Instant Start 18 fet Infineon Boards are here...

[methods]

DAMN! :?

Looking at the picture I now realize that I could have scraped the ground pad and had a much better fit.
Oh well, - Next time

-methods

 

[steveo]

DAMN! :?

Looking at the picture I now realize that I could have scraped the ground pad and had a much better fit.
Oh well, - Next time

-methods

Hey methods

What is this resistor for? ... Is it nessasary for the c/a to read correctly? .. I presently have a direct plug in model.

-steveo

 

[methods]

That is a "Shunt Resistor".

A Shunt Resistor is precision resistance used for measuring current draw.
Your Direct Plugin model also uses a shunt resistor for measuring current.
The trouble is that the exact value of your shunt resistor may be off by +/-10%, 20%, or even 50%
For this reason, even though you can just plug in your CA, the current readings will be way off if you dont tell it what the shunt value is.

This is where all that "calibrating" comes in that I am always grumbling about.

The resistor I soldered in place here is 0.000250 ohms exactly (ok, +/-1%)
By using this resistor I truly can just plug in my CA, tell the CA that the shunt is 0.000250 ohms, and I am on my way.

Does this make sense?

Other factors are that this inline resistance is much lower than what you commonly find. Most controllers have more like 3.8mOhms of inline resistance.
When you run large currents (say 100A) through such a large resistance you get a huge amount of heat generated (power) and this swings the resistance value (up with heat) thereby swinging your measurement.

Square your current and multiply that times the resistance.
That is how much heat you create.
If you make your resistance 10x smaller you create 10x less heat.

-methods

 

[methods]

FYI -> I will be unavailable this coming weekend so PM's etc will get a slow response starting now.

I made up a few kits so that I can still ship same day during the wedding weekend.
Unfortunately most of them already sold today
Here is what is left:

(1) CA V2.11
(7) 18 fet controller kits
(3) 20 fet @ $2 packages

Once those are gone that is it.
I will be ordering another large batch of the CA's but those are the very last 7 controller kits I will sell (probably ). Same with the fets.
I kept a few kits for myself and I am about 80% done building up 5 of them. Those are mostly spoken for already though.
I am going to use them to trade with people for goods/services.

I have a surprise up my sleeve for when I return though.... I think I may be able to bring prices down by taking on a little risk.
We will need to cash in on our Translators for this one as my Chinese is a little rusty. It will be slow going too because I dont want to step in shit.

-methods

 

[bikeman]

I'm all new to this, but after reading the thread I'm very impressed with the work you all have done. In this thread as well in previous threads on the infineon controller, lots of thoughts was spent on the low voltage sections with different voltage dividers set up. Forgive a silly question: But why can't one just tap the batterie pack at the right places and have a separate wire feeding, for ex 6.6 V to the 7805, and 13.2V for the LMxxx for 12 V or introduce a 7812 for the 12V supply? The battery cells specs vary a lot, it seems, so it can't (i'm guessing here ) be important whether a couple of cells has to contribute with some mAmps more than the others.

I'm thinking of getting my filthy hands on the last CA and one of the kits I planning to start with 48V (yes I know now that this is the only place where that sounds insane :lol: ) with a Crystalyte dual winding 4011/408. Since 48V and 4011 is reasonably legal in Europe, and with a "cop-switch" change to the 408 which isn't legal at all @ 48V. All though I will start out (fairly) legal I know myself, It has to go faster! So starting with a badass controller like this seems like a good idea.

In europe we need pedelec as well. I haven't seen anything about it, but I've seen the Crystalyte controller with pedelec option (not sure if its the 18 pcs fet or not) Does anybody know if and how supports for pedelec would be implemented?

 

[methods]

why can't one just tap the batterie pack at the right places and have a separate wire feeding, for ex 6.6 V to the 7805, and 13.2V for the LMxxx for 12 V or introduce a 7812 for the 12V supply?

The main reason people dont tap mid pack is keeping the pack in balance.
If you tap the middle those cells will always end up low and that is a pain because then you have to equally discharge the rest of the pack at charging / balancing time.
Another issue is wiring complexity and the chance for error.

It is good that you are thinking outside of the box though. Every idea that is new is a good idea.

RE: Pedal sensor -> I have been looking at those. I dont know off the top of my head how they work though. I am sure someone around here does.

RE: Controller -> You know I love to talk my way out of a sale.... I dont mean to spill any beans but there is someone around here who will soon be creating a controller that is probably better suited to your needs. You dont need to spend all your cash on this crazy controller. A 6 fet can be had for $20 + mosfets and you can run that at 48V 65A all day. That would be my suggestion. For under $50 USD you could put down 3KW of power (which is WAY illegal ).

The controller is TINY too... like the size of a credit card.

