtag button)

this
   text
      is
        really
           pre

 

[CKidder]

Is there any information available on how to update the charging profile yourself? I need to change the parameters on a TCCH 4kw charger and I'd rather not have to send it to California. Would this be possible with the same tools required to work with the firmware or is there some different pathway? Is it just an EEPROM edit?

 

[kennybobby]

...I need to change the parameters on a TCCH 4kw charger

Same tools are needed and you could easily do this. If this is a factory TCCH Lithium charger then it might be a good candidate for reading the firmware before making changes--we suspect that the elcon versions are slightly different than the TCCH and it would be a big help to see for sure.

 

[CKidder]

Same tools are needed and you could easily do this. If this is a factory TCCH Lithium charger then it might be a good candidate for reading the firmware before making changes--we suspect that the elcon versions are slightly different than the TCCH and it would be a big help to see for sure.

Oh, sorry, it's an Elcon PFC4000. Is there any write up somewhere about how to tweak the EEPROM settings or would I have to figure out the format?

 

[pdove]

Oh, sorry, it's an Elcon PFC4000. Is there any write up somewhere about how to tweak the EEPROM settings or would I have to figure out the format?

You do not need to Change EEPROM.

The curve data is located in the firmware.

You modify the curves.c for your requirements. Then you'd modify the curves.Uv2 Microvision project file to use that file name, and save that as a new file. Use Microvision 3 to compile it to a .SRC (assembly language) file.

This .SRC file has to be massaged a bit; modify the edit_curves script to produce a new .a51 file from the .SRC file. Save this new script and use Cygwin to run it and actually produce the new .a51 file compile and link it... produce a new ordered hex file.

 

[Coulomb]

Oh, sorry, it's an Elcon PFC4000.

They are all much the same inside, except that there is a 1500 W model and a 2000+W model. In the USA (Elcon branded chargers) there seems to be a distinction between 2000W and 2500 W, with the former somehow optimized for 120 V operation. A 4000 W charger will have two 2000 W units inside it, a "master" and a "slave".

Is there any write up somewhere about how to tweak the EEPROM settings or would I have to figure out the format?

As mentioned above, you don't need to change the EEPROM values (as you would expect), but it's all in the 8 KiB firmware image.

I do intend to publish full details on how to do the firmware upgrades, along with all the firmware source code needed. I'm just a bit snowed under (not literally; it's the middle of summer here).

I note that this is moderately complex software; this will not be for everyone.