INFINEON CONTROLLERS FOR DUMMIES

[rkosiorek]

what might be easier is to design an 18FET output board and heatsink. then take the 12fet board, remove the fets and pass the drive signals from it to your new output board.

might also be a good idea to move the current shunt to the output board. somewhere else, i saw a thread where someone had done this with a different controller. i think that it was a brushed controller actually, but eh same basic idea would apply.

not sure if 18fets provide that much more advantage over 15. whichever side (high or Low) has the PWM signal needs fewer FETs. it is operating more efficiently than the side that is only being switched on and off.

but what the hey, bigger is better, right?

rick

 

[methods]

Doc,
I am running 50A on the teeny tiny credit card sized 6 fet box.
Hardly gets warm.

-methods

 

[Doctorbass]

Doc,
I am running 50A on the teeny tiny credit card sized 6 fet box.
Hardly gets warm.

-methods

Nice!!... but that may not be safe for long term use...

remember.. this is Infineon controlelr FOR DUMMIES... so all newbi here must know that this is not a recommanded use otherwise that could just create more problem than helpinhg them to well understand how they work and what are their real safe limit.. :wink:

But i admit ... thsi is encuraging!!.. after all.. i'm using 100A on the 12 fets version of a Crystalyte .. pretty similar no?

Doc

 

[John in CR]

not sure if 18fets provide that much more advantage over 15. whichever side (high or Low) has the PWM signal needs fewer FETs. it is operating more efficiently than the side that is only being switched on and off.

Ah, so that's why the Infineons that I have contain 15fets. Now for a dummy question, can different fets be mixed on one controller? eg Can I just replace the 12fets that are working harder as an upgrade or do all HAVE to be identical?

John

 

[Julez]

Hi!

I just tried to flash my controller with the device shown here:
http://endless-sphere.com/forums/viewtopic.php?p=134200#p134200

I hit the "Start Transmit" button, and connected the TTL concerter to the controller.
I did not connect anything else to the controller.

How do I know if I was sucessful?

The "transmit gauge" never shows anything.

I have some questions about the settings, too:
What does "phase current" "rated current", "tolderance", "Block Time" and "Auto Cruising Time" stand for?

Thanks for the help in advance,

Julez

 

[Julez]

Hi!

I just tested my push trailer yesteday, and I logged arounf 24A max current. So, clearly the flashing did not work.
What did I do wrong?

Cheers,

Julez

 

[philf]

OK. Weird question time...

In playing with the Infineon's regen braking option, I've made a weird observation...

When I spin up the wheel, and then engage the e-brake line into the controller, the wheel slows down abruptly (which it should), and all seems happy enough. In practice, though, I've noticed that - with the e-brake engaged - the bike is MUCH harder to roll backwards than forwards. You can even HEAR it complain about rolling backwards.

Is this a software "feature", or is it actually possible that I have a non-optimal working combination of phase/Hall wiring? Trying to get my head around what's happening, here...

This feels like software to me, but I maybe someone has a better answer?

Cheers!

 

[John in CR]

The hall combination IS non-optimal for the wheel spinning backwards. Is the added braking force strong enough to work as a hill brake at a stop light? What goes on with the electrical output in that instance, ie does it throw some reverse polarity at the battery or something or just throw it back at another phase and more like a dead short of the phases? IOW could rolling backward with your brakes on be harmful to something with regen?

John

 

[fechter]

This is pretty common with most brushless setups. When you roll backwards, the low side FETs are commutating, and the high side FETs are conducting through their body diodes, which resembles a short circuit. If you put enough force on the motor, the current in the phase wires could get high enough to blow a FET, since there is no current limiting. As long as the reverse torque never goes higher than what you can get in the forward direction, the FETs should stay happy.

 

[methods]

Hi!
I just tested my push trailer yesteday, and I logged arounf 24A max current. So, clearly the flashing did not work.
What did I do wrong?
Cheers,
Julez

Attach RX, TX, Ground, and +5V
Start the software and press transmit.
Then attach your battery pack to the controller
Half a second after powering the controller you will see that status bar fill on the software
Then just disconnect everything.
You dont have to restart for the flash to take.
That is how I did it and it has worked well.
Others have other methods.

