Version 2 Crystalyte Controller information

[Sol]

Just wanted to add some further insights and explanations for the need of individual gate resistors for the power MOSFETs.
All the posts here are truly remarkable. So I took the opportunity to collate all the suggested changes into a single PDF along with some design reflections of my own. Please feel free to add or modify anything in the attached document.

Cheers,
Sol

 

[fechter]

Nice job and welcome to the forum. I'd like to take a summary like that and post it in the Technical Reference section. I'll see if I can port the .pdf file.

 

[Elektrofreak]

Hello,

I've read nearly 7 hours about the Controller now. I must say it hides most of it's features if you do not ask or read about it :wink: .So i will aks about a nice feature.
On my papers (pin settings) it says that you can use a kind of tachometer. Can anybody tell me how to use it?

I must tell you that this forum is exceptionally great and that you really help me in developing the hardware for our "multi-powersource-vehicle", e.g. Fuel-Cell, LiIonMn and Solar.


Thanks a lot.

Elektrofreak

 

[methods]

Just wanted to add some further insights...

Good job Sol.

When a forum gets as big as this one is starting to get we almost need someone to do a nice job like you did, then post the most current "summary" to the first post. You see this a lot on RC Groups or Reef Central where the threads regularly get to be hundreds of pages long.

I guess the counter argument would be that people should be forced to do the work to extract the information themselves thereby gaining an education in the process. You see this a lot in automotive forums where it is important for people to understand exactly *why* they are doing this or that to an engine.

In this case, the subject is already so "DIY" that if someone is looking for the information and prepared to use it then we can assume that they are the type that will educate themselves without being forced to do so.

Ok, now hurry up and get to work, you have about 50 other major threads that need servicing :wink:

-methods

 

[Doctorbass]

Welcome Sol,

That's verY generous from you to spend time like that and to share that at your first post!! Nice summary!

Maybe i should give that to John (Oyeah), he have direct contact with the crystalyte staff and he is now working on asking them to build a 18 mosfet version of the existing 12 mosfet controller for a future mooded version of the 5305... the 5303/06 serie parallel motor!!.. just like tsome NEDRA EV race use for getting more troque and then speed!

Doc

 

[markcycle]

crystalyte has a 18 Fet controller see picture

A) 18 Fet

B) 12 Fet

 

[Doctorbass]

crystalyte has a 18 Fet controller see picture

A) 18 Fet

B) 12 Fet

Hey Mark, that's great!

Does the A) 18 mosfet is made of 4110??.. and why it is only rated 48A???

the one i'm talking about will have higher current capability using 18x 4110.

Where did you got it?

Doc

 

[methods]

Hey doc, are you aware that electric rider has a new prototype X5 with disk brakes?
They said it will be available in a few months.
Custom designed through Crystalyte

I tried to talk to them about all the things we want . . . you know. . . rolled threads, better wire, built in torque arms, longer shafts, better heat management, etc.
As expected they have to cater to 90% of their customer base which are "normal people" who want to get into a powerful ebike for as cheap as possible
Cant knock that, business is business.

Ya know Doc, I know a lot of RC guys who re-wind motors. There is one fellow in particular who is "the man" not far from me. Perhaps I could convince him to rewind my X5. Do you have any sort of plans or calculations? He once rewound a 16 gram Hobby City motor for me and the difference was night and day. Only charged me $20 too! (for the motor, parts, and labor)

-methods

 

[markcycle]

crystalyte has a 18 Fet controller see picture

A) 18 Fet

B) 12 Fet

Hey Mark, that's great!

Does the A) 18 mosfet is made of 4110??.. and why it is only rated 48A???

the one i'm talking about will have higher current capability using 18x 4110.

Where did you got it?

Doc

I got the picture from Crystalyte when I was considering there controller for my motorcycle hub motor but I decided to stay with a Kelly controller

I don't know the type of Fet in the controller. I also think they derate these units because they don't use large enough wire size for the phase wires and the PC board traces can't handle the current even with wire overlay.

Mark

I don't know the type of Fet

 

[rkosiorek]

I wonder if they set it up as 2 fets on the high side and 3 on the low side?

i have seen some of those cheap scooter controllers set up as 9 fet units. ie. 1 high side fet and 2 low side fets on each phase. these ones had no labels or markings. but i was told they were supposed to be 48V 30A units. board layout and part numbers were similar to but not identical to the Crystalyte V2 controller.

rick

 

[ElectricEd]

Ever wondered how to connect the Cycle Analyst CA-DC throttle over-ride to a Crystalyte Ver2 controller?
It's simple if you have some soldering skills and good eyesight. My circuit is a little different from the Cycle Analyst manual. I found that connecting to XS+ and a diode parallel to the on-board 20K resistor R39 (as shown in some photos in this thread) disabled the controller. After a bit of tweaking I got it going. Here's how:



I have found that a 5K6 resistor works best, the throttle has full linear control from off to full on.
Make sure that that the diode direction is correct and be very careful when removing the surface mount resistor R39. Heat both sides evenly and remove using fine tweezers.

Getting this going has now made it a no brainer for me to get home with a 10AH battery for 27 Km and 1200ft of total climb. Just set Speed Limit to 35 Kmh and Max Amps to 20. Prior to getting this going I had to be careful not to floor it from each stop and not cruise at 40 Kmh or I would run out of charge.
IMHO I reckon that the CA must be the best thing since sliced bread and Vegemite. Hats off to Justin. Justin, you are the ultimate Ebike hero.

 

[rkosiorek]

back in August of 2007 Justin Lemire of e-bikes.ca and the designer of the CA posted this schematic of the mods to the V2 controller for the additional e-brake line.



the link for this post is:
http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=2038&st=0&sk=t&sd=a&hilit=rfw4850&start=15

it shows that the R39 resistor is replaced by the diode. i did this on a couple of V2 controllers and it worked fine as the e-brake signal. until yesterday i did not have a CA to play with. i have a couple of the good old reliable (read "antique") drain-brains that do not have the built in limiting functions.

so i am wondering why is the 5.6K resistor necessary? is the limiting action too harsh without it? does it work without the 5.6k resistor in series with the CA?

rick

 

[ElectricEd]

Hi Rick,
You are suggesting that it is to be used as the ebrake signal with R39 replaced by a diode. If you are using the brake lever switches on this line, you are grounding the throttle line which is the same as turning the throttle off. The correct ebrake inputs on the Crystalyte Ver2 controller are HH (switched to +5V to brake) or HL (switched to ground to brake). If the ABS line is grounded, either HH or HL when activated will also activate regen.
It may be that using the term "ebrake" here is a misnomer created by the manufacturer's "Chinglish", with Justin using it initially, but now using the correct term "throttle over-ride."

I initially used the schematic supplied in the CA manual Ver 2, but found it put the controller into an error mode and wouldn't start again without a power down reset. (On seeing your circuit, it was probably just the CA killing the SP line, if I had disconnected it, the controller would have started again). I then tried the version shown in the picture of lazarus2405's article in this thread where the R39 resistor is replaced by a diode between XS+ and the R39's inner pad.
http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=3896#p58220

That gave the same result of stopping the controller. I then set up a small break-out cable to fit between the CA's connector and the controller. A 10K trimpot set up as a reostat was placed in the CA throttle over-ride line. At zero ohms the fault remained. I increased the resistance and found that a measurement where it worked properly without killing the controller was measured at 5416 ohms. I then installed the 5K6 resistor into the throttle over-ride line, removed the break out cable and the problem was gone, the controller worked and the Cycle Analyst did everything as promised. It maybe that the CA settings for ITermMax and ITermMin are wrong for my throttle. It was a generic brand supplied by solarbbq, I don't even know if it's a Hall-effect or potentiometer type. All I know is that it works. Depending on the throttle type, the 5K6 may or not be required. I'll have to investigate this aspect futher.