Crystalite controller info?

E-HP

10 GW
Joined
Nov 1, 2018
Messages
6,199
Location
USA
I've been looking for another controller that will either go with my existing motor when it gets transplanted to a new project, or to go with my new Leaf motor once I finally get around to upgrading my current bike. Similar to the current PowerVelocity controller, I wanted it to be sine wave, decent current, and a flexible voltage range. Regen and push button cruise control options are nice-to-haves.
I spotted this last night and decided to take a chance on it and picked it up for the $95 I offered:
CL 72V.jpg
Cl 72V 2.jpg
It looks to be something similar to the 18FET version here:
https://ebikedeal.com/controllers/
or here:
https://kobois.com/index.php/product/crystalyte-controllers/

Here are some questions for anyone who knows about these Crystalite controllers. I read a couple of thread on the forum already.

  • The voltage range of 48V-72V seem to be the nominal voltage ranges, so 13S to 20S. One part of the listing says max input voltage of 78V, which I think may be tied to the max regen voltage. My guess is it will go to the 84V of a 20S battery, but I plan to open up the controller and check the caps and fets unless someone confirms the ratings to be nominal battery voltages.

  • The label says it supports regen up to 78V, and I noticed that the brake switch connector has 3 conductors. I'm wondering if the controller is capable of variable regen.

  • For the thermal JST connector, I'm wondering if that's an input or output (input from thermistor in the motor, or output from one in the controller).

  • Any idea where to get the software or hardware for the USB programming connector.

  • Any idea about the phase amps the controller is capable of. I'd be moving from 70A battery/200A phase current, to 60A battery, but I usually only see 60A now under acceleration anyway.

It looks like the LVC is 40V, for the 13S voltage, and if everything is unconstrained in the default settings, the controller should work fine for my purpose, without having to access any program settings. The controller has a CA connector, so I may only need to run the CA cable, plus the on/off and 3-speed switch cables to the handlebars, so a fairly clean installation. Anyway, seemed like a decent deal for 72V 60A sine wave/FOC, so worth taking a chance on. If it's genuine and has the 4115 FETs, then that's even better. No cruise, and I don't like the CA auto-cruise, but I'll live.
 
E-HP said:
  • The label says it supports regen up to 78V, and I noticed that the brake switch connector has 3 conductors. I'm wondering if the controller is capable of variable regen.
Probably not, but if it is true FOC, then it might be. However, it probably does this via the throttle like slip regen (if at a high throttle position, moving throttle to lower position causes braking to slow to that level; letting go of throttle completely doesn't brake to a stop; you'd turn it down slowly enough to cause braking effect instead), or done like the Grin method of engaging brake lever and then the throttle controls braking amount instead of motor power amount.

If the brake connector reads 0v on both the signal wire and the ground wire, but 5v on the other wire, then it is more likely to be a throttle-style input for variable regen, and you can connect a throttle to it to see how it behaves.

If it reads 5v on two wires and 0v on the ground wire, then it is probably just a switch input, and is just capable of powering a hall-sensor type of brake switch (not common, but they're out there, like the HWBS inline-cable type, or the ones that came with my Fusin kits, etc). These do the same thing most two wire brakes do, which is to just ground the signal line to activate the brake. Easy to test. You can also put a potentiometr between the signal and ground and see if it reacts to variable input, but if it's 5v (rather than 0v) on the signal that's unlikely.



  • For the thermal JST connector, I'm wondering if that's an input or output (input from thermistor in the motor, or output from one in the controller).
Almost certainly an input from the motor. There's no reason to output anything from the controller, as the controller itself would be able to do it's own limiting / protection if it has it's own internal temperature sensor.

If it's been modified by a previous owner to work with the Cycle Analyst, etc., then it might have an independent thermistor, but you can trace that out when you open it to check FETs and caps.


  • Any idea where to get the software or hardware for the USB programming connector.
If there is any, it would have to come from Crystalyte. If they don't have that on their website, you'd need to contact them. If they dont' respond, you could check with Crystalyte dealers (if there are any) to see if they have it. (I've never run across any myself).



  • Any idea about the phase amps the controller is capable of. I'd be moving from 70A battery/200A phase current, to 60A battery, but I usually only see 60A now under acceleration anyway.
My guess is it's something around the same ratio, but unless Crystalyte has the specs, you might only be certain if there is setup software for it and it has a setting for that you can access.

