Custom KMX recumbent trike with QS205

tonystark20 said:
EDIT: also wanted to say that I get tones of compliments on the suspension. and I'll admit I like it too. (and I most certainly give credit to you when people ask about it... just saying.) I also notice with my bike being in the sun a lot the anodized coating has taken on a copper-ish color. which I think looks cool. and the hex bolts haven't rusted anywhere near what i thought they might. (easy enough to clean and lubricate to prevent further rusting)

Thanks for the kind words :D
 
Sounds like the BMS in the battery shutdown to protect it from something, real or imaginary (defect). Could be a broken sense wire to a cell; cheap BMSs don't usually detect that but a good one should. Could also be overcurrent if something shorted inside the controller (but that would probably lock up your wheel, as it would probably be FETs in a phase to get that much current).

tonystark20 said:
The 5v line coming out shows the power going up and down with the blinking of the power light.

That probably means it is charging up with the leakage current from the battery's BMS in BMS shutdown mode, then it reaches the point it tries to activate, which drains the charge, and repeats the cycle.

If you have a power supply or other battery of enough voltage, you can at least see if the controller will power on.

You can also see if the battery reactivates it's output if the controller is disconnected from it.
 
The cycle analyst will boot up normally and stay stable during my testing. The software of the BMS also has no log of shut down except for an over charge condition. (Which I know is from driving downhill on one of my trips (after I'd just charged) by all accounts the BMS appears to not be the cause.

But thanks for the suggestion
 
I really didn't want to wait to hear back about the warranty... I have an event Saturday and next Thursday and next Saturday/ Sunday that I really wanted to drive to. So I took the controller apart hoping it would be a simple capacitor that was burned out that could be easily replaced...

Turns out 4 of the mosfets are completely destroyed.

Lets see how quickly I can get replacements...

EDIT: me being impatient I was looking online the mosfet currently in mine is the TDM31026... which only comes from one company in china, no ebay, no nothing. could I swap all 24 mosfets in my controller for the IRFB4110. The specs seem better on the 4110 in terms of dissipation and current handling but everything else appears the same for Rds on and such.
 

Attachments

  • burned mosfets.jpg
    burned mosfets.jpg
    141 KB · Views: 1,858
  • heatsink.jpg
    heatsink.jpg
    104.5 KB · Views: 1,858
  • main board.jpg
    main board.jpg
    136.8 KB · Views: 1,858
  • mosfets.jpg
    mosfets.jpg
    118.6 KB · Views: 1,858
That's some pretty extensive explodies.

is that just one phase? I wonder what caused it?

As long as things like the gate charge value (Qg, I think) and such are the same, or at least similar, the FETs will be able to be driven by the existing electronics.

YOu may have to also replace the gate drivers--they are often taken out by FET failures, especially ones as dramatic as that.


tonystark20 said:
The cycle analyst will boot up normally and stay stable during my testing. The software of the BMS also has no log of shut down except for an over charge condition.
Ok, just that you said the CA shut off during the incident, whcih means power was cut to it. That means the battery was disconnected from it, and that usually means the BMS shut off.

There must be something different about how your setup is wired; something that disconnected the power to the CA. If the CA is powered by the controller's output that is only activated when the controller is actively operating, that would do it, but that would be unusual in itself. (most controllers have no battery-level output that they control that way.
 
amberwolf said:
That's some pretty extensive explodies.

I'd say the cost of the 24 replacements of 4110 cost me 115 (half what I paid for the controller) will be delivered tomorrow.
amberwolf said:
is that just one phase? I wonder what caused it?

Yep, just the one phase. Ground side. From what I gather on these forums is that sabvoton stop being made in 2014, anything after is a knock off... and apparently these FETs are the know weakness. One or 2 sets will blow. And from what I saw the high the amps your running (even if within spec) the faster they burn out.

amberwolf said:
You may have to also replace the gate drivers--they are often taken out by FET failures, especially ones as dramatic as that.

as far as I can tell everything else is in great condition. I'll find out when I power it up again after replacing parts. unfortunately all my equipment is in storage so I don't really have any way to test parts. (been moving too much since college and haven't had a permanent place yet)

amberwolf said:
Ok, just that you said the CA shut off during the incident, whcih means power was cut to it. That means the battery was disconnected from it, and that usually means the BMS shut off.