If this does not work swap RX and TX

-methods

 

[geoff57]

Hi
Care should be taken when connecting your controller to the computer for reprogramming.
Your MCU (the infineon chip) should be powered down the best way is to disconnect the controller battery pack from the controller at the very least the ignition switch should be off isolating pad Vcc-L from the battery pack, I disconnect the whole battery pack myself and put ignition in the off postion.

The wiring I have no problems with as I use one of the original programming leads not the newer cheaper ones.

the method is the same for both direct connection for Rx, Tx and GND and put a switch in the +5v wire, some use a momentary make switch I use a normal type switch it stays on til i say off.
method of use is as follows set up program link to the correct com port there is a choice of 4 I am using a usb to rs232 converter this links to com port 4, ports 1and 2 report as not found com 3 has no error mesages but does not send the information or report the information as being sent.

So having set everything including the com port connected up the cable with the switch open I then click on the transmit button on the screen then close the switch while looking at the screen, you will see the bar go across and a message come up of if the transmit worked. Open the switch click on the stop transmit and that is it.

remember to save your .asv files do not use any spaces and keep screen shots or text files with the settings in just in case.

The parameter designer will not remember what com port to go to you will have to reset each time.

Geoff

 

[solarbbq2003]

luckily we're in the dummies thread: how do you link a program to a particular comm port? if using usb port will same apply?
or only it applies to rs232?

 

[Jeremy Harris]

On the subject of Infineon controller MOSFET upgrades and heat, does anyone happen to know what the FET driver chip part numbers are on these Infineon controllers please?

Also, can anyone confirm that the controller chip is really the Infineon XC864 (rather than quoted XC846, which seems not to exist), please?

Part of the heating issue already mentioned may simply be down to the FETs not being turned on and off fast enough, so increasing the switching losses. I'm working on making an add-on power board for the little controller, so knowing whether or not the on-board driver is up to driving bigger devices (which will have a bigger gate charge and hence higher gate drive current) would be useful.

As an aside, it looks pretty easy to design and build a power board that will massively increase the current for the little Infineon, which may make it a reasonable choice for a hall-modded RC motor. 48V at at least 200A looks pretty easy, going up to 60V at 400A looks reasonably OK. I'm not sure there's much merit in going to higher voltages with the readily available (and affordable) RC motors, as few seem able to safely run above about 50V or so (I may be wrong here, but the gearing problem starts to get tricky anyway, due to the high motor rpm).

Jeremy

 

[Julez]

Hi methods and Geoff,

thank you for your replies.

methods, if everything else fails, I will try your way.

Geoff,

I did exactly as you describe.
The controller was lying on my table, it was brand new and has never been connected to anything.
The first time I tried I simply stuck the 5V contact cable last into my makeshift connector.
Now that I have soldered everything together with neat connectors, but without a switch, is it possible to just plug the 4 wires from the controller in at the same time, after hitting transmit, or is there no chance to make it work without a switch?

Thanks,

Julez

 

[geoff57]

luckily we're in the dummies thread: how do you link a program to a particular comm port? if using usb port will same apply?
or only it applies to rs232?

Hi
I dont know about other programs but this is how the parameter designer works.
It was made when com or serial ports are still used as standard, the com ports were the 9 and 25 way female D connections on older computers, we now have the USB or Universal Serial Bus and so the com ports are no longer used and so are very often not installed on the case of a full computer though the ports are on the motherboard and with laptops are not even considered.

if you buy a USB to RS232 adaptor the computer will usually find and install the software to work it its self mine took about 5 min's i did think somthing was wrong but waitedand it sorted itself out the drivers are found on the net all over the place. Once it is done every time you plug in to the USB port (any on the computer) the computer will set up the com port. I belive this is what they call "plug 'n' play".

Jeremy
The infineon chip is the 846 as quoted we have all been searching for information on it and can't find any. I can't see any other chips on the board each bank of FETs are driven from a collection of components all marked on the board by number and phase eg R1A, R1B, R1C fed from one pin of the MCU hope this helps I can post a closeup of one phase if you want B would be best to get in the MCU pin as well.

