RELIABLE rc-hobby city setup that doesnt catch fire??

Joined
Mar 27, 2010
Messages
214
Location
el paso texas
hello everyone im new but have been a reader of this forum for some time i find it very interesting, but ive noticed alot of you guys have issues with using hobby city stuff especially controllers catching fire, working fine andthen not working the next day, sync problems etc, these arent exactly cheap to be getting this kind of thing and in general pretty confusing to set up for a non electrical engineer,

i really want to do this kind of setup for my bike, i was about to buy a hub motor until i found this site, i like these more because they seem much more upgradeable than the hub motors for the same price, isnt the crystalite 1000 wat phoenix is up to like 1000$ now?

anyway im very interested and i figured ok ill just go with the largest motor for sale on hobby city, but i dont want to order just to have that catch fire on me the next day, so whats the deal can you guys that have already gone that route reflect on ur experiences here? i thought this thread could work as a nice reference point for us new guys, i like what aussie jester is doing i want to do something comparable parts wise on my mountain bike, who else has used the turnigy stuff?
 
one way to make a setup 99% more reliable ( and a tad cheaper ) is to fit sensors onto the motor a buy a sensored controller either from someone off this forum or e-crazyman/keywin on ebay ..
 
I would suggest in addition to obtining a keywin or infineon controller and fitting your motor with sensors, to look for the largest you can fit with the lowest kV... then convert it to wye while you are adding sensors, this will be not only a bit more efficient but also produces 1.73 x the torque and runs 1.73x slower...

The reasoning behind this is simple...

If you have room for a single stage reduction (say 9t to 92t #219) for a total of 10.22:1 reduction ability, take a motor with a kV of 215 RPM per volt and run from a 10S 2P pack - now look at LVC RPM at 3.3v per cell or 7095 RPM @ LVC now convert that motor to wye and the kV changes to 124.277457 or 4,101.15607 - assuming you want 25mph minimum speed at LVC (you will need to keep the 25mph speed via throttle durring pack discharge) then divide by RPM needed for 25MPH on your wheel size... at 20" mine is 420 so I divide the LVC RPM by this and come up with 9.76465731:1 reduction required... we can acheive up to 10.22:1 and that is what we should go with (it will factor to about 23-24 at LVC).

Now figure out the torque your motor creates -

Divide 1355 by the new kV to calculate the inch oz per AMP of torque generated at the motor...

1355 / 124.277457 = 10.9030232 in oz per amp.

Assume 50A (just for example purposes)

50 * 10.9030232 = 545.151161 inoz from the motor shaft @ 50A

Factor for the reduction:

545.151161 * 10.22 = 5571.44487 in oz or
5571.44487 / 12 = 464.287072 ft oz
464.287072 / 16 = 29.018 ft lbs torque

33v * 50a = 1650 watts minimum
41.5v * 50a = 2075 watts peak
37v * 50a = 1850 watts nominal or
1850 / 756 = 2.48 hp

You can always add a stage of reduction and increase voltage to obtain higher torque with less current.

Hope this helps!

-Mike
 
wow that was a very thorough answer im gonna have to look up some more stuff to fully understand whats going on, but first off, i just learned about the sensors today, so yo uguys actually open up the motor and put these inside? what do they connect to? they dont lock up tangle things inside? anybody have any pictures of this? i was thinking of simply trying out the largest turnigy motor but i read something that the hv-100 model, with sensors could almost equal its output for much cheaper once sensors where installed, is this right?

so why do the controllers have these issues of catching fire? do the hub motors have this as well? because id never heard of it so its either a quality thing or that people here in the forum really push their bikes

finally, what dc brushed option would be comparable? if the mars "etek" had a smaller brother with maybe half the power what motor would that be
 
You need to read up in the non hub motor section i think all these questions you have are covered there, your jumping in at the deep end.

KiM
 
thedarlington said:
so why do the controllers have these issues of catching fire? do the hub motors have this as well?