-methods

 

[bikeman]

why can't one just tap the batterie pack at the right places and have a separate wire feeding, for ex 6.6 V to the 7805, and 13.2V for the LMxxx for 12 V or introduce a 7812 for the 12V supply?

The main reason people dont tap mid pack is keeping the pack in balance.
If you tap the middle those cells will always end up low and that is a pain because then you have to equally discharge the rest of the pack at charging / balancing time.

I do see that, the question is whether the controllers mA is detectable in this mayhem of amperage. I suspect that cells normally vary more than what the uC uses. With a BMS it shouldn't be any problem. But it is unsymmetrical and hence not aesthetic. On the other hand you don't carry around electrons with the sole target of life to make wasteful heat.

Another issue is wiring complexity and the chance for error.

True.

RE: Controller -> You know I love to talk my way out of a sale.... I dont mean to spill any beans but there is someone around here who will soon be creating a controller that is probably better suited to your needs. You dont need to spend all your cash on this crazy controller. A 6 fet can be had for $20 + mosfets and you can run that at 48V 65A all day. That would be my suggestion. For under $50 USD you could put down 3KW of power (which is WAY illegal ).

Sounds great - for the added front wheel drive when the CrazyController isn't enough!

 

[glv]

RE: Controller -> You know I love to talk my way out of a sale.... I dont mean to spill any beans but there is someone around here who will soon be creating a controller that is probably better suited to your needs. You dont need to spend all your cash on this crazy controller. A 6 fet can be had for $20 + mosfets and you can run that at 48V 65A all day. That would be my suggestion. For under $50 USD you could put down 3KW of power (which is WAY illegal ).

The controller is TINY too... like the size of a credit card.

This sounds like it would be an awesome drop-in replacement for those giant headline/cyclone controllers... rated for twice the Amps with the same number of FETs and a quarter (or less!) of the size. This wouldn't happen to be the fancy controller Justin's been working on, would it? :wink:

Incidentally, can anyone say if 65A@48V is enough juice to make one of these new sensored Astro 3210s run ok with the right gearing, or is that too wimpy? If (when) my cyclone gearbox chews itself up again I gotta replace it with something. A 3210+matt drive+methods controller (or this mysterious new one!) sounds like a nice combo, provided my V3 ping can deliver the juice (and I can find the funds!).

 

[methods]

nope. I am talking about a 6 fet Infineon with 4110 mosfets.
I have posted about it in many other threads.

You can even make a 100V version for just a few dollars more.

-methods

 

[John in CR]

nope. I am talking about a 6 fet Infineon with 4110 mosfets.
I have posted about it in many other threads.

You can even make a 100V version for just a few dollars more.

-methods

Can I place an order with you? I'll buy 5 right now if they can really take up to 100V and 60A "all day long". If it has all the programmable options too, then it should be THE ebike controller for the vast majority of users. What's the catch?

John

 

[methods]

I have only tested 12S lipo which is a nominal 44.4V, 50.4 off the charger
You can run 12S lipo "all day long" at 65A and the box gets "hot" but thats about it.
I put down about 500 hard miles on my box and never a hitch.

As far as buying one: I am dealing only in the 18 fet controllers. There is someone else who will be offering 6 fet 100V controllers. I will let him announce that if he wishes to.

If you step it up to 100V you may or may not have to run a lower current limit. I dont know because I have not done any testing myself.
Not because the rdson is creating any more heat, but because you are going to be chopping the hell out of the signal and the bypass diodes are going to be going nuts.
Any time you run a lower current limit on 100V things get hot... If people ran 100V on an ultra-low KV motor things would be fine - but everybody wants to get a 5302 and then run 100V Just makes for increased I^2*R losses due to the ripple current. I try to set up all my bikes to run with no current limit - better that way.

Anyhow - whats the catch?
The catch is that people are sleepwalking

Keywin sells the 6 fet controllers on ebay for $20 + shipping (which is probably another $20 lol...)
I sell 6 4110's for $12 + $5 shipping

Solder in the fets, run some larger gauge wire, solder the shunt, program the chip, smoke off the tires at 2.5KW
Thats the catch.

A bad ass, credit card sized, programmable, controller that can put down serious power all day long.

But who would want to do that when you can pay $250 for a piece of shit crystalyte in a shoebox that is only rated to 72V and 50A using 4310 mosfets?

btw: I dont think Justin is going to set the current limit very high on his controller. Justin is into reliability - not performance. He will build that thing to be bullet proof in the worst Canadian weather. Kind of a different market than we are in... I expect my stuff to blow up

-methods

 

[PhoenixOSU]

EXPECT it to blow up???...

I'll test the V2 and find out just what that "explosion" looks like on a 6KW kelly hub. Got the kit in record time btw, thanks!

 

[tostino]

EXPECT it to blow up???...