Julez
If you don't fit a switch it is not a loss many including me have just found it easyer. If you have no switch then do not connect up the controller to the connector untill the transmit button has been clicked, once the transmit button is clicked a message transmit started should come up whenyou connect up to the controller the transmit bar will go across and then you will see a message transmit finished then click on the stop transmit button and a message comm is closed will come up. if transmit failed comes up instead of of finished disconnect from controller and reconnect, the switch makes it easyer to keep an eye on what the screen is doing which is a main reason for it.

one more thing to try set up the link to transmit click on the start transmit button then mesure the voltage across +5v and GND to make sure it is 5v a TTL board can transmit at several diffrent voltages as low as 3.3 v can be used, if this is the voltage set then the controller may well not program, I have checked my programming lead that uses 5v.

Geoff

 

[Jeremy Harris]

Jeremy
The infineon chip is the 846 as quoted we have all been searching for information on it and can't find any. I can't see any other chips on the board each bank of FETs are driven from a collection of components all marked on the board by number and phase eg R1A, R1B, R1C fed from one pin of the MCU hope this helps I can post a closeup of one phase if you want B would be best to get in the MCU pin as well.

Geoff

Thanks, Geoff. Don't worry about the photos, I'll just wait until my controller arrives. In some ways, having no on-board driver chip makes it easier, as I think I will just make a board that will accept direct phase connections at logic level. That way it may well be a multi-purpose power board that will work with more than one controller.

I've just been doing a few sums, and it looks like a reasonably good starting point might be to use either 18, or perhaps 24, IRFB3077 FETs, driven by three IR2181 drivers. The IR2181s are good for 1.4A of high side drive, which should be OK for up to four high side FETs, I think. The IRFB3077s seem to be the best bet for low Rdson and high current, as they are rated at 120A continuous, 210A peak, 75V and a max Rdson of 3.3mOhm. 18 FETs should allow 360A continuous, which I'd personally derate to around 200A just to keep within reasonable heat dissipation conditions. Still, 200A continuous at up to around 70V (14kW) should be enough for most purposes................

Coupled up to a pair of HXT 80-100 motors, synced to a common shaft and fitted with Hall sensors this should make a fairly neat combination (and not to expensive, either).

Jeremy

 

[Julez]

Hi!

I have now tried everything, but nothing works. My usb converter works well, as I have used it to update other devices.
I tried to swap RX and TX, I switched the controller on before or after the converter was plugged in.

The worst thing is, the controller does not work anymore, too.
I cannot exclude the possibility that I accidently swapped the 5V and GND contact during my various tries.
Is it possible to destroy the controller by this?

Before, the idle current consumption was 50mA, now it draws 150mA.

So something is likely to be fried, right?
Great.
My sunday is now officially ruined

 

[philf]

The worst thing is, the controller does not work anymore, too.
I cannot exclude the possibility that I accidently swapped the 5V and GND contact during my various tries.
Is it possible to destroy the controller by this?

Before, the idle current consumption was 50mA, now it draws 150mA.

Ack. This is never good. Yes - reversing the polarity of the power, at the level we're injecting it into the board, here - is never a good thing. Before jumping to the conclusion that it's really fried, though - look carefully (preferably with a magnifier) at your work. It's not impossible that something else is or has been shorted out during your attempts to get this to work. A wire hair, solder ball, or even a solder bridge... you never know...

Based on what you've written, though - I'm not feeling too hopeful.

 

[solarbbq2003]

thanks for info geoff on usb/rs232, I'm still abit confused, why use a rs232 connector at all? why not just use a usb-ttl connector? am I missing something? and could this be related to julez prob?

 

[geoff57]

thanks for info geoff on usb/rs232, I'm still abit confused, why use a rs232 connector at all? why not just use a usb-ttl connector? am I missing something? and could this be related to julez prob?

Hi
That is a anomaly with computers although everyone thinks of comm ports as RS232 that was just what IBM chose, previous computers used other serial standards, RS232 was chosen for speed and distance (about 100m at slow speeds but keep it to 10m or less and you could go as fast as you could) the information could be sent down as little as 3 wires ( a guy at a computer firm made a cable out of 3 core mains domestic cable), computers have advanced and USB was the final nail in the RS232 coffin for genral use.
RS232 still has specialised uses but now through the USB port and a USB to RS232 converter, this is what i use as my PC link is the first one, before Philf reverse engineered the device in the regen thred.

