Infineon controller REVIEW+tests with 18x IRFB4110

hi guys,
I've just been on skype for 3 hours with keywin to sort out this problem with the 18 fet controllers, getting RSI in my hands from typing and earache from the wife for working on the computer for so long, i hope you like the results.

Yesterday I managed to get keywin to start stocking the 12 fet hard start controllers again,over night the debate on the 18 fet controllers being soft start got too bad to belive, so I desided to have a word with keywin and see if we could work out a deal.
I got one for you, the 18 fet at the moment is SOFT start, NOT programmable so fixed LVC alterable only via a pot.
Keywin will now be getting in new 18 fet controller boards, the exact spec is still unknown but this is what i do know:-
the new board is going to be 18 fet (same spacing I presume)
it will be programmable ( the software may be the same it may be diffrent I do not know)
functions as yet unknown but probably the same as for the rest of the Infineons many of the hole /pad names are the same as on other boards keywin had a look at the board he has a sample for testing with a puma at high volts no load conditions.
It uses a new chip still an infineon but not the standard xc846 this one has more built in memory than a standard xc846

allowing for tests and getting in a stock of the new controller boards he should be ready in a week (he said less I am alowing for problems)
Now this may help future 18 fet owners but what about me you will say this is what I though as well
So I have worked out a deal with keywin anyone who has bought an 18 fet board for conversion to 4110 fets can buy the board on its own with the small caps fitted but the big caps missing for $40 USD per board including shipping to the US transfer the caps from your old board to the new board, or you can get a board with the big caps fitted for $45 USD including shipping to the US.
replacement controllers cant be offered to everyone so are being limited to the 4110 fet converting gang.
the same pclink will work on both controller boards but the new board has a diffrent pin order.
The new board has 4 shunt resistors fitted at the factory before keywin gets them, lets see what they are like before we shun them.

I hope this has been helpfull.

Geoff
 
Thanks Geoff! for the time spent with Keywin to help about this issue!

This is great new! the best new would be if they could supply many 470uF 160Vdc capacitors with them!! the 100V are not really the best choice for the 4110 used at above 72V..

Please put me on the list to get one of these 40$ board.. so could have one

+ two spare board (the first board version i got) to repair it in case it would blow... making every tests i'll do unstoppable and fixable!!!

Doc
 
good work geoff.

i just bot big caps off ebay, 2 boxes of 100ct, 100V1000uF electrolytics like the rubycons.

my cost was a little over 30 cents each. will share out to anyone at 30 cents each, plus postage. 18mm diam x 40mm long. i think i can tape them on a piece of cardboard and send them through the mail in a flat envelope.

i did see small lots of 2 to 8 electrolytics in the 150V, 470uF range, but usually over $2 each. was thinking of steveo when i saw the 160V2200uF for $4.25.
 
Sorry about the duff link, I'm not sure why it won't tinyURL properly.

Here's the full link to this useful site:

http://www.infineon.com/cms/en/product/channel.html?channel=ff80808112ab681d0112ab6b826c0867&tab=2

The key to decoding the exact chip in these controller is the part number decoder in the presentation in this pdf:

http://de.sitestat.com/infineon/inf...90004&fileId=db3a304412b407950112b41994aa28ab

The decode is:

The first three characters are SAF for -40deg C to +85degC, SAK for -40degC to _125degC

The 4th, 5th and 6th characters are XC8 for the XC 800 family

The 7th character is 6 for Flash Module 1 (1 wait state), 5 for Flash Module 2 (0 wait state), 4 for ROM only

The 8th character is 6

The next group of characters can be:

C for CAN
L for LIN
A for ADC
M for MDU

The next two characters are:

2F for 8kB Flash
4F for 16kB Flash
2R for 8kB ROM
4R for 16kB ROM

The last character is:

R for TSSOP
F for TQFP
G for DSO

The chips on my 18 FET boards are only marked "F-XC846", so I'm assuming they are -40degC to +85degC parts with ROM only.

Jeremy
 
Interesting. . . I was wondering why a pot was soldered to the bottom of my boards.
I had assumed the boards were programmable.
I now understand that perhaps they are not.
I need to go test that because if that is the case I need to change the ad I have posted.

New boards sound interesting. 4 shunts in parallel would be very close to 1mohm so much better than 3 shunts.

-methods
 
geoff57 said:
I got one for you, the 18 fet at the moment is SOFT start, NOT programmable so fixed LVC alterable only via a pot.
Click to expand...

Hey Geoff,

Thanks for the hard work.
We got off on the wrong foot so I dont want you to think that I am trying to argue with you.
Try to take what I say below as an attempt to get to the right answer, not start an argument.

Please confirm the information that you stated above.
I just programmed my 18 fet board so I think there may have been a miss communication between you and Keywin.