The display did turn off, (its wired inline to the power going to the controller) my guess is that these FETs which were the ground side had failed to ground (they are on the ground side) which triggered the BMS and as I was trying to accelerate it likely dumped voltage back to the charged battery triggering overcharge condition on the BMS. (if that makes sense. but after a reset the display would come on, the controller wouldn't.

Either way thanks for the information. shall be interesting if new FETs will fix and sustain the issue. just glad it didn't break on one of my long trips. (longest has been 200km one way) in this case I was a 5 minute push back to home.
 
So i replaced all the mosfets, that was tedious as my soldering iron was having difficulty keeping up (its 60w) but the heat kept being pulled away from the thick power traces. I also think it would have been easier if i had a de-soldering station or wick. but alas I did complete the task and it wasn't a half bad result. assembled controller plugged it in (power wires only and turn it on...

Nope same blinking red light. took it apart to spot any issues... nothing visual no blow ups. I noticed the usb power has the power wire connected so I used my programmer (ch370 that I use for my arduino mini's) and was able to have it turn on, so I could actually communicate. went back to wire it to the trike.

Was able to see all the information when connected and powering the 5v externally. I tried to hit the throttle to see if it would drive. But no luck powering it with 5v externally (with and without com port attached) My next issue is to figure out why the power supply isn't working but that is a task for another day. rally wish I had the budget to buy another controller. will likely buy an adaptto controller. but don't have the budget to really buy anything right now.
 
I'm not sure Adappto is a good choice; what I have seen here on the forum is that they may no longer be around other than sometimes eventually repairing stuff sent back to them, not making new stuff anymore. Some people have had a hard time getting stuff back from them after sending it back for epair.

You'd have to check that out yourself, before buying, to be safe.



For the power supply problem, if a gate driver did blow up, it could be shorting the supply out, keepng it from starting up (if it's an SMPS). Find the power supply pin to the gate driver chip, and disconnect that. If the PS comes back then the gate driver chip is toast.
 
Thanks for the information about adaptto. It would explain why I'm having trouble finding retailers with stock, most either are sold out or have the listing removed. That's a bit of a shame. I like the idea of the BMS being linked as well as the mode where you tell it how far you need to travel and it keeps the watt hours where it needs to be. I only wish it had reverse. plus the cost was up there. But the weight and space savings would be nice. especially for my next design. Guess I might start looking at open source projects and adding my own flair to some designs.

I finally got an email back from QS. they wanted me to ensure battery was charged (it is) and to power it via ground and ignition only. which confirmed what I already knew. (computer / power supply is bad) I keep talking to them maybe I'll get lucky and they can send me one. if they want it sent back well that not happening. its far to expensive to ship back. (that and I've already tried fixing it, and it'll be noticeable that its been modified.

On to the controller:

the orange arrows are pointing to the IR2110 gate drivers. There are not shorts on any of the chips. they all display ~180ohms.
the green is the two AMS1117 one is 3.3v the other 1.8v. They don't indicate any shorts either. upon power up these show 1.5v incoming. (fluctuates between 1-1.5v) so the problem is upstream.
I'll have to take the heat sink off to get at parts on the bottom. None of the parts on the top get warm when powered.
 
update. took the heat sink back off to see the underside.

I found a diode on the transformer that was reading wrong (not blocking voltage) so i removed it and powered the controller. it turned on. put a replacement diode and powered it... and it wouldn't turn on. So i removed the diode and started testing. as far as I can tell there is a short to ground when the diode is connected. my best guess is the short is thru the transformer itself. I checked the original diode and its still in working order, so the issue i had detected before removing it was the short not the diode itself. I'm temped to assemble the heat sink again with the diode off and see if it will power the motor with the diode removed. but I'm unsure the part of the supply the diode was for.

diode is near the yellow spot (removed in photo has silk screen diode symbol)

EDIT: assembled the heat sink powered the motor, no luck. but I was probing the pins. where the diode is, it is supposed to provide power to the gate drivers. so I definitely need to figure out why there's a short but as far I can tell the drivers aren't shorted but one of them must be, or some other component on the line is.
 
Hello,

thanks for the Controller pics.