EXPECT it to blow up, just like you would expect a minivan to blow up if you put it in neutral, and held the pedal to the floor for an extended period of time.

Can't be on the cutting edge with performance, and have all the safety features you would expect from a production scale model. It's not like the failure would be because of bad design, it would just be pushing the limits of the parts too far.

You probably know that already, and it's too late for me to be making any intelligent discussion on here anyways.

 

[methods]

I just have it in me to push and push and push until something blows up. I always want more.

I will be surprised if someone can actually blow up one of these controllers due to running it under normal conditions.
By normal conditions I mean with a reasonable load (say 100V 100A cont), adequate airflow, and no anomalies (like a phase wires pops open at full load)

If someone does blow up one of these controllers it will likely be due to something lame like:

* Plugging in the battery backwards
* Shorting Battery voltage to some random pin
* Controller in a bag with no cooling in 100 degree heat under full load
* Phase wire pops off under 100A load and blows everything up
* Part comes lose inside of controller and shorts everything out
* Water
* Failed component (like a capacitor pops)

I really doubt anybody has a big enough pack, a big enough motor, and a long enough stretch of road to pop this thing by power draw alone.

(haha tempting fate!)

lets see... the 4110 runs up to what? 175C operating temperature? That is 350F!
When people say their controller is "HOT!!!" what they really mean is that it burnt their hand... so... maybe 150F?
I would like to be the fist member to get his controller up to 300F (lol...)

I already got my 5305 up to 420F and melted the paint right off of it....

-methods

 

[liveforphysics]

My normal pack can do 50kw. If I do my extended pack with my normal pack, I think I could peak 100kw if the connectors and 4 and 2awg leads didn't fail.


I challange you to a contest to be the first to make one fail just from over powering. The rule is that you must run the controller out in open airflow, and do none of the things you listed above to cause a failure. It's not going to be easy, and it will likely be the death of a few motors before I can get it done, but I WILL kill one.

I HAVE to push my toys to the point of failure. Even when it's a $10,000 race engine build, I will continue to push things harder until it fails. I don't trust something that I haven't destroyed. Until you find the limits, things feel unproven.

 

[methods]

Too bad that is Luke's philosophy with women too....


-methods

 

[liveforphysics]

Too bad that is Luke's philosophy with women too....


-methods

8) :lol: :wink:

 

[John in CR]

Methods,

Blowing stuff up may be great fun for those into souping up and repairing electronics, but while many of us sleepwalkers like performance as much as the next guy, we also want absolute dependability of electronic items with no moving parts to wear. My intrigue with the 6 fet controllers is the convenient size compared to the big 15 fet Infineon I have hanging on my swingarm.

Thanks for that bit about current limiting. Now I think I understand why my controllers run hotter than my motors. The motors are wound faster than 5302's, and the current is limited to probably 30A, so there must be lots of "chopping". As a practical matter, don't we want our motors to stay cool, not just to keep resistance lower in the windings, but more importantly not to kill the neo magnets by overheating? How could melting the paint on your X5 not ruin the magnets?

John

 

[Lyen]

I have a surprise up my sleeve for when I return though.... I think I may be able to bring prices down by taking on a little risk.
We will need to cash in on our Translators for this one as my Chinese is a little rusty. It will be slow going too because I dont want to step in shit.

-methods

Methods, let me know if you need help to bargain with someone from China in Chinese. My wife is one of the best negotiator I know and she helps me bargain on things all the time by making phone calls to China & using the popular Chinese instant messenger called QQ.

 

[methods]

I was thinking of you.
We may need to negotiate with the factory on a case of boards.
I am still on the fence about how to get the next shipment...

Kenny
Keywin
Direct from the factory

Build it here
Build it there

Shipping costs
Duties
Brokerage fees
Taxes
risk
time

Control
Communication
sample units
getting ripped off
counterfeit parts

So many trade offs... People think it is simple "Just order a box from china" lol..... "Have them completely built over there" lol lol lol...

The first (and toughest) order of business is not getting ripped off.
Sending a few thousand to a guy in China is always stressful..... and that is just the start. Every aspect of the transaction must be correct else I eat the cost.

Speaking of eating the cost....

If any of you sleep walkers are interested I still have a pile of "18 fet slow start boards" that I would be willing to let for next to nothing.
Comes with a populated board (with caps), extruded case, end caps, and end cap gaskets.
I suppose I could build them into kits but I am not sure it is worth the trouble.
I will probably end up using most of these slow start boards as parts boards for when things start breaking.


-methods

 

[methods]

Methods,

Blowing stuff up may be great fun for those into souping up and repairing electronics, but while many of us sleepwalkers like performance as much as the next guy, we also want absolute dependability of electronic items with no moving parts to wear.

