#crossbreaks' midmotor CB120

1. IMHO there's not enough room between bearing seat (I'm not sure the exact term in English) and holes/threads for screws. Very thin wall tends to end up be torn or distorted during machining.
2. Unless the rim for timing pulley is a separate part, how do you install the outer bearing?

photo_2017-04-21_09-48-18.jpg
 
minimum said:
1. IMHO there's not enough room between bearing seat (I'm not sure the exact term in English) and holes/threads for screws. Very thin wall tends to end up be torn or distorted during machining.
2. Unless the rim for timing pulley is a separate part, how do you install the outer bearing?
Click to expand...

1. True. the part will be turned on the lathe first, then the holes are drilled. I hope that will work out fine. The metal parts arrive next week, so i can try it. the construction is quite ambitioned. i try to make it as robust as possible. if the bearings fail then the magnets are dead.
2. the pulley is a separate part. I will change colors in the model to make this more clear. The sideplate of the pulley that holds the belt in place axially is a seperate part, too. It is held by the 6 M5 bolts that also fasten the pulley.
 
I'm assuming the housing is aluminum, so there is a tendency for very thin wall features to be pushed or formed out, instead of being cut (that includes drilling and thread boring). Or in another words, the last part just gives away before cutter, instead of being cut. As the bearing seat should be cut with proper tolerances, so it's a press fit; so drilling holes later may ruin the bearing seat.
Over the years working with aluminum I have learned to avoid features less than 1mm thick. Then again I may be wrong and everything turns out just like it's supposed to be :wink:

Few more chamfers would be good:
photo_2017-04-21_09-48-18_1.jpg

First, to guide the timing belts (as the alignment can be never perfect) or the sharp edges will eat up the belt. Download CAD for commercial timing pulley with flanges for study.
Second minor one is for inner edge of timing pulley.

Large enough chamfer could be used for the clearance issue with bearing, already pointed out earlier.
 
sdp-si might be a good place to get CAD's.

Here's front view of a random pulley cad:
HDT5M-24F.JPG

Consider adding chamfers or fillets to (almost) all edges - it helps to reduce stress areas, makes post-machining processing easier (no need to deburr) and may be helpful while assembling.
 
I agree with above from minimum, wall thickness is too low towards bearing seat. I drilled out my revolt motor to fit 8awg wire and the surface buckled towards bearing side due to too thin walls. I think the drilling operation might be better done first and turning later as drilling creates more forces on the material than the turning (depending of cut depth :D )
 
I've found the easiest way to get the most copper out of the axle is using magnet wire, so instead of three sets of phase wires each with insulation all of the phases can be combined in one bundle (along with the hall sensor and temperature wires) for more copper but using multiple layers ending up with more insulation protection. There's no need for flexibility until outside, so I'd suggest leaving extra length when winding and just bring the phase windings outside of the motor.
 
macribs said:
Video not viewable.
Click to expand...
*fixed

larsb said:
I agree with above from minimum, wall thickness is too low towards bearing seat. I drilled out my revolt motor to fit 8awg wire and the surface buckled towards bearing side due to too thin walls. I think the drilling operation might be better done first and turning later as drilling creates more forces on the material than the turning (depending of cut depth :D )
Click to expand...

thx, i'll try it that way. And thicken the wall, ATM it is .8mm which might be too athletic i agree. i'll try making it >1mm by shifting the BCD by .5mm (bolt circle diameter)

John in CR said:
I've found the easiest way to get the most copper out of the axle is using magnet wire, so instead of three sets of phase wires each with insulation all of the phases can be combined in one bundle (along with the hall sensor and temperature wires) for more copper but using multiple layers ending up with more insulation protection. There's no need for flexibility until outside, so I'd suggest leaving extra length when winding and just bring the phase windings outside of the motor.
Click to expand...
i know revolt and turnigy do it this way. I had broken strands on the cable exit once or twice, which is hard to fix if you are in the field without striping the whole motor. something you defnitly dont want. hubmotors have those connections inside, my design is more like a hubmotor, so id like it to be that way, too. I think the extra effort is worth it, as long it does not add axial length to the overall design

minimum said:
sdp-si
Click to expand...
thx but they have no PolyChain pattern. A HTD M8 belt would be too wide to fit the buildspace. I know a polychain/DeltaChain is expensive ($50+ for a 150T, 12mm wide 8M belt) but it's the best you can buy for the purpose, and i should last ~10.000 miles
 
crossbreak said:
thx, i'll try it that way. And thicken the wall, ATM it is .8mm which might be too athletic i agree. i'll try making it >1mm by shifting the BCD by .5mm (bolt circle diameter)
Click to expand...

now it's 7pcs M5 bolts instead of 6, resulting in much thicker walls. Thanks for all of your input, this really helps to improve it before i built the "final" prototype next week
 

Attachments

  • .5mm1.png
    .5mm1.png
    33.1 KB · Views: 4,384
  • .5mm2.png
    .5mm2.png
    45.4 KB · Views: 4,384
For the halls i would recommend to use a shilded wire. It would be also good in terms of signal quality to add a pcb inside with capacitors (to stabilize the +5V supply) and resistors (like the mac motor has or some bafang).

Or what about an encoder like larsb has put into his revolt?
 
i use this cable: UNITRONIC® LiYCY 6 x 0.25 mm², shielded

https://www.conrad.de/de/datenleitu...-grau-lappkabel-0034406-meterware-600719.html

sorry no time ATM for adding the encoder PCB..this will have to wait till August. Which encoder do you want? The 14bit SPI one? I dont have the model number at hand...? Or do you want the renishaw rmc22? if you plan to put the motor into a robot axis drive, then i suggest using a renishaw one that is supported by the Aerotech NDrive-HPe with encoder option https://www.aerotechgmbh.de/produkt...p=/produktkatalog/antriebe-antriebsracks.aspx Just give tech Aerotech support a call, they can tell you which encoders are supported

anyway, for actor drive, i always tend to put the encoder as close to the tool as possible...not on the motor but on the driven belt pulley. otherwise you have the belt in the middle which distorts position feedback measurement

i plan on releasing a special "robot drive" version of this motor and start shipping them in November. The robot version will be more simple and a bit cheaper, including the 14bit SPI encoder, ODrive BLDC controller and the A123 systems coulomb buffer. I also participate in the ODrive controller development and plan on selling those controllers along with the robot version of this motor. The encoder interface is done already and works quite well. The motor can be used for scara type robots, rotary tables and there will be a horizontal, belt driven 500mm linear stage with optical linear encoder attached directly to the stage, including an adjustable counterbalance in the center of the stage, for payloads up to 100kg with micron precision.

here is a very early pic of the 500mm horizontal stage and the counterbalance that is hidden inside. Works just like an elevator but with a belt drive instead of steel cables :D dont worry if you dont find the belt in those pics, it's still missing in the model
 

Attachments

  • horizontal-stage-500mm.png
    horizontal-stage-500mm.png
    21.1 KB · Views: 4,380
  • counterbalance.png
    counterbalance.png
    41.8 KB · Views: 4,380
i'm setting up the lathe to make proto side covers. The fan blades will be added later on a 3axis CNC that gets retrofitted ATM. Cooling design is almost done and i have a good quote on 0.35mm laminations and the magnets. The rotor backiron will be cast into alu for saving weight and to easily add thin fins of 2mm height. Watercooling design isnt done yet. I hassle with my supplier since he refuses to cast an aluminium stator carrier that i need to fit channels for watercooling. We still have the pressfit option though: The inner bore of the stator gets an epoxy coating, then a lathe-turned stator carrier is cooled down using frozen CO², the stator is heated to 120°C and both parts are pushed into each other. Tolerances must be very tight and am not sure if i can do this on my lathe. Maybe it is worth a try. a temperature difference of 150°C+ may be ok with 50micron total tolarance...at least I calced 80microns should be fine. Here is a pic of the new lathe:
file.php
 
I love new tooling. When I was a kid my tool and die/mechanical engineer, certified welder father had a complete machine shop in the basement. He would sit down there pretty much my whole child hood building. He made hit and miss motors, pop motors, hot air "Stirling" motors in beam, combined and some where made into fans, compressors and just about anything mechanical. He made me custom parts for my Tamiya Sand Scorcher R/C cars when the factory parts wore out. He would redesign and make it last including bearings, universal joints, shocks and gears. He was heli-arc welding before tig was a name and a master at it. I spent a thousand hours will him in the shop and he had 5 times more down there. I sure miss him!

I was always afraid to by that stuff (lathe, milling machine and everything else) in fear of getting lost in building stuff in the basement after all the years I built off road car stuff and tinkering. I envy you for having that stuff and the time.

Sorry for the off topic walk down memory lane.

Looks like a nice piece of equipment.

Tom
 
Yo! Crossbreak, and Litespeed!

I found the last few posts about the machining and Litespeed's father highly inspiring. Thanks for the thoughtful and insightful posts. This illuminated a world of stuff, that I was vaguely aware must happen when designing a product, but with concrete examples of the thought process, tools, and effort needed to get anywhere near achieving something useful!
 
Hey Crossbreak,
this looks very promising. I have a Mondraker Summum DH Bike and an 14s7p 25r battery sitting here they are both waiting for a big Motor :)
I would like to buy a complete kit.


Gesendet von meinem GT-I9505 mit Tapatalk
 
This is an impressive project :shock: , after months reading ES, I've just decided to sign in to congratulate your work.
 
sorry for no progress ATM. the weather is too good to sit in the workshop :D but it will go on as soon as the temperature in the workshop drops. Good news: There will be a better lathe soon, i'll get it from an old gunsmith, it is more precise than the chinese thing i got recently.
 
Hey CB, just one question. Where did you get the raw-material for the lams here in Germany / Europe ? I can hardly find a dealer around :shock:
 
ok, when you built the prototype, how did you get these lams ? Or what is the type/name/code of the material. A close friend of mine has 2 laser cutting companies in the stuttgart area. If I can let him know the material, he probably can get the lams and cut them out.....
 
You must log in or register to reply here.
Back
Top