He gave detailed discriptions of everything I will summarise:-
1: From the computer comes a signal intended for a RS232 port.
2: In the device there are 2 chips the signal goes into the first chip and is converted into a TTL signal.
3: The second chip takes the signal from the first and converts it to a RS232 signal.
4: that is outputed through the D connector.

now Philf looked at the pics of the chip in the device I have and found it was a RS232 to TTL converter so deduced correctly that removing the TTL to RS232 chip in a USBto RS232 converter would allow for a direct output to the controller, he had one of these converters onthe shelves doing nothing so tried this and it worked, thus removing the last 2 chips as they are not needed.
We now have knuckles finding a USB to TTL converter on the net so saving a device reworking. This is the same type of board as the red board that julez is using.

So this is the whole point people assosiate comm ports with RS232, this is not fixed when sending the information via a USB port, a signal sent out intended to go to a RS232 port does not need to go to a RS232 port, Philf showed how the USB to RS232 converter does this in 2 stages (of which he removed the second) so we can quite easily use a converter that has just one stage the USB to TTL converter but tell the computer it is a USB to RS232 converter. the comm port to use will probably be comm 4

Geoff

 

[Julez]

Hi!

I just had a careful look at the PCB. The positive contact goes right to the chip without anything to burn in between. So it must really be the chip that is fried.
So I'll have to buy a new controller.
D'oh...just when everything started working out right...

 

[John in CR]

Now you guys understand why I'm scared to touch the stuff. BTW, my computer still has a serial port, and I even use it (for my brushed Kelly controller settings). Does that make tying into the software even easier for me?

Now for my "Dummy" question. I see all these different holes on the board for GND. Are they all common? Is it the same as battery negative? I ask because my version has a wire going from BK to outside the controller. It expects +12V as the electronic brake cutoff, but I'd like to try it to battery neg (if that's ground) to see if it then does regen braking. That is if you E wizards think it's a good low risk idea. Should I measure voltage across it first?

John

 

[geoff57]

Now you guys understand why I'm scared to touch the stuff. BTW, my computer still has a serial port, and I even use it (for my brushed Kelly controller settings). Does that make tying into the software even easier for me?

Now for my "Dummy" question. I see all these different holes on the board for GND. Are they all common? Is it the same as battery negative? I ask because my version has a wire going from BK to outside the controller. It expects +12V as the electronic brake cutoff, but I'd like to try it to battery neg (if that's ground) to see if it then does regen braking. That is if you E wizards think it's a good low risk idea. Should I measure voltage across it first?

John

hi John
the software links to a comm port or serial port but the controller takes its information in TTL format, there is a RS232 of serial port to TTL PC link avalible I use it, the price is not cheap and I thought it would put off customers and that USB should be used now anyway. When Philf came up with the USB to TTL converter at so much more reasonable a price.
You might be able to use the Kelly PC link to connect to the infineon you will have to find what sort of link it is if it is just wires then you are out of luck if there is a chip in there then there is a chance, this is a BEWARE DANGER idea the best thing would be to get a USB to TTL converter they are not that much .

Now for the Dummy answer.
Yes all the GND pads/holes are common and linked to the negitive wire of the battery.
What controller do you have if it is a 12 fet then I can say the wire that is connectet to the BK pad is designed to be connected to GND that is why ther is a row of GND holes next to all of thoese holes marked with letters BK is the switch for regen when that is linked to GND then ifwhen the ebrakes are applied regen will be activated as well if the controller has been programed for regen.
There is a wire that is expecting 12v+ as an ebrake this wire is yellow and on its own in the controller it is connected to a hole/pad marked EBS+ next to the standard ebrake yellow wire that is connected to EBS-.
hope this helps
Geoff

 

[solarbbq2003]

to geoff57
i'll try get a pic posted here shortly of the pins related to alarm, having problems replying to forum emails when i hit submit it dont go no where, grrrrrr

 

[solarbbq2003]

chip info
http://www.infineon.com/cms/cn/product/channel.html?channel=db3a30431b0626df011b315ce2eb6a27