-your Trike is great
-I use into my new Trike a Nucular 12FET Controller, have a look to his new Controllers, very easy to programm and robust
-https://endless-sphere.com/forums/viewtopic.php?f=31&t=75494

BR
 
ecross said:
I use into my new Trike a Nucular 12FET Controller, have a look to his new Controllers, very easy to programm and robust

Thanks for the information.

amberwolf said:
if a gate driver did blow up

Once I found the pin on the transformer taht was shorted, I removed a diode. that diode was apart of the power supply for those gates. Using a boost converter I had on hand i attached it to the functional 5v rail to boost to the 12v for the chips. sure enough the chip with the blown mosfets went hot right away. I'm gonna order the chip and test it.

As much as i really should stop trying to fix this controller I do find its a great test for me to analyst and understand something.

As far as getting another controller I'm not sure what I'll do. I'm considering building my own controller as There are a bunch of features I wouldn't mind adding to mine. The only thing I'm having trouble finding is how to program the FOC control. I know how to sequence the phases with the hall effects and do simple control but not FOC. also how the heck does the controller limit current. (does it turn off the mosfets sooner?) If i build a controller I'll likely turn it into it's own thread.

I tried to weight the bike with a bluetooth bathroom scale I have but was getting wacky readings. If I get it running next week I'll take it to the scales. but I was able to measure each wheel individually. The front was showing 26kg each side. and 63kg on the rear tire. When I start doing upgrades I would like to know the weight so that hopefully the upgrades will removed weight. this will also help me in my long term quest to build a light weight long distance trike. (know what weight and efficiency i need for it to work. I'm also going to use paper mache to mock up a room design and fender covers (which would be molds for fiberglass or carbon fiber a later time) I've also been looking at flexible solar panels. (I'm thinking to try and get as much on the bike as I can, ideally I'd need 900watts to run without battery (which would equal 30wh/km) but with inefficiencies I know I'd need more solar than that. I'd like to aim with a system reaching 2kw maybe. but I know I'll have to crunch numbers and design stuff.
 
tonystark20 said:
I'm considering building my own controller as There are a bunch of features I wouldn't mind adding to mine. The only thing I'm having trouble finding is how to program the FOC control.

You could build one that uses Lebowski's firmware (preprogrammed onto an MCU shipped to you). There's a number of threads about different versions of that, different powerstage designs, etc., if you're interested.

https://endless-sphere.com/forums/search.php?keywords=lebowski&terms=all&author=&sc=1&sf=titleonly&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search

There's also the KT series controllers, that have the Casainho / et.al. open source firmware for the controller and for the LCD.
 
thanks for the information amberwolf. Definitely looks like I'll have to put in a bunch of research.

I received more parts from digikey. I powered up the controller... hit the throttle and the motor moved a few degrees, then stopped and the mosfets went bang. The exact mosfets as before too.

at this point the circuit board is starting to get some heavy damage from the mosfets blowing up so i'm going to stop trying to repair it. I was looking at the goldenmotor controller (I've had great success with their products in the past https://www.goldenmotor.bike/produc...volt-brushless-motor-controller/?currency=CAD

a little on the pricey side but it has screw terminals which would save me from doing a bunch of work to adapt the current connections on the motor.

I was probing the 3 phase wires on the motor, each phase was giving me about .5 ohms is this normal?
 
It's just a bunch of wire, all connected to each other inside the motor, so a regular ohmmeter or multimeter gives no useful reading--it's a dead short as far as it's concerned. ;)

To get a really useful reading you'd have to use something that reads very very low inductance (microhenries) accurately, and something that reads very very low resistance (milliohms) accurately.

The only reading a regular meter can make is to see if the phases are completely shorted to the stator (measure from axle to any phase without opening motor, or to stator laminations if you have it open), and if there are any open phases (measure from each phase lead to each other one).
 
so I purchased the vector 200 from goldenmotor.

The controller overall is quite nice, but the acceleration is terrible. no matter what settings I change I can't make it go faster. I feel as if it's programmed to avoid hard acceleration. which is annoying. it does work and once at speed It works nicely. But acceleration it goes .5kw then after a little while goes up to 1kw, then goes to full power 3kw. but it takes a good 5 seconds to get there. I plan to go driving tonight to see how it performs during normal driving and on hills. I'll likely keep the controller but will have to try another controller for long term use. But I do want to experiment with the astro 3220 motors, and the vector controller is good for 6000rpm, and if I'm not mistaken it can do sensored and sensor-less. (at least earlier versions could do sensor-less)

The long term goal is going to be to have the trike drive-able, get the actual weight then spend the winter doing changes and reducing weight. Then I'll be able to have an idea of the power and weight that I want to have in order to design version 3 of my trike, Ideally the final version would be light weight (maybe 2 seater) and have solar and able to drive long distance (thinking cross Canada camping)
 
Picked up this bad boy today.