True FOC controllers should be monitoring actual phase current, so there should be current sensors on at least two of the phases (preferably all three), so if you can find out what those sensors are, and the scale of their output, you can make a monitoring system to actually read what the real phase current is, realtime. :)
 
amberwolf said:
If it's been modified by a previous owner to work with the Cycle Analyst, etc., then it might have an independent thermistor, but you can trace that out when you open it to check FETs and caps.

Thanks, great info.

Based on the label and the similar 18FET controller from the link, it looks to come stock with the CA connector, so now I'm wondering if I connect the thermistor from the motor to this input, if it would feed the info to the CA though the data conductors, of if I need to wire to the CA separately via the temp sensor plug.
 
There is no data communication of any kind between the CA and any controller. (not in the sense that KT / etc types communicate with their displays)

The only connection between them in the common 6pin JST-SM "CA connector" are the two shunt wires from controller to CA, battery positive and negative from controller to CA, a throttle signal from the CA to controller, and a hall signal for speedometer from controller to CA. No "data conductors" exist.

In the phaserunner/baserunner waterproof wiring for the CA, there is also a thermal sensor wire, IIRC, but that isn't present in the JST CA connector that you would typically find on "CA compatible" controllers.

If you prefer the CA to do the monitoring and rollback based on motor temperature, and don't want the controller to monitor motor temperatures and do it's own rollback per however it is designed and setup, you could cut the thermal sensor wire from the motor and connect it instead to the CA's temperature sensor input (assuming your motor has a compatible sensor).

I don't know if the electrical design of the controller's temperature sensor input is compatible with teh way the CA's temperature sensor input works, or if it will screw up the readings for both of them (or even damage one or the other, though that is unlikely), so unless you know it will work to hook it up to both, it is safer to just use it for one of them. If you need sensing for both, it would be safest to isntall separate sensors in the motor for each.


BTW, there are series of Crystalyte controllers that use their *own* display system (like the KT, etc type controllers do) that use a connector that *looks* like a CA connector, but is not, and is not wired the same. I think they even called it a similar name, at least when they started out using them--don't know what they do nowadays. Not knowing what voltages are present on the Crystalyte display connector and on what pins, I don't know what would happen if you plugged a CA into it--you could fry the CA if battery voltage is present on the wrong pins. :(

So be very sure that it really is a CA connector (and not that it is just labelled as such), by verifying the actual wiring inside the controller, before you plug one in.



EDIT: found that they now call it "Crystalyte APM Display":
http://www.crystalyte.com/Crystalyte%20controllers.htm
 
I went through a bunch of Xlyte controllers, I found them to be less reliable than grintech generally, sometimes the caps blew after a month of moderate use, all the Xlyte lasted a relatively short time.

I'd stay within spec, the Xlyte company is very responsive AFAIK, you can write to them for spec pdf's and specific questions, they have multiple knowledgeable dealers and multiple business email addresses.
 
zzoing said:
I went through a bunch of Xlyte controllers, I found them to be less reliable than grintech generally, sometimes the caps blew after a month of moderate use, all the Xlyte lasted a relatively short time.

I'd stay within spec, the Xlyte company is very responsive AFAIK, you can write to them for spec pdf's and specific questions, they have multiple knowledgeable dealers and multiple business email addresses.
Thanks for the insights. Were you running at the higher end of the voltage range when you had the failures, or pushing beyond the current rating?
I don't ever run high power continuously, but only in bursts.
 
I was always well within the voltage specification. The time the caps blew after 1-2 months was because of low quality caps, it was my 3rd/4th xlyte in about 2013. Their hubs used to have the reputation of being slightly out of alignment, and the hub would go up and down by 1mm as well as going round and round, so eventually the aluminium could fail and lock the bike at traffic lights, whereas today their motors are very good afaik, so perhaps it's the same for the motors.
 
zzoing said:
I was always well within the voltage specification. The time the caps blew after 1-2 months was because of low quality caps, it was my 3rd/4th xlyte in about 2013. Their hubs used to have the reputation of being slightly out of alignment, and the hub would go up and down by 1mm as well as going round and round, so eventually the aluminium could fail and lock the bike at traffic lights, whereas today their motors are very good afaik, so perhaps it's the same for the motors.

OK thanks. I'll open it up when it arrives and take a look at the components. Might be better to replace the caps then, rather than wait for them to fail. I'm not in a hurry, since it was an impulse buy anyway. If it delivers anything close to the rated power for a minute at a time, and 35ish continuously, I'll be happy.
What's programmable via the USB?
 
That was ages ago that I switched away from Xlyte!!! as soon as something as good was available. Perhaps they use high quality caps these days!?. I'd write to them, it's a nice company, I figure they are trying to put quality gear these days! Some folk on here will know what their reputation is, I'd do a new question "Are Xlyte Controllers reliable?"
 
The controller arrived and the first thing I noticed is how light it is, and smaller than expected. My 18FET KT controller is a little bigger and has more mass, and my 18FET PowerVelocity controller is much more hefty.

Opened it up to take a look. Silicone seals on the covers, plenty of thermal grease, but pretty sparse inside.
Controller 1.jpg

Looks like it's set up to handle the rated current
Controller 2.jpg

The larger caps are 100V, and smaller are 50V, including one tiny one on the back of the board. No bulging or signs of problems. The listing did say it's a New* unit.
Caps.jpg

The "Thermal" JST is actually on the motor harness, so more likely an input from the motor thermistor
therm 2.JPG

In the controller it terminates on the main board. This makes me think the CA connector may be capable of transmitting temp signal, perhaps another option to the thermal JST coming out of the CA. CA docs don't say it can, but this controller is supposed to only work with a CA. Looking at the circuits on the back of the board, one wire seems to go to ground and the other over to the MCU.
therm 1.JPG

Also, it looks almost like there's a thermistor attached on one of the FETs, but hard to really tell. The FETs themselves don't seem to have any markings, at least looking at them using the magnifying app with light on my phone.
FET Therm.jpg

To me, it looks like it could be a decent unit, but wonder if the light weight will be a factor, since there's less mass (or fins) for heat dissipation. I ordered some XT150 connectors for the phase wires so I can test it out, and will probably replace the Andersons that are on my PV controller so I can swap it back later if I fry this one.

I remembered that I have my extra triangle pack in my living room, for power outage battery backup. Since both have XT90 connectors, I could test, and plugged it in. It beeps on start up and the LED flashes. Disconnecting the 2 pin jumper on the "Switched" JST and it turns off, so that's where the one off button will connect to.

The only thing I'll need to deal with is the hall sensor cable, which has a HiGo connector, so even before deciding which connector to change out, I need to determine the pinouts. I can access the area of the board just by sliding it out slightly and not fully disassembling, so I can test the pins. I recall seeing the board is labeled for those pins. There's a separate throttle cable which is also HiGo, which makes me wonder if this controller is of the same vintage as the CA 2.4, so the connector carries a throttle override signal. If so, I think I may need to add a diode or 1k ohm resistor to the controller board so it works as a throttle input. :shock:
 
E-HP said:
The controller arrived and the first thing I noticed is how light it is, and smaller than expected. My 18FET KT controller is a little bigger and has more mass, and my 18FET PowerVelocity controller is much more hefty.
The heavier ones may use a denser aluminum for the case and FET moutning bar (which AFAICR will generally conduct the heat away better), and / or busbars of some type on the battery/FET busses, or even a heavier-copper PCB. This one has at least thin busbars on the battery busses to teh FETs, visible in the pic of the bottom of the PCB. They look like copper; cant' tell thickness. 1mm?

plenty of thermal grease,
That's actually "bad". You only want exactly enough to just barely fill the scratches in the mating surfaces, so that the metal itself makes all contact possible, and the thermal grease only fills places where air would have been (because air is even worse at conducting heat than thermal grease is).

If the mating surfaces are so bent or distorted that they require visible thermal paste to make a heat path between them, rather than metal to metal contact, that's not very good. You can improve it by lapping them together but it is probably not worth the effort for a "little" controller like this. ;) For a hefty controller for multi-kW applications, it might be.


The larger caps are 100V, and smaller are 50V, including one tiny one on the back of the board. No bulging or signs of problems. The listing did say it's a New* unit.
Never heard of that "Chanox" brand, or "Huahong"; if you do have problems I recommend replacing with Rubicon or Panasonic, from a distributor like Mouser, Digikey, Farnell, etc. (ebay, and the like have plenty of fake/remarked/defective/scrapped parts of all kinds, and without a test lab you can't know if the parts you get are the parts they say they are). Get the Low ESR versions of whatever you do get, in the higher temperature ratings (at least 105C).




In the controller it terminates on the main board. This makes me think the CA connector may be capable of transmitting temp signal, perhaps another option to the thermal JST coming out of the CA. CA docs don't say it can,
I've never seen anything (at least in non-Grin-Tech-designed hardware) that can pass a thermal signal thru a "CA connector".

If the CA connector is a 6 pin (usually JST, nowadays I think they use Higo on the WP models of CA), then it doesn't have a wire for any temperature sensors. It only has shunt + and -, Battery + and -, Speed (one of the motor hall sensors), all five come out of the controller to the CA, and the only input to the controller from the CA is the Throttle wire.

The CA's temperature input is a separate connector.

Looking at the circuits on the back of the board, one wire seems to go to ground and the other over to the MCU.
Most likley the controller monitors motor temperature so it doesn't damage the motor (same reason you can do it with the CA). You'd have to check the setup program for it, or it's specs, from Crystalyte. If it doesn't have a setup program, you would probably ahve to have one of Crystalyte's APM displays to change it's settings.

but this controller is supposed to only work with a CA.
SInce the CA doesnt' do anything required by any controller, you can use the controller by itself just fine. If it is "made to work with a CA" it just means it *can* work with a CA, in that it would have the CA-specific-wiring of the signals I listed prevously on a connector the CA can connect to, so you don't have to use the external shunt and speed sensor on the CA and separate throttle connector from CA to controller. (which is what the CA uses when there's no CA-specific connector).


Also, it looks almost like there's a thermistor attached on one of the FETs, but hard to really tell.
Certainly possible; that would only be read by the controller for it's own purposes, and not be available for any outside uses.
There's a separate throttle cable which is also HiGo, which makes me wonder if this controller is of the same vintage as the CA 2.4, so the connector carries a throttle override signal. If so, I think I may need to add a diode or 1k ohm resistor to the controller board so it works as a throttle input. :shock:
If it's new enough to use higo connectors, it's unlikely to be a CA2.4 wiring setup; that was quite a long time ago. ;) If it was that old, it would probably have all JST connectors, and Andersons on the phases and battery (all the ones I've had did).
 
Thanks for the great info!

amberwolf said:
That's actually "bad". You only want exactly enough to just barely fill the scratches in the mating surfaces, so that the metal itself makes all contact possible, and the thermal grease only fills places where air would have been (because air is even worse at conducting heat than thermal grease is).

Well when I say plenty, I mean enough, but not gobs of excess.

amberwolf said:
Never heard of that "Chanox" brand, or "Huahong"; if you do have problems I recommend replacing with Rubicon or Panasonic, from a distributor like Mouser, Digikey, Farnell, etc. (ebay, and the like have plenty of fake/remarked/defective/scrapped parts of all kinds, and without a test lab you can't know if the parts you get are the parts they say they are). Get the Low ESR versions of whatever you do get, in the higher temperature ratings (at least 105C).

I have a small assortment of Ribicon caps in my parts box, so I may have a match, but if I don't I'll order replacements for at least the two big caps.

amberwolf said:
If the CA connector is a 6 pin (usually JST, nowadays I think they use Higo on the WP models of CA), then it doesn't have a wire for any temperature sensors. It only has shunt + and -, Battery + and -, Speed (one of the motor hall sensors), all five come out of the controller to the CA, and the only input to the controller from the CA is the Throttle wire.

Ya, I've already decided I'll just split off the thermistor wire from the hall sensor wires, and connect it to the thermistor input of the CA. I'll need to do some testing to make sure it works, since I only have the single white wire, and hoping the CA works with a common ground and white wire for the thermistor input.

amberwolf said:
If it's new enough to use higo connectors, it's unlikely to be a CA2.4 wiring setup; that was quite a long time ago. ;) If it was that old, it would probably have all JST connectors, and Andersons on the phases and battery (all the ones I've had did).

Thanks, that's promising, and less work!
 
E-HP said:
Well when I say plenty, I mean enough, but not gobs of excess.

OK, that's good.

Just remember that if you can see the thermal paste clearly, especially as a solid color, rather than the metal surfaces, it's creating a thermal barrier between the metal (compared to metal-to-metal contact), especially as it dries out over time and leaves an air gap. It's better than air, but not nearly as good as direct metal contact over the full surface area.

Sometimes this is the only way it can be done when the mating surfaces are not correctly machined to mate directly with each other across their entire surface area, and is pretty common with FET mounting bars bolted to the inside of controller cases. :/



Ya, I've already decided I'll just split off the thermistor wire from the hall sensor wires, and connect it to the thermistor input of the CA. I'll need to do some testing to make sure it works, since I only have the single white wire, and hoping the CA works with a common ground and white wire for the thermistor input.
Thermistors aren't polarized, so it doens't matter which way you hook it up.

As long as the controller will work without a motor thermal sensor (probably will), that should be fine. If it doesn't, you could fake it by just connecting a bare thermal sensor to the controller connector for that. :) It will always read ambient, but at least hte ocntroller will operate.
 
amberwolf said:
Thermistors aren't polarized, so it doens't matter which way you hook it up.

As long as the controller will work without a motor thermal sensor (probably will), that should be fine. If it doesn't, you could fake it by just connecting a bare thermal sensor to the controller connector for that. :) It will always read ambient, but at least hte ocntroller will operate.

I was concerned about the CA end, since one of the two thermistor leads would be connecting to ground. But thank goodness that Grin's documentation is great, since the pinout diagram alleviates any concerns. My white wire needs to go to the CA yellow, done!
Temp Sensor JST.jpg

I noticed the thermistor behavior is that the resistance goes down as it heats up, so I'm guessing an open circuit looks like a low temperature, so guessing if the controller itself can provide thermal protection for the motor, an open circuit wouldn't trigger rollback. Just guessing; something to test later. Now I just have to see if I still have some 2 pin JSTs.
 
E-HP said:
I was concerned about the CA end, since one of the two thermistor leads would be connecting to ground.

Oh, yeah, I forgot that most motors use the hall ground for one lead of the temperature sensor. :oops:

FWIW, some controllers, like the SFOC5, expect the sensor's non-signal lead to be on the hall 5v instead, which caused me some problems when testing that controller. But I don't know of any specific motors that actually come wired that way (only with common ground, or completely independent wiring).


I noticed the thermistor behavior is that the resistance goes down as it heats up, so I'm guessing an open circuit looks like a low temperature, so guessing if the controller itself can provide thermal protection for the motor, an open circuit wouldn't trigger rollback. Just guessing; something to test later.

If that thermistor does that, then its' called an NTC, if it helps. (negative temperature coefficient). PTCs go up with temp. Note that there are many types of each, the most common are called 10K, meaning they are around 10kohm at room temperature (72F, 25C). Each has a "Beta" meaning how many ohms they go up or down per degree C of temperature above or below that. I forget what the default in the CA is, but if your motor's sensor is different, you can calibrate the CA for it.
 
amberwolf said:
If it's new enough to use higo connectors, it's unlikely to be a CA2.4 wiring setup; that was quite a long time ago. ;) If it was that old, it would probably have all JST connectors, and Andersons on the phases and battery (all the ones I've had did).

I finally reopened the controller to trace the CA connector wires and throttle wires. They use some odd color combos. The throttle cable uses white to ground and black as the throttle signal, for instance, which caused some confusion, making me triple check.

Even with the couple of HiGo connectors, the controller still seems to be on the cusp when it comes to when it was made. The clue I see that there's a 1K resistor in series with the CA throttle wire and the throttle input on the board, which I believe was needed for pre CA3, when adding a CA connector to a controller. Either it's pre-CA3, or the resistor is there for backward compatibility. I'm thinking of bypassing the resistor and see if it works that way or if it makes any difference.

The resistor from the CA connector terminated on the SP point on the board. Interestingly, the throttle cable (black) that terminates at the same point, also appears to have a resistor in series, although I haven't peeled off the shrink wrap to see what resistance is there yet.
throt.jpg
 
Back
Top