I am unable to test the board under power, so it is possible that I am mistaken, but my suspicion is that if it can be programmed it will take to the settings.
Perhaps by "programmed" keywin meant that the boards can not be re-programmed with an instant start.
Keywin actually told me that he could ship the 18 fet boards back to the factory to be reprogrammed, but that is a different subject.

So it is my understanding that the LVC, Speed, current limit, etc. are all still programmable on the current 18 fet board.
This is an important point as it eases the adjustment of current limits quite a bit.
Otherwise people have to tune the shunt to tune the max current.

remember - not trying to argue :)

thanks,
-methods

EDIT:

Now I am really confused:
geoff57 said:
the same pclink will work on both controller boards but the new board has a diffrent pin order.
Click to expand...

Are you saying that the old boards ARE programmable but the LVC is not programmable?
 
methods said:
geoff57 said:
I got one for you, the 18 fet at the moment is SOFT start, NOT programmable so fixed LVC alterable only via a pot.
Click to expand...

Hey Geoff,

Thanks for the hard work.
We got off on the wrong foot so I dont want you to think that I am trying to argue with you.
Try to take what I say below as an attempt to get to the right answer, not start an argument.
Click to expand...
ok we start with a clean slate, as if we had not chated on the forum.
I found where you pull that quote from that was me telling everyone I contacted keywin on skype to make a deal for new hard start,18 fet boards/controllers the start of the quote was I got one as in I got a deal for you guys.

The rest of the quote tells you what type of 18 fet board is sold at the moment.

methods said:
Please confirm the information that you stated above.
I just programmed my 18 fet board so I think there may have been a miss communication between you and Keywin.

I am unable to test the board under power, so it is possible that I am mistaken, but my suspicion is that if it can be programmed it will take to the settings.
Perhaps by "programmed" keywin meant that the boards can not be re-programmed with an instant start.
Keywin actually told me that he could ship the 18 fet boards back to the factory to be reprogrammed, but that is a different subject.

So it is my understanding that the LVC, Speed, current limit, etc. are all still programmable on the current 18 fet board.
This is an important point as it eases the adjustment of current limits quite a bit.
Otherwise people have to tune the shunt to tune the max current.

remember - not trying to argue :)

thanks,
-methods
Click to expand...
to quote William Shakespeare "aye, there's the rub" by programable we mean that the infineon family of controllers can have some of the settings changed by the end user using this pclink and the program that is avalible free, you have both or you would not be asking the question. these settings are stored in part of the memory of the chip that is FLASH memory so rewritable like an SD card in your camara. The problem is when keywin bought these boards the infineon chip that was used was a variant without FLASH memory but a difrent type that requires a difrent way to reprogram the setting if at all, remember keywin looses alot in traslating english to chineese through google( he told me that was how he could communicate with us)so some of the things he tells you in an email may not be correct on skype i can ask the question and if the answer sounds like something was lost in translation i alter the question again. Now the program and pclink that some of us have to alter the settings on other fet sizes of controllers when pluged in to this 18 fet board (you have 6 I belive) and set to send the information to the board first it checks all the connections are made then when that is done the information is sent, this is where the problem is, there is no return comunication as to what is in the chip "the link is a one way street" although there is a Tx and a Rx the chip was programed not to give up its secrets by the Germans that designed the chip not even the board makers, so when you send settings to the chip the chip sends a signel back down the RX line i recived it and it was ok but when it comes to storing it the chip can't because the wrong memory is on the chipbut there is no code to tell the program that, or maybe there is and it is ignored as once a recived signel is sent the program stops looking at the RX line that I don't know. I DO know that there are difrent levels of settings some we can set some the factory sets dependent on the amount of FLASH memory on the chip, a standard chip found on the 12, 9, 6 fet boards at the moment has 2k Flash I presume the 18 fet was unfotunate to have no FLASH memory the good news s the new 18 fet boards I told keywin to get that are hard start have 8k FLASH memory this may meen more settings or finer settings of the ones we have I don't know.
The program saying it has altered the setting in the 18 fet when it has not is a pain to say the least but this new board should fix that.
as I said just a board will be offered to customers who have bought controllers without fets for insaine use and want an upgrade at a resonable price.


EDIT:
methods said:
Now I am really confused:
geoff57 said:
the same pclink will work on both controller boards but the new board has a diffrent pin order.
Click to expand...

Are you saying that the old boards ARE programmable but the LVC is not programmable?
Click to expand...
no the old boards are NOT programmable but if they were the port on the board to connect up the pclink is identical in pinout to the ports in the rest of the family but the new boards port has a difrent pin order.

sorry it is so long and some may be as clear as mud.
Geoff
 
Got it.
I am surprised that they wrote the code such that it would accept the programming but not write the info to the flash memory.
Too bad really, but no big deal. When prices are this good it is bound to happen.