ASI BAC4000 - this thing should have no problem powering the trike, or even the high power mid-drive system I want to experiment with later in the long future.

ASI BAC4000.jpg


I also plan to go through the trike and make changes to the wiring to hopefully make it more reliable long term.

Had a few ideas with the cycle analyst display, I've already tested that I'm able to disable the regular back light and can use neopixels to shine in instead - this does however mean that it won't fit in the regular case, but i plan to make a custom shroud anyways to hold and extra front light (not that i need extra lights)

https://endless-sphere.com/forums/viewtopic.php?f=6&t=94721

after looking tah that amazing thread, I want to model a front cycle analyst mount to be similar in style, and add the extra row of leds as well. looking tah the arduino the poster added inline to the data logger, I tested with mine and realized I could get enough dtaa from the cycle analyst to possible design code that could tell me if i'm driving too lead footed to make a possible distance that i tell it i want to make, the other possibility would be to use a DAC to also give it throttle override to try and make a certain distance to a destination. either way it'll be fun to prototype and test.
 
The Toecutter said:
How much does this controller weigh?

according to my fairly decent scale its 1432 grams.
Which is lighter than the Sabvoton 72150 and its smaller in dimension too (~6.75" x ~5" x ~1.5" size of the ASI BAC400)

I'm looking forward to digging into the software and playing with the settings. In the future I also want to play around with some other motors so this thing gives lots of room to play and I'm hoping it
won't break in 1200km or 3 months of driving.
 
So the new controller is awesome, managed to get it working after having a session with ASI engineer...

I like that the controller has so many options and parameters that can be adjusted but the software should be easier to set up and understand instead of having to use an engineer. (thankfully the engineering time was paid for with the controller package)

However today... I managed to shear the axle...

and no I wasn't over running anything.

broken axle.jpg
 
I had the controller set to 250 phase amps and 3300 watts normal. the brake was set to 50% of motor current and disables below 7.5km speed (the brake was tough to set and i was still playing with it because on flat ground it would lock up really easily but on a down hill in traffic where the regen can really save wear and tear on the brakes the limit was way to weak.

I'm hoping that with having the v3 motor that maybe the axle is a separate piece that can be swapped out, but i'll have to open it to check.
 
I wonder if the axle slots went up and down would be technically stronger? Seems like there could possibly be more flex with the way your drop outs are or if you just got a poor quality axle?

Bummer man.

Tom
 
Based on my multiple experiences with sheared axles :( it doesn't make any difference which way the dropout is oriented.


The most likely cause of this type of shearing is one of the axles being not *completely* immobilized, so that it can rock back and forth, even just a tiny bit, while the other *is* immobilized, and cannot move.

The axle that cannot move would be more stressed than the one that can, and would be the most likely to break at the stress riser of the axle shoulder.

This is what I've seen on my SB Cruiser trike, whenever I've broken axles on any motor--the end that breaks is the fully-secured one, and I can feel the rocking of the other end, after the break disconnects the fixed end from it.


However...it could also be a specific problem to some of the QS205 axles; I've seen this problem posted more than once, and I have one here (from Shortcircuit911) with *both* axles sheared off at the shoulder (though this one was more likely because of *both* dropouts being loosened as the axle worked the metal of the dropouts over time). It is, however, more likely in cases like that, where much higher power was being used than the motor is sold as rated for, that the problem is in the amount of torque placed on the axle, which was not designed for that much torque.

The latter shouldn't be the case for this thread, as the
 
tonystark20 said:
the brake was tough to set and i was still playing with it because on flat ground it would lock up really easily but on a down hill in traffic where the regen can really save wear and tear on the brakes the limit was way to weak.

Do you have an analog control input for the brake? Or is it on/off?

If it's analog, is it a brake lever, or a throttle?

And what is the voltage range of the brake input, vs the voltage range of the control device?

And what is the mechanical range of the control device vs it's electrical range? (meaning, do you get the full electrical range over the full mechanical range, or is the electrical range a small portion of the mechanical motion range?)
 
Back
Top