[RIPPERTON]

Having custom ordered the wrong voltage charger, was left with the only option of shipping the charger back to China to have it adjusted but thanks to Ian Bruinsma here in Sydney was able to make the voltage adjustable.
Originally the voltage cut off at 80.1v (LiFePo4 24s 1p 72v 8Ah)
This charger is for my Giant MTB with hub motor.
Putting a trim pot and resistor in was able to make the output DC voltage adjustable from 83v to 89v. Ideally it should cut out at 87.6v for 3.65v each cell but I still havnt got a BMS for this battery yet so have left it at its minimum 83v.
Ian is starting to convert a Smart car to AC drive so you might find him around here asking questions so help him out if you can.
cheers Ripperton

 

[DIYguy]

Having custom ordered the wrong voltage charger, was left with the only option of shipping the charger back to China to have it adjusted but thanks to Ian Bruinsma here in Sydney was able to make the voltage adjustable.
Originally the voltage cut off at 80.1v (LiFePo4 24s 1p 72v 8Ah)
This charger is for my Giant MTB with hub motor.
Putting a trim pot and resistor in was able to make the output DC voltage adjustable from 83v to 89v. Ideally it should cut out at 87.6v for 3.65v each cell but I still havnt got a BMS for this battery yet so have left it at its minimum 83v.
Ian is starting to convert a Smart car to AC drive so you might find him around here asking questions so help him out if you can.
cheers Ripperton

Thanks Ripper! you never know. I stuffed it away in my charger file. I have a 6 kw Elcon set up for 192 volt AGM's. Lithium is in my future and it will have to be modified.... but I was wondering about this also...

 

[SmartArseEV]

OK - It was me... I am now visible as SmartArseEV...

Thanks RIPERTON for the acknowledgment.

This was not my best work (it was a quick and nasty modification), but it was worth posting. If it helps any of you EV converters out there well excellent!

 

[Dougnutz]

... was able to make the voltage adjustable.

OH! thank you thank you, this is exactly what I needed

 

[FWD]

Having custom ordered the wrong voltage charger, was left with the only option of shipping the charger back to China to have it adjusted but thanks to Ian Bruinsma here in Sydney was able to make the voltage adjustable.
Originally the voltage cut off at 80.1v (LiFePo4 24s 1p 72v 8Ah)
This charger is for my Giant MTB with hub motor.
Putting a trim pot and resistor in was able to make the output DC voltage adjustable from 83v to 89v. Ideally it should cut out at 87.6v for 3.65v each cell but I still havnt got a BMS for this battery yet so have left it at its minimum 83v.
Ian is starting to convert a Smart car to AC drive so you might find him around here asking questions so help him out if you can.
cheers Ripperton

is it still working? no problems over time of use?

i am going to do this mod to mine is 292 volt max and i need 262 so i do the down adjustment

 

[FWD]

double post

 

[FWD]

ok my elcon with 292 end voltage is not the same
between pin 2 and dc+ there is not 75k ohm but 600k ohm.

i geuss thats because the voltage is higher

so the 10x R1 of 75kohm to lower the voltage wil not work.

but is it 10 x 600kohm so 6Mohm or something else?

 

[Otmar]

Has anyone figured out how the Elcon charge profiles are changed?

I've heard that it may be possible through a serial interface inside the charger. I've done a little research, but am not getting too far. In case it helps the others this is what I have figured out:

My charger board is a TCCH1K5W A ver 1.4
It has a 5 pin header for which I assume is for programming profiles.
It has a NXP 89LPC938FDH uC near the connector, this shares battery pack ground.
As best I can tell working around the potting:

Pin 1 = VCC 3.3V
Pin 2 = RST pin 6 on uC, 105K to gnd
Pin 3 = Pin 23 on uC, 10K to gnd
Pin 4 = Pin 22 on uC, 15K to gnd
Pin 5 = Gnd.

None of the pins send out data when powered up, and Pins 3 and 4 maintain the approximate resistance to ground (they seem to be inputs when powered up).

I suspect they may be programming the whole chip through a bootloader, but these are not the pins that NXP uses for that.

I would buy a number of these chargers if I could adjust them myself, but as it is I'm very hesitant since my pack voltages tend to vary. If anyone has ideas for reprogramming the profiles, I'd love to hear them. In the end, once my BMS is done I may just order them turned all the way up and use the analog or CAN inputs to regulate the finish charge.

 

[Coulomb]

In case it helps the others this is what I have figured out ...

Thanks for the info, Otmar.

... I may just order them turned all the way up and use the analog or CAN inputs to regulate the finish charge.

My understanding is that the CAN versions are very different from the others, and have *no* profiles built into them. You have to provide the instantaneous voltage and current limits yourself. If you like, the CAN versions are "turned all the way up" from the factory.

Until recently, I did not believe that the CAN versions listen to the analog control signal. So we're not planning on using it at all. Something I read recently seemed to contradict that. Anyone know for sure? We have a pair of 410 V CAN version chargers, so I could do some experiments, but exploring the charger internal workings, while interesting, is way down the priority list at present.