The controllers are made for flinging 10lbs RC toys around in the air. They do a good job at what they were made to do.

Pushing a 300lbs vehicle around and up hills and things is a drasticly different application than they were made to be doing.

Mount sensors. Use a cheap Infinion. Problem solved. No fire. No starting jitters. Mega low-rpm torque. High reliability.

You can get a 6fet for like $50, buy some IRF4110's from Methods for $20. Swap them in. Get $10 of hall sensors, and buy a RC sensor mount block from Burtie for whatever price they sell for. Put the pieces together, and you end up spending at least an extra 4-5hrs of personal labor, but the over all price is cheaper than using an RC speed controller, and the end result is SOOOOOOOOOOOOOOOOOOOO much better. I can't even express how much better they run sensored than un-sensored.
 
liveforphysics said:
The controllers are made for flinging 10lbs RC toys around in the air. They do a good job at what they were made to do.

Pushing a 300lbs vehicle around and up hills and things is a drasticly different application than they were made to be doing.

Mount sensors. Use a cheap Infinion. Problem solved. No fire. No starting jitters. Mega low-rpm torque. High reliability.

You can get a 6fet for like $50, buy some IRF4110's from Methods for $20. Swap them in. Get $10 of hall sensors, and buy a RC sensor mount block from Burtie for whatever price they sell for. Put the pieces together, and you end up spending at least an extra 4-5hrs of personal labor, but the over all price is cheaper than using an RC speed controller, and the end result is SOOOOOOOOOOOOOOOOOOOO much better. I can't even express how much better they run sensored than un-sensored.

A couple of followup questions:
Will any of the known bike controllers let me run an RC motors up to 12-14000 rpm?

I run a 6FET 4110 on my 72V 9C 9 turn. Phase current peaks at 88A even with this moderate setup.

What I'm looking at building is an RC setup (see other thread ) http://endless-sphere.com/forums/viewtopic.php?f=6&t=17242#p251774
This RC setup be rated 100-150A (depending on wind and motor, and may easily peak twice or more amps at startup.

The high end RC controllers seem to have a massive number of FETs in parallel, (though I have no idea of specs). So would a 160A Castle or a 200A Turnigy be a bad idea? Hard to believe it has less current capacity and is worse than a 6FET 4110.

Now another plan might be to get an 18FET BMC (ought to run at fairly high frequency since the BMC's are 80 pole -- but not sure if it will let me run an Astro at 12000rpm??) Then if needed can I just double up each FET to make it 6FET per driver and 36FET total? Or is that likely too much for the FET drivers? (I built an audio amp with 6 FETs in parallel on each stage, but this runs in class AB, not switching the FETS) I would also be tedious to add a cooling path for the new FETs, but not impossible. (I was thinking to just screw groups of 3 to an L-shaped aluminum piece. Screw the aluminium piece to the controller heatsink, and solder the extra FETs directly to the pins of the original ones.

I'm also watching Jeremy's construction thread with FET's directly on copper bar. I could try that also. All hinges on finding a 6FET controller board that will go high frequencey though.
 
jag said:
A couple of followup questions:
Will any of the known bike controllers let me run an RC motors up to 12-14000 rpm?

I do 10,000rpms with big 12p 14magnet HXT motors using sensored infinion controllers.

So would a 160A Castle or a 200A Turnigy be a bad idea? Hard to believe it has less current capacity and is worse than a 6FET 4110.
I've tried all of them, and burned them ALL into plasma balls. Same goes for everyone else on this board who tries to run real current through the RC controllers. You can get away with it for a while on a high turn-count astro motor or something with higher winding resistance, but the big HXTs just EAT controllers like popcorn.