I have refunded all paypal payments sent, corrected my ad, removed the programmer from the kit, and lowered my price to $57 on the 18 fet controllers.
Thanks for the clarification Geoff.

This is not a deal breaker. Most people dont really need the programming anyway. It is a fun tool but the boards are still an amazing value without that feature.
We will now need to get more handy with our soldering irons to tune that current limit up :mrgreen:

-methods
 
Hey guys, I'm just interested in whether the "soft start" 1 second delay can be fixed or not. Can't see why the LVC is such an issue (got something against hardware (pot?). :?:

If the start delay can't be corrected, wonder if Keywin will exchange the boards I just got?

DK
 
Am I mis-understanding something? Is somehow the LVC issue connected to the soft start (delay)? I'm runnning 80V right now, planning on 100-120V.

DK
 
Go back and read the last few pages again.

-methods
 
Well, page 11 people mention the small cap, and finally decide it's the software.
By page 12 Jeremy is saying our boards is the ROM version that can't be changed.

Geoff57 makes this comment which seems to tie the 2 subjects together - "I got one for you, the 18 fet at the moment is SOFT start, NOT programmable so fixed LVC alterable only via a pot".

Then you talk about the pot - which I assume only concerns LVC.

Page 13 - the "argument" and long narration where evryone figures out again the 846 chip is not programmable (in the way y'all want it).

So I read it - and again ask - how is LVC (low voltage cutoff) related to the "soft start" issue. I am only concerned about the throttle delay, and can it be corrected or do I have to replace the board. :|

Bryan
 
Deepkimchi said:
So I read it - and again ask - how is LVC (low voltage cutoff) related to the "soft start" issue. I am only concerned about the throttle delay, and can it be corrected or do I have to replace the board. :|

Bryan
Click to expand...


Nice summary.
They are not related.
nearly every single newish Infineon out there has the soft start and a throttle delay
I think that most "normal" people like the soft start (though my wife found it unsettling)

The only reason LVC came into the discussion is because the boards do not support user programability, hence the LVC pot soldered on the board.
Two totally different subjects.

Bad news for you is that you can not alter the soft start of your controller. (or any controller with soft start)
It is burned into the firmware of the chip.
Your only choice is to pick up a new board. (or enjoy the soft start)

The good news is that the boards are cheap ;)

Does that answer your question?

-methods
 
Hmm.. i'll oeder one new board for sure.. but i would really appreciate if the new board upgraded comnponents couls be just transfered to the old board smt or not!... cause my actual old board is just perfectly soldered with perfect beffy cooper braid and all mosfet soldered on it... I would appreciate if it would be possible to just take the new component appart from the new board and to replace them on the old board? :?

Doc
 
I agree docbass. I've got a lot of time soldering in many layers of solderwick over all the current traces in my board.

I have a hunch the new hardware isn't compatible though. Which means more hours of boreing soldering, but at least when we finish we will have responsive programable boards :)
 
Thanks, that answered it. And I believe correcting the problem is impossible, at least for someone like me with a soldering iron & multimeter.

That 846 has fine legs, I don't think even the above average guy could unsolder it, and replace it with an 866.

DK
 
It would be a piece of cake actually...

Mask the area around it, hit it with a heat gun, lift the chip with tweezers. Then clean up/prep the board for the new chip. If you are a pimp, you would solder paste it. If not a pimp, carefully re-tin each pad before setting the new chip down, then set the chip down, touch the outside corners to locate the chip, then touch each leg. A fine tip would be nice, but the edge of a regular tip would also do the job.

This chip and it's legs are giant in compairison with cell phone parts, which pro's can still do by hand using magnification.
 
liveforphysics said:
It would be a piece of cake actually...

Mask the area around it, hit it with a heat gun, lift the chip with tweezers. Then clean up/prep the board for the new chip. If you are a pimp, you would solder paste it. If not a pimp, carefully re-tin each pad before setting the new chip down, then set the chip down, touch the outside corners to locate the chip, then touch each leg. A fine tip would be nice, but the edge of a regular tip would also do the job.

This chip and it's legs are giant in compairison with cell phone parts, which pro's can still do by hand using magnification.
Click to expand...

That is what I do (heat gun) it's also real way of doing it...but usually it's a more complex heat gun (I used to work with that every day) with adjustable temperature and perfect adaptor for each chip packaging ....but the heat gun work perfectly if you know what you are doing!

Robin
 
hi
I hate this world time delay you all post while I am in bed.
To answer the questions raised since my last post.

Deepkimchi
"soft start" which in effect is a 1 second delay was and has always been part of the coding of the chip done at the factory by the manufacturer BEFORE Keywin got the controller boards. It is not alterable by the computer link on any of the infineon family of controllers and never was.
By programming a controller we are using the wrong term really the correct term would be:-
user alterable settings via a TTl interface linked to a com port and a computer program to controll the settings that can be changed and the values that can be used.
this "programability" worked on all the infineons keywin had got but when he got the 18 fet he had to have all the settings that we se on that program screen set by the manufacturer and it was fixed so the only way to alter LVC is via a pot, as on most of the family LVC is set by the programming tool. The two were not linked it is just that with the new boards we get "programability" back as well as hard start.

doc, methods
I have some bad news for you, the chips ar 100% incompatable for swoping, keywin sent me a pic of the new board with orders to keep it off the net for now I can tell you the new chip has legs on all four sides so will not fit the old boards.
I can tell you this when having to change fets on a 12 fet board once the mass of solder was removed from the fet removal was easy with a solder tool shaped like an S all of the fets were ok after removal they were 4310's.
You also asked about more powerfull caps I will ask keywin I'm sure we can work out something.
If the same program is used as for the rest of the family we are going to have a problem with the shunts, the pics I was sent had no shunts fitted but keywin told me the shunts come already fitted to the board so we can't ask keywin to install difrent rated shunts there, the max amps on the program we use at the moment is 57A keywin had this batch of controllers set to 65 so 57 is not the limmit, first we wait to see if a difrent program is needed for programming and what its limmits are then we will work from there, changing shunts and recalibrating the program settings as a last resort, Doc you will need to do this regardless.

One thing more about the infineon family of controllers none of the boards are designed with a particular mosfet I know putting less than a 4310 on a 18 fet would be pointless but it could have 75nf75 or worse and still work fine so going the other way with the 4110's and the 4115's will probably not cause a broblem either.

Geoff
 
Thanks Geoff for these clarifications.

I realize that Now I have a 100-150A controller able to take around 90V..... WITH SOFT START :lol: :cry: :roll: :x

I am just wondering WHY can't someone Wake up and desing a TRUE powerfull ebike controller :?

One that NOBODY could bother with these features.. LVC, Shunt max current, voltage etc , Board with enough large trace to handle the current.. no need mod it to make what we want and need...?

Having a controlelr that is STOCK 100-150A, up to 100-150V with adjustable LVC, current limit, speed limit, C-A output..etc

It just seems IMPOSSIBLE from now?

I'm sure that this product that could be offered to 200-300$ from Infineon could satisfy many persons!

Cause from now.. The hub motors are still ok.. the X5 dot really need to be modified to work as we want.. maybe just having stronger torque arm.. 1/4" steel with...

But for the conmtroller, everybody that want toi get the real capability of the X5 NEED to buy a controller and to mod it... no controlelr at decent price can do what we want from now... The kelly is just too expensive...

Doc
 
You know the answer to your gripes doc

For every 100 people who want to ride and ebike only one or two want to have crazy power
The factory is just trying to produce a product that will satisfy the greatest number of people while minimizing the cost and variation of stock.
They dont make much money off of us doc, I think they make their money off of the nOObs.

I think these new controllers coming out next week are going to be the closest we have ever come to having the ultimate inexpensive high power controller.
I have the pictures of the new board too and improvements have certainly been made. I am very excited about these new boards.

All we really have to do is solder in the fets, solder up the traces, and run wire.
That is not that bad if you make the job easy.

-methods
 
Hey doc how are you.

i wont pretend to know a lot about controller or anything but have you seen those controller from kelly maybe if you modified them with 4110 it would satisfied you ...if not well.... i tried

by the way anybody here review them
http://www.kellycontroller.com/shop/?mo ... t_id=41,39
http://www.kellycontroller.com/shop/?mod=product&cat_id=41,39&product_id=525
kbs72101 24-72 volts 100 amps 199$
 
I agree with Patrick here, I think the new 18FET boards will be pretty close to "perfect", and will still be way less expensive than anything else out there, like the crap Clyte offerings. The only thing I think would be a great mod would be to go to a higher voltage. I'd really like to be able to use two 16s5p a123 packs in series, which would be about 105v (about 117V fresh off the charger...). The Kelly 120V/75A (continuous) model is the only option for this right now, but it is $500. :shock:

-- Gary
 
If just the new boards are $ 40 each (w/o cases) then it's not too bad. I could have a back up soft start controller with my ebay 4110s (don't y'all fuss).

Real plan is to use 2 packs in my WALD panniers, and my 12ah pack on the back rack, for total of 120V, but at 50-60A with 4115 FETs.

I wrote Keywin about an exchange but no response. If it is like Korea, then no answer, changing the subject is the polite way of saying "no". :?

DK
 
You must log in or register to reply here.

Similar threads

oldswamm
A Lebowski controller in a 9 FET 'Infineon' case
Replies
3
Views
2,241
oldswamm
oldswamm
oldswamm
An economical Hi Amp DIY BLDC controller
2 3
Replies
65
Views
10,795
kevbo
K
Back
Top