 

[bjfreeman]

Has anyone figured out how the Elcon charge profiles are changed?

not familar

My charger board is a TCCH1K5W A ver 1.4
It has a 5 pin header for which I assume is for programming profiles.
It has a NXP 89LPC938FDH uC near the connector, this shares battery pack ground.
As best I can tell working around the potting:

Pin 1 = VCC 3.3V
Pin 2 = RST pin 6 on uC, 105K to gnd
Pin 3 = Pin 23 on uC, 10K to gnd
Pin 4 = Pin 22 on uC, 15K to gnd
Pin 5 = Gnd.

None of the pins send out data when powered up, and Pins 3 and 4 maintain the approximate resistance to ground (they seem to be inputs when powered up).

I suspect they may be programming the whole chip through a bootloader, but these are not the pins that NXP uses for that.

this is known as JTAG and is used to program the code to run the micro. unless you have the ability to compile the code do not use these.

I would buy a number of these chargers if I could adjust them myself, but as it is I'm very hesitant since my pack voltages tend to vary. If anyone has ideas for reprogramming the profiles, I'd love to hear them. In the end, once my BMS is done I may just order them turned all the way up and use the analog or CAN inputs to regulate the finish charge.

I am not sure of the Can messages, I would need to see the stream an reverse engineer it base on what I know. here is what I use CanCharger spec

 

[Coulomb]

this is known as JTAG and is used to program the code to run the micro. unless you have the ability to compile the code do not use these.

If that is a JTAG port, then that's what we'd like to understand, it seems to me. It's an 80C51 compatible processor, so if we could get to the code, it could be readily disassembled.

But looking at the datasheet for the processor, I'd say it could also be an SPI, I^2C, or serial (RS232/485) port. The datasheet mentions:

• Internal fixed boot ROM, containing low-level IAP routines available to user code.
• Default loader providing In-System Programming via the serial port, located in upper end of user program memory.

The flash memory (8kB) seems to be byte programmable; that's almost certainly where the charger parameters ("curves") will be stored. If the flash can be manipulated from the serial port, then it makes sense to use the serial port to change the "curves". JTAG is not mentioned in the data sheet I found:

http://pdf1.alldatasheet.com/datasheet-pdf/view/123811/PHILIPS/P89LPC938FDH.html

So the first thing might be to determine where the two lines that are not power or reset connect to on the processor. That will determine what sort of port it is.

But after that... who knows. Each 1kB sector of the flash memory has a "security bit", which will probably mean we can't read the contents, only write to it. Writing blind into 8 kB of memory isn't going to work.

I am not sure of the Can messages, I would need to see the stream an reverse engineer it base on what I know. here is what I use CanCharger spec

The CAN messages are extremely simple; they are documented here as well as on the TCCharger site:

http://lithiumate.elithion.com/php/controller_can_specs.php#ElCon_messages

They just set the maximum voltage and current, and allow the charger to be turned off. That's it. The CAN version of the charger doesn't have any algorithms/curves; you have to send it CAN messages to set the voltage and current.

Actually, since this chip has a UART and is close to the 7-pin connector where the "CAN port" is connected, and the "CAN port" is just 2400 bps RS232, the UART is likely used to talk to the CAN interface box.

So I'm thinking that the most likely candidates for programming the parameters would be SPI and I^2C; the former is far more likely.

 

[bjfreeman]

If that is a JTAG port, then that's what we'd like to understand, it seems to me. It's an 80C51 compatible processor, so if we could get to the code, it could be readily disassembled.

Jtag is more a protocol the is implemented with a serial port and the reset. the device that connected to the JTAG hits the reset to enter into boot loader mode then send data to the CPU.
do those pages represent the CPU you have?
if so they show how to upload an download Files which I assume is the code. So you can reverse engine what pins are being used
they even show how to program
http://liionbms.com/php/teaclipper.php
this follows most avr an arm open source IDE systems like Ardiuno, Maple. The difference is the target micro.
There is a DFU utility that you can upload an download to CPU using the USB port.

But looking at the datasheet for the processor, I'd say it could also be an SPI, I^2C, or serial (RS232/485) port. The datasheet mentions:

• Internal fixed boot ROM, containing low-level IAP routines available to user code.
• Default loader providing In-System Programming via the serial port, located in upper end of user program memory.

The flash memory (8kB) seems to be byte programmable; that's almost certainly where the charger parameters ("curves") will be stored. If the flash can be manipulated from the serial port, then it makes sense to use the serial port to change the "curves". JTAG is not mentioned in the data sheet I found:

http://pdf1.alldatasheet.com/datasheet-pdf/view/123811/PHILIPS/P89LPC938FDH.html

So the first thing might be to determine where the two lines that are not power or reset connect to on the processor. That will determine what sort of port it is.