Now another plan might be to get an 18FET BMC (ought to run at fairly high frequency since the BMC's are 80 pole -- but not sure if it will let me run an Astro at 12000rpm??) Then if needed can I just double up each FET to make it 6FET per driver and 36FET total? Or is that likely too much for the FET drivers? (I built an audio amp with 6 FETs in parallel on each stage, but this runs in class AB, not switching the FETS) I would also be tedious to add a cooling path for the new FETs, but not impossible. (I was thinking to just screw groups of 3 to an L-shaped aluminum piece. Screw the aluminium piece to the controller heatsink, and solder the extra FETs directly to the pins of the original ones.

It's simply not needed. I don't think anyone has even had a 6-fet fail running high power RC motors yet. You need to be running a hell of a lot bigger RC motor to be running into even melting down a 12-fet before the motor itself melts down.

I'm also watching Jeremy's construction thread with FET's directly on copper bar. I could try that also. All hinges on finding a 6FET controller board that will go high frequencey though.

They do.
 
I am planning on trying a sensored rc setup soon since I have an old 6 fet ecrazyman controller sitting here. I actually think it would be fun to make an rc setup that uses the pedal assist sensor so it feels as though you can pedal like superman, but I digress...

For me, I like the small size of the rc controllers. My cc hv85 can be mounted just about anywhere on my bike with nothing more than a couple zip ties and it's easy to hide. And so far, knock on wood, it's never lost sync (even from a dead stop) or exploded. Of course, I'm only running at 36 volts and my setup is probably pretty easy on it.

I think your choices should be based on what your intended use is. Don't buy the biggest motor hobbycity has just because it's the biggest one they have.

Btw, are that many controllers catching on fire? You've got me freaked out now, I need to do a search! lol
 
EVTodd said:
Btw, are that many controllers catching on fire? You've got me freaked out now, I need to do a search! lol


I think I'm the only one who has exploded them into plasma balls. Most folks just seem to let the smoke out and they stop working in a harmless fashion.
flameons.jpg

carnages.jpg
 
EVTodd said:
I do remember seeing those pictures. I just wanted to make sure they weren't going up left and right. My controller is too close to my ass and a big fireball in that area wouldn't be good! :lol:


lol, I feel you bro. :)

It actually burned my leg, but it was just a sissy little sun-burn like burn. I don't think it would do any real damage through clothing unless you had your junk resting on it, like a heated seat. lol :)
 
dayuuummm! i remember reading a post from liveforphysics where you pulled like 50mph running 2 big turnigy motors, where those fried controllers the result of that? u wherent kidding when you said ud had them explode!

ok so im understanding this as that, adding the sensors one would not need the weak controllers that the big turnigy motors eat thru? and i know these motors are made for model planes yet i see so many projects here using them, must be a good bang for your buck right? i think ill just go the less efficient brushed currie motor with a simple on off switch to get my feet wet but im definitely interested in pullin off the turnigy in the near future
 
Here's my take on the pros and cons of the two types of controller:

1) RC sensorless controllers use large paralleled arrays of FETs, which reduces controller losses at full throttle, as they have a very low total Rdson value.

2) RC sensorless controllers have poorer low speed torque and tend not to start as cleanly as sensored controllers.

3) Large paralleled arrays of FETs need high gate current drive to ensure fast transitions. This is hard to achieve, which means that switching losses in this type of controller tend to be higher. Switching losses have most effect at low speed, high phase current, operation, an operating area that model aircraft don't get into, because the power needed by a prop is proportional to the cube of it's rpm. A prop on a model aircraft can't demand high torque from the motor at low speed, a bike definitely does.

4) The smaller the number of FETs the better, in terms of reducing switching losses, provided that the total gate charge remains lower than that of a paralleled array of FETs. The challenge is finding FETs that have adequate thermal ratings so that they can handle high current and stay within junction temperature limits.

5) The limitation on all existing readily available (and affordable) controllers, of any flavour, is the FET package type. The standard surface mount D pack FETs used in RC controllers have a limited package current capability and are mounted in such a way that the total thermal resistance from the FET junction to free air is very, very high. The TO220 package FETs used on virtually all bike controllers also has finite package current and thermal resistance limits, but can at least be coupled to a heatsink fairly easily.

