Crystalyte Controllers - Repair and Modification information

[Doctorbass]

The 18k feeding the zener is a small one, not 2w. I'm not sure about the other one. I can check later. What are you going to uppgrade?

yesterday, after repairing my controller, i also mooded the small 18k because i saw that when i tested my controller on my 100V power supply, it had to dissipate 0.4W !!!! and it's a smt...... so i replaced it by a 1/2w to be sure.

Also, i added a 1.2K 2W resistor in serie to the existing 3k 2W to absorb more power when runing at 100V and avoid over heat and prematured aging of it. so at 72V that 3k should is ok so i expect that by adding that 1.2k, the ratio for 100V is the same... finally i replaced the 15V zenner by a 3W zenner of 14V... so now, the 15V regulator should be proof too.

I also connected the reverse key switch wires in serie with the power botton to add a "security" level .... because the reverse switch is not usable for me......

i added a 6A heavy diode in parallel to the Rail to avoid reverse spike that could blow the mosfet in case where the fuse would blow and that the indictive effect of the motor coil would generate too much reverse HV...

and finally replaced the 3 220uF caps by 3x 330uF, 105C, low esr serie.



now.. it should be ready for 100V .. i just hope the 4110 will accept that without blowng.. i know it's borderline..

Doc

 

[The7]

now.. it should be ready for 100V .. i just hope the 4110 will accept that without blowng.. i know it's borderline..

When the battery is increased from 65V to 100V, the maximun motor current will also increase by the same ratio if the battery current limit remains unchanged. So go figure!

 

[Doctorbass]

now.. it should be ready for 100V .. i just hope the 4110 will accept that without blowng.. i know it's borderline..

When the battery is increased from 65V to 100V, the maximun motor current will also increase by the same ratio if the battery current limit remains unchanged. So go figure!

yess, thanks for this advice!

 

[The7]

now.. it should be ready for 100V .. i just hope the 4110 will accept that without blowng.. i know it's borderline..

When the battery is increased from 65V to 100V, the maximun motor current will also increase by the same ratio if the battery current limit remains unchanged. So go figure!

yess, thanks for this advice!

One further suggestion:
If you want to keep the same max motor current (same max motor torgue) and aim only for a higher top speed when up-volting from 65 V to 100V, the battery current limit is to be reduced by an inverse ratio( say 0.65:1).
In this way, your controller will have the same degree of robustness.

 

[Dee Jay]

Long story short: I am TEH BOMB! 8) (but you know that)



My controller took a sh!t On my way to pick up my son after school, I hopped off a curb and knocked something loose, one phase wire wasn't firing. The motor was running but slower and cogging. It persisted even after I :

- switched front and rear motors back and forth, so that eliminated the possibility of motor damage

- checked the phase wire and sensor wire connections, cleaned them out

- checked my homemade switch box

I didn't want to chance damaging the controller so I turned the battery off and pedaled all the way to school on a sweltering humid evening. It's only a mile and a half but too many stop and go traffic so it's not easy for a semi-recumbent with the weight of two motors and a battery pack. I was worried even more if I wasn't able fix the problem at the school because I'd have to carry my son and his two school bags!

After picking up my son we pedaled to a nearby 7/11 and got him some pudding and water, then we rode to a park where I had him eat and play while I took my switch apart using nothing but a nail file off my nail clipper. I really should invest on a good multi-tool.

My switches and its connections seemed ok so I detached the controller and plugged directly to the motor.

Still NO GO grrrrr . . .

It had to be the controller but didn't have tools to open it up so I smacked it like a redheaded step child and . . . AND . . . IT WORKED!

I didn't want to ride slow on my NiMH pack and chance the controller acting up again so I jumped hyper-space with my A123 pack and zoomed all the way home, wee hoo! 8) My son loved it too.

The next day I bought 10 gauge wires to replace the controller's phase wires as this was something I'd been planning to do whenever the next time I needed to open up the controller. So mote it be! DONE! Unfortunately, after several hours of phase wire work and inspecting the controller (no obvious damage) I was still running on two (and what seemed like *a half*) phases.

Did some searches on ES and didn't find any posts with similar problems.

Checked the voltage on the sensor wires, and then ran the motor while monitoring the voltage (if I remember correctly) :

Yellow 14.6v - 7v
Green 1.2v - 0v
Blue 14.6v - 7v

Disconnected the sensor connector from the board and measured directly off the connector pins: same readings. The problem was obviously ON the board but nothing was burnt so it seemed to be some kind of defective chip. But I have no full understanding of how controllers actually work!

Man, I was sh!t out of luck. Karma is in full swing the opposite direction.

I looked at some schematics to get a general idea of what could possibly go wrong, at this point I was burning out so I took a few hours break, ate and chilled on the sofa . . . I was thinking it would be easier for me to see the board closer if I took a macro photo of it and zoom in on it with the laptop but my wife took the camera to my son's classmate's farewell party.

After my break, with refreshed eyes, I traced the connector pins and noticed at least two surface mounted (micro) resistor with a dull gray solder on one end and a shiny solder on the other end.



There were about 5 of these resistors located along side the connector pins and most of their solder were shiny. Both dull solder seemed ok until I noticed *upon closer inspection* a CRACK on the dull solder, thinner than human hair!

Fnck me, I thought . . .

That's gotta be the problem. With surgical precision I was in there with a soldering pen and fixed them up. Fixed the other resistors as well just to be sure and
.
.
.

~ WHAAAAAAA ~ I was up a running again!




I heard birds chirping. I stepped out into the veranda and saw the clouds parting and the sun was shining once again. I folded my arms, heaved my chest and took a sigh of contentment. I looked off into the horizon with beaming eyes, nostrils flaring and thought . . .

"Gosh, I love Me . . ."



No seriously though, Lessons Learned:

- Take breaks.

- Must have: Magnification

- Must have: Bright White Light (like my bike's LED lights) to SEE clearly to find problems on these boards.

- Must have: A multi-tool!

- Must have: A positive attitude. Narcism works too! :wink:

- Dare I say, "Crystalyte controllers are not bullet proof"? Duh...

EDIT:
- A spare controller would be nice if you can afford one



Happy Friday!

J

 

[fechter]

Excellent job. The fact that whacking it made it work was a clue.
Finding a crack like that is amost impossible since the things are so tiny. Must have been a bad solder joint at the factory.

 

[Dee Jay]

Thanks fechter!

Finding a crack like that is amost impossible since the things are so tiny.

Very true.

Plus, I was just lucky that the damage was in plain sight and accessible because it could've happened under the board or somewhere under the wires. I was not up for taking the board apart!

J

 

[Mike B]

Hi Everyone,

I have been mostly lurking aroound for the past few months. NOW I have a question.

Would it make sense to split the standard Crystalyte (etc) into two sections?

The power section would have the mosfets and all the other watchamicallits (mosfet drivers, diodes, I don't know what)--a decent heatsink and fan. This higher voltage stage would be whatever is needed for the battery/motor combo (big wires, bit fets), so in Doctorbass's situation, it would eventually reach 200 amps at 300V And when it blows up, it is easy to troubleshoot, or do the swap out test. And now it is generic, or we could just build them from parts and a simple board.

The other section is just the balance of the controller. If I am reading the posts here right, this side of things doesn't usually lose smoke. It would run cool, and would be easy to weather proof. Now, it's generic too, so ECrazyMan's people could start running these up.

I am guessing about ten people are going to point me to the section where this was already hashed out in detail.

Mike

 

[fechter]

It's not a bad idea and it has only been breifly discussed before. FrankG did something like that with a brushed controller. With proper resistor or optical isolation, any damage on the power stage could be isolated from the control electronics. You could make a modular system that way.

Right now we're having enough trouble just getting a basic controller to stay reliable. I think that idea will eventually come about, but not for a while. You're just ahead of your time.