Then you better get your wallet out :wink:

Performance + Reliability = Expensive

I know it is worn thin, but here are your three options -> pick any two:

Powerful
Reliable
Cheap

-methods

 

[methods]

I just tuned in my Magura Throttle

I was going to make a post for the "Technical Reference Area" on increasing throttle resolution so I made up this big fancy spreadsheet, calculated all the ideal values, measured all the real values, procured parts, then got slapped down.

The general idea is that the Infineon only has a throttle window where it will work:

Dead zone 0V - 1.3V
Active zone 1.3V - 3.2V
100% throttle zone 3.2V - 3.8V
Upper Dead Zone 3.8V - 5V

This is tuned to the typical hall sensor throttle.

The throttle MUST be below 1.3V for the controller to initialize
The throttle MUST stay below 3.8V else you will get throttle cut-out at WOT

So... The idea is to use series resistances before and after the throttle to tune in the maximum resolution
There is no point in having a region of your throttle that moves with no reaction out at the wheel.

wel... After all those fancy calculations...
I failed to consider the impedance that are put in parallel with the throttle when connected to the controller

Long story short, I decided to tune the lowest point of the throttle to around 1.1V and the highest to around 3.3V
After testing I found this to be too close to the edges so I ended up tuning up the high end to around 3.4 - 3.5V
This gives an area around WOT that will garantee full throttle with no chopping.

Anyhow, ends up that if you put a 1K resistor in line with the +5V and a 3K resistor inline with GND it is near perfect.
You will see:

~1.15V at the lowest setting
~3.65V at the highest setting

Not ideal, but using uber-standard values is always a good thing. Each throttle will be different as well so... Values may need to change.

IMPORTANT:

i AM INSERTING THESE RESISTORS IN BETWEEN A 3PIN MAIL AND 3PIN FEMALE CONNECTOR SO THAT IT CAN BE INSERTED AND REMOVED FROM THE CIRCUIT.

Soldering these into the perminant circuit and then finding out on the trail that you were off by 5% could mean that either your bike wont start up or it will creep at the lowest throttle point. That could be a real bummer so one would want the option of removing the resistors! Dont forget that resistance goes up with heat so you may get different results in 30F than you get in 100F.

-methods

 

[BenMoore]

I think my cue has arrived :wink: ...

• Yes, its true: the Methods Monster Controller is having puppies
  
  The birth of compact and cheaper mid power (2-5kW?), programable controllers is expected within weeks.
  With their '116' control chip and 4110 FET pedigree, the 6-transistor pups are predicted to share their fathers' 'instant start' and high voltage capabilities and run with up to 65A of continuous current.
  
  Photos and specs to follow...

At the mo I'm waiting on parts but will proceed as per methods' inspiring example by documenting my work (in a new Thread), inviting feedback and offering discount prices to forum membs.
As has been suggested, the higher voltage versions especially will need testing to determine their current limits.

Feel free to PM to express interest.

 

[liveforphysics]

If any of you sleep walkers are interested I still have a pile of "18 fet slow start boards" that I would be willing to let for next to nothing.
Comes with a populated board (with caps), extruded case, end caps, and end cap gaskets.
I suppose I could build them into kits but I am not sure it is worth the trouble.
I will probably end up using most of these slow start boards as parts boards for when things start breaking.


-methods

I don't want to be the guy to do it, cause I value my free time, but those boards would be IDEAL for hosting the brains of an RC controller to switch that lovely set of 18 fets. Solves pretty much all of the problems with getting 100v 200amp RC based controllers. Voltage insualted FET/heatink setup. Caps. Durable case. FET connection board layout. All you need to do is cut the traces from the gay slow-start chip that had controlled the FETs, wire up 3 x 2channel FET driver chips and some little voltage divider crap to power the couple mA's that the RC controller brain uses, and you've got 200amp 100V RC controllers in a neat aluminum package. Just something to think about.

 

[liveforphysics]

I think my cue has arrived :wink: ...

• Yes, its true: the Methods Monster Controller is having puppies
  
  The birth of compact and cheaper mid power (2-5kW?), programable controllers is expected within weeks.
  With their '116' control chip and 4110 FET pedigree, the 6-transistor pups are predicted to share their fathers' 'instant start' and high voltage capabilities and run with up to 65A of continuous current.
  
  Photos and specs to follow...

At the mo I'm waiting on parts but will proceed as per methods' inspiring example by documenting my work (in a new Thread), inviting feedback and offering discount prices to forum membs.
As has been suggested, the higher voltage versions especially will need testing to determine their current limits.

Feel free to PM to express interest.

Congratulations Ben! I think those little 6-fet controllers with 4110s are fantastic! Much better than so many of the bulky crude and less efficient offerings in controllers I've seen. Not suited towards the power nutters like myself, but an outstanding choice for reasonable people who are making a reasonable bike. I bet you sell a ton of them.