6) If you want a reliable high current, high voltage controller then the best bet is to look at using a small number of FETs that can be mounted easily, that have a high current capability, good thermal characteristics and an adequate balance between total gate charge and Rdson. The big SOT227 packages are pretty good, particularly in terms of thermal performance, although their total gate charge is a bit high. The D²Pack 7 pin FETs I'm looking at using have a lower total gate charge, similar current handling, but poorer thermal properties, plus they're harder to mount.

7) As there's no production controller like this (yet) the best alternative is to go for a well-proven set up that will deliver what you want, unless you want to spend a lot of time doing R&D. I'd suggest fitting Halls to the motor and using whatever flavour of Xiechang controller meets your needs.

Jeremy
 
I use a bmsbattery 6fet controller, it's really a good one. works sensorless, great start-up, cruise control, regen, possibility to run sensored if needed. and it's really cheap!
if you need more power, they have up to 18fet units, that works sensorless!
Mine run great on a kollmorgen inruner on 12s3p a123.
 
liveforphysics said:
You can get a 6fet for like $50, buy some IRF4110's from Methods for $20. Swap them in. Get $10 of hall sensors, and buy a RC sensor mount block from Burtie for whatever price they sell for. Put the pieces together, and you end up spending at least an extra 4-5hrs of personal labor, but the over all price is cheaper than using an RC speed controller, and the end result is SOOOOOOOOOOOOOOOOOOOO much better. I can't even express how much better they run sensored than un-sensored.
Thanks LFP!

For anyone - OK, I am ready to sensor my new 3220 and HXT. What exactly will I need to make this happen? Where can I get the 6fet controller, which hall sensors, ets?
My guess is this info is already on E-S. Can you point me to the link or give me the full low down?
Is there a manual of sorts to put it all together?
Thanks so much,
Roy
 
Roy,

I would also find it very helpful to find all of that info in one place. I think I have a pretty good understanding of how it all goes together but a little how-to would be really nice.

Btw, did you ditch the TowerPro 5330? If so let me know if you ever want to sell it.
 
Hey Todd,

I did find this thread:
Adding hall sensors to outrunners by Bertie
http://www.endless-sphere.com/forums/viewtopic.php?f=30&t=15686
and am starting my way through it.
But a how to with sources and products would be great :D

If you send me a pre paid box for shipping the TowerPro 5330, it's yours gratis. PM for details.

Roy
 
I've looked for links to the info you're interested in for many many hours, and was able to get enough hints to put everything together. Here's a brief summary of where I'm at, based on memory(don't have time to look everything up right now).

I though I was going crazy for a while because the site search was very erratic, one night a thread would come right up, and the next night, only posts from 2008 and back would show up. I told my wife that I have to search every topic every day because it changes constantly. It seems like the issue is fixed because my searches are much more productive now????

Motor Controller:
Sources: ebay?, a couple of guys on here, still basically searching
There is no step by step for controller mod thread that I have been able to find. I used to repair car audio equipment and motor controllers are very similar in construction to amplifiers so figuring out the controller mods was almost more common sense that learned on here. It's interesting to see the differences in fets used. The IRFB3077 seems better for applications using 36-50Volts because of its lower RDS-on(current required to switch) while the IRFB4110 is better for 60V+ applications. It's all about increasing the amperage capacity and then reconfiguring the controller to tolerate the higher volts/amps. This is done by messing with the value of the shunt resistor(one member mentioned half value) and with programming. The circuit traces on the board need to be beefed up. This can be done by laying a pool of fresh solder on the traces, cutting a stranded wire to length, tinning it, laying it over the trace, and flowing solder over it. You have to be careful that you don't overheat the other components on the board while doing this. They can be removed, or cooled externally(pieces of heat sink clamped to them, air cooled, etc). I have also noticed that manufacturers and modders alike are putting ceramic resistors right down close to the board. They will transfer their heat to the board and weaken solder joints over time. This is one of the differences between a controller that lasts ten years instead of two though. If you've ever had a ford truck with a cd/tape player that has a burnt out display, the resistors being too close to the board are the cause. I never put a resistor that heats up closer that 1/2" to the board. Also, quality fets and caps help a lot. Ratings don't mean to much these days. Digikey is a great source for components. Also a good source for hall effect sensors for the motor.