 

[solarbbq2003]

I think mike has made a good prediction, I've been thinking along these lines with v2 controller, would be great if you could get a 12 fet board and another low voltage board to run it, maybe with some fairly easy way of connecting the two,
so opto isolating the switching is fast enough in the optic part?
that might be the only way we can get really reliable easier to fix controllers, if it was only the fets that were the parts to potentially fail, having a spare fet board would make testing high volts/amps alot easier.
wish list:
1. separate fet board, easy access to change caps/fets when testing different voltages with high amp fet connections ( wish there was some push in pull out fet system would make life much much easier)
2. v2 or similar low voltage board run on 9v battery
3. 'simple' opto coupling system between the two boards

 

[solarbbq2003]

ah what the heck while i'm dreaming, lets make it a 24 fet board

 

[Mike B]

All & Fechter & Solar BBQ,

Would the "opto bridge" be similar for the V1 and V2 controllers?

I understand the basics of how a three phase brushless motor is working, and that the controller IC determines timing and amplitude (negative and positive?) by observing the Hall effect sensors (slow) or by watching the voltage(?) coming back out of each phase (faster).

Then, the controller IC modulates the fets. How? Please point me at the best description of this on this site.

Of all the hundreds of circuit diagrams on Endless, which one(s) should I be looking at?

Here is something. I have the 36-72V 20a controller, a Sanyo NiMH pack, and the 4011, all this on a 1995 Trek 930. Last year, I noticed The Watts Up meter show the voltage going hi on descents--so I put another meter in series and backwards. I am routinely seeing 15a going from the controller to the batteries, and "regen" is in the 20% range for the commute. (am i due for some smoke?)

This makes me wonder if the "power stage" could also have a regen function that would bypass the "controller stage" altogether.

Thanks, Mike

 

[fechter]

Opto couplers are way fast enough to do motor control.

The controller uses PWM (pulse width modulation) on the high side FETs to control the speed. Basically the power is turned on and off very quickly (around 15khz). At full throttle, the power is always on. At half throttle, the power is on about 50% of the time, and so on.

If you are going downhill at a speed that is greater than the no load full throttle speed of the motor, it will regen through the body diodes in the FETs. The controller could be powered off and this would still happen.

To get regen at lower speeds requires switching the FETs in such a way that they essentially short out the motor windings. This is done at the PWM frequency and is modulated to control the amount of braking.

 

[XJOULESTER]

Hi all I have an issue with my 36-72V 48A Start Immediate Digital Controller (IRFB4110 Mosfets)

I'm running dewalt(bypassed) 2s2p for 66v.
On my 5305 26" the controller runs ok on slow accel rates, however if you demand faster acceleration it seems to trip out (no indicated fault blink) a power down/up restores the controller.

Monitoring the voltage on a scopemeter and Wattsup shows no appreciable volt drop (59v) and current 43Amax (verified by wattsup and limited by ca)

This also happens on my 408 26" motor......


Any help would be appreciated.

UpDated.... It seems that my current limit is 12 Amps... if i keep the current under 12 amps controller runs with no issues....

 

[rf]

Hi all I have an issue with my 36-72V 48A Start Immediate Digital Controller (IRFB4110 Mosfets)

I'm running dewalt(bypassed) 2s2p for 66v.
On my 5305 26" the controller runs ok on slow accel rates, however if you demand faster acceleration it seems to trip out (no indicated fault blink) a power down/up restores the controller.

Monitoring the voltage on a scopemeter and Wattsup shows no appreciable volt drop (59v) and current 43Amax (verified by wattsup and limited by ca)

I have a CycleAnalyst on a similar setup that lets you set the max amperage. When I set it to too low a number (say 20 amps) it makes for very jerky acceleration anywhere beyond about half throttle. Especially bad on hills.

Richard

 

[solarbbq2003]

i would try removing the diode shown in pic below, I haven't seen that diode on the speed control pin on any v2 controllers I've got from crystalyte, worth a shot, seem to work fine without it

 

[pwbset]

if you have a current limited power supply and a good meter, you can run an amp or two through the shunt on the controller, and you should find it is very close to a milliohm, inferred from the 1 mv/amp voltage drop across it. you can add solder until the voltage goes down to about 80% of the original value, then reprogram the cycle analyst with the new shunt value.

Is there any other way to calibrate a new Rshunt value for a CA after you've applied some solder to a couple of the legs? I don't have a current limited power supply and I'd like to modify my 72v/20a 4110 pedal first controller to 30-ish Amps, but I have no way of knowing what to calibrate my CA to afterwards (do I?). Currently my Rshunt value is 1.910 mohm. Is that even the correct setting to be modifying? I know as a noob I shouldn't even be attempting this, but we all have to start somewhere right? I'm already pushing my Milwaukee pack at 3s 84v/20amps with diodes and it's been awesome, but they are rated to 20a per pack I believe so in 2p I should be able to run them 25-30a no problem right?

Climbing a mountain every morning is still kicking my ass physically so I could you a few extra amps until I get in better shape.

 

[fechter]

Yes, the shunt resistance is how you calibrate the amp reading.

The only way to calibrate it is to either measure the shunt resistance or use a separate ammeter in series with the battery to make a measurement. Most DVM ammeter will fry at 10 amps or less, so it's better to use a meter with a higher rating.

You could possibly use a 10 amp meter and lift the wheel so the motor has no load or just use less throttle. Take a measurement with the ammeter, then note the reading on the CA at the same time. From there, you can do some math to calculate the shunt resistance. Once the CA is calibrated, the ammeter can be removed.
Perhaps you can borrow one if you don't have one.

 

[pwbset]

The only way to calibrate [the CA] it is to either measure the shunt resistance

The photo below doesn't reflect it, but I filled the space between the two shunts with solder. This is a Clyte 72v/20A controller from ebikes.ca. On my ride this morning I kept adjusting the CA Rshunt value lower and lower to try and get in the ball park of accuracy based on previous ride's data. I went from 1.910mohm, which was the stock reading as shipped to me, down to 1.1mohms after several ride adjustments and it still seems too high and isn't reporting high enough amp/hour usage and my max amp reading was 16A and should have been much higher than that during my 15% climbs.. at least double.. so I need to just take the time and calibrate a new value...

Last night though I attached a little homemade AA battery setup from a tip Justin posted and measured the two points in the photo below in mohms to get an Rshunt for my CA. Problem was that the reading skittered all over the place so I couldn't get a consistent number. .6 to 1.7 to .9 etc. etc. Should I be putting my multimeter contact points as shown below or somewhere else? Thanks for any tips.

 

[fechter]

I suppose your multimeter should go to the same spots where the CA is going to connect. It should be on the copper (scrape off some of the green stuff to get a good connection).

 

[mikereidis]

My controller is 72v, 20a, pedal first Crystalyte sold by ebikes 1 month ago. Is it v1 ? :



I want to convert this controller to instant start if possible.

The empty connector near R26 seems to be for brake switch motor cutoff. Is it ?

The 3 holes in corner of board look like they are hall sensor connections because they go to what looks like commutator chip with the correct pins. I'm hoping I can hook my hall sensors to these 3 connectors, add ground and as suitable voltage source for halls (like same source as for throttle.)

Has anybody done this or have any insights to offer ?


Thanks,... !

 

[docnjoj]

What about Knuckles/Keywin new add on board that replaces the Halls with a feedback starting method? I dont know if it can be added to Crystalyte.
otherDoc

 

[fechter]

I think he wants to do the opposite.

I've never seen one of those controllers first hand. It would be v1 of sorts.

Look on the board and see if there is a 16 pin chip marked C1246. That's the commutator chip used on the immediate start controller. If it has one, then there should be a way. I've never fully traced out the hall sensor input circuit, so I don't have a schematic of that part (it may be posted earlier in this thread?).

 

[solarbbq2003]

does that board have a manufacturer name or any markings on it?

 

[mikereidis]

What about Knuckles/Keywin new add on board that replaces the Halls with a feedback starting method? I dont know if it can be added to Crystalyte.
otherDoc

I read that thread but can't find it now. If it produces simulated hall sensor outputs, it's no good for me (unless my halls are bad.) I might as well run my halls to my hall inputs, if I need hall inputs anyway. What's that board do exactly again ? Thread ?