But after that... who knows. Each 1kB sector of the flash memory has a "security bit", which will probably mean we can't read the contents, only write to it. Writing blind into 8 kB of memory isn't going to work.

think that is doing it the hardway.

The CAN messages are extremely simple; they are documented here as well as on the TCCharger site:

http://lithiumate.elithion.com/php/controller_can_specs.php#ElCon_messages

They just set the maximum voltage and current, and allow the charger to be turned off. That's it. The CAN version of the charger doesn't have any algorithms/curves; you have to send it CAN messages to set the voltage and current.

Actually, since this chip has a UART and is close to the 7-pin connector where the "CAN port" is connected, and the "CAN port" is just 2400 bps RS232, the UART is likely used to talk to the CAN interface box.

So I'm thinking that the most likely candidates for programming the parameters would be SPI and I^2C; the former is far more likely.

Basically the CAN is a Serial port that is then interfaced through a CAN chip then through a Can hardware intface to the CAN Bus. The Can chip formats the the stream.
so if you read the Serial port going into the Can bus chip you will see the stream that sets up he CAN chip

Thanks for the links will add this to my bridge node for Chargers.

I noticed that is says that it hangs if CAN is used. I believe this is more to the terminaltion of the CANBus network but could be from the CPU not being able to handle the CAN Chip interface.

 

[bjfreeman]

found this that should help in understanding the memory layout,
http://lithiumate.elithion.com/xls/eeprom_data.xls

 

[Elithion]

I am happy to see that the data on the Elithion website is so useful to your work.
Glad to be of help!
Davide

 

[ruckus]

Has anyone figured out how the Elcon charge profiles are changed?

Follow this link, scroll down, it tells how to select another curve that has already been programmed in:
http://www.electricmotorsport.com/store/pdf-downloads/charger_elcon_pfc3000.pdf
(link now fixed)

When you power on the charger it blinks several times to tell you which curve it is using.

I am having an issue where the charger works fine on short trips but if we drain the cells down quite a bit and try to do a fill-up then it flashes an error that we must have over-filled the batt or be using the wrong curve. I think this must be because we are filling 260ah batts which are bigger than most (they are certainly not full when the code happens). If it is unplugged and replugged then it fills them all the way up perfect (109.5v).

Currently the charger is on curve #3 (3 blinks) any hints whether I should go towards 2 or towards 4?

 

[Coulomb]

Follow this link, scroll down, it tells how to select another curve that has already been programmed in:

My understanding is that Otmar wanted to know how to program in other curves, or the same curve with different limits (e.g. after adding or removing 1 or 2 cells).

 

[bjfreeman]

Follow this link, scroll down, it tells how to select another curve that has already been programmed in:
http://www.electricmotorsport.com/store/pdf-downloads/charger_elcon_pfc3000.pdf
(link now fixed)

When you power on the charger it blinks several times to tell you which curve it is using.

I am having an issue where the charger works fine on short trips but if we drain the cells down quite a bit and try to do a fill-up then it flashes an error that we must have over-filled the batt or be using the wrong curve. I think this must be because we are filling 260ah batts which are bigger than most (they are certainly not full when the code happens). If it is unplugged and replugged then it fills them all the way up perfect (109.5v).

Currently the charger is on curve #3 (3 blinks) any hints whether I should go towards 2 or towards 4?

if your draining the pack below he 80% curve no profile will accommodate that.
that is one of the reason for a BMS.
did you put in your battery profile?
did you change the timeout?

 

[ruckus]

if your draining the pack below he 80% curve no profile will accommodate that.
that is one of the reason for a BMS.
did you put in your battery profile?
did you change the timeout?

Not draining that much, have a bms. I am saying 2 hr charge, no problem. 6 hour charge, problem. Not sure how far it gets, but suspect about 4hrs.
Did not put in a bat profile or change the timeout. Just the name timeout makes me think that is likely the problem. I will read your info linked above and then ask dumb questions...

It charges to the correct voltage, but does not seem to know we are using rather large batts (260ah).

Got the charger from Kois/Morrison, so who knows...

 

[Coulomb]

here is how to do charger profiles
http://www.orionbms.com/charger-integration/interfacing-elcon/
Tte profile is a text XML file. you an open it in notepad or any text editor.

Err, that seems to be a file that is processed by the Orion BMS, which sets the curve for use with an Elcon/TC charger with the Can bus option. So these would appear to have nothing to do with the curves stored inside the non-Can bus models, which I assume are the more common variety, and are the subject of this thread.

Or an I missing something?

 

[ruckus]

Ok, here is an "example" of possible loaded charge curves for a particular elcon.

http://www.evsource.com/datasheets/3kw Charger/Elcon 3kW Charging curves.pdf

You can see that both voltage and amp rate can be adjusted by selecting a different curve. HOWEVER, there is no way to tell what you have in your charger (that I know of), so it is a bit of a crap shoot.

I thought the elcon dealer in the U.S. could reprogram your desired curves so you would know what you have...