Motors:
Astro3210, 3220, etc.
Turnigy,HXT, etc
Source: just search online, easy
On the cheaper motors, it's a good idea to take them apart, make sure the magnets are in stout, re-glue with high temp expoxy if necessary, replace the bearings(many sources, just search miniature bearings), and make sure windings are tight and secured. It has also been recommended that the big turnigy's be rewound/reconnected from delta configuration to wye. This increases the torque that the motor puts out and lowers the rpm, allowing for less reduction in your transmission. BLDC Motor winding configurations can be searched online for more details. The hall effect sensors need to be mounted in the correct position in relation to the magnetic fields produces by the motor. One of the members here did some gauss testing and if I remember right, the optimal placement was 5-8 mm above the outrunner can and away from the edges, towards the middle. There have also been a few different strategies posted about the radial positioning of the sensors. You'll just have to search this stuff. I seem to remember, a 120 degree even spacing(which makes the most sense to me, and something like 17.34 degrees. I definitely see how the 17 degree one would work, but it seems like the 120 spacing would "spread the load" out on the different coils, and potentially make the motor run a little smoother. It seems like putting the spacing at 17 degrees would make the controller "throw the motor around" with three nearby coils every time??? I don't know, I haven't had time to get into this area yet as I'm still ordering parts. I found an awesome PDF somewhere that is a must read.

Transmissions:
Sources:banebots, matex, stanton, neutgart, recumpence drive

This is wide open territory. Recumpence's drive is beautiful, and does the job very well. I like the gates powergrip2 belts for the first stage. This is an area where, if you can't design something yourself, you're best off getting a drive from recumpence. He's pretty much nailed it. I've been researching this for four years and came up with a very similar design. I've looked in to every alternative, and dollar for dollar, you can't beat it.

So there it is. I hope it helped a little. It helped me just typing it all out.
 
mdd0127 said:
I seem to remember, a 120 degree even spacing(which makes the most sense to me, and something like 17.14 degrees. I definitely see how the 17 degree one would work, but it seems like the 120 spacing would "spread the load" out on the different coils, and potentially make the motor run a little smoother. It seems like putting the spacing at 17 degrees would make the controller "throw the motor around" with three nearby coils every time??? I don't know, I haven't had time to get into this area yet as I'm still ordering parts. I found an awesome PDF somewhere that is a must read.
each hall signal tells the controller to power the relevant phase winding so it will run just as well/smooth with either 120 or 17.14 degree spacing, if the hall sensors are to be mounted inside the motor then 120degrees is the only real way to go, however if mounted externally the 17.14 degree spacing makes it a smaller more compact design

I think you have done a good job putting this info together. Hopefully all the relevant info for fitting halls to motors will come together in this thread 'Adding hall sensors to outrunners' http://www.endless-sphere.com/forums/viewtopic.php?f=30&t=15686 , the controller info is another issue and it is all over the place on this forum.
 
I am also on the track of running outrunners on the E-crazyman controllers. I have all the parts in hand & will post a build thread for dummy's by the biggest dummy on the board in that regard.
I am also going to try may hand at controller modding......I have a capacitor kit from methods, & some good book marks I'll share right here.

Controller modding info:
#1-resistor values for different pack voltages
the last post has the info
http://endless-sphere.com/forums/viewtopic.php?f=16&t=7361
#2)how to program your controller:
has lincks to download the programs & not to much techno speak to confuse dolts like myself
http://www.ebike.biz/infineon.htm
I bought the usb-ttl adapters on e-bay for $6 a copy

A souce for small conectors for hall sensors & programing connetions:
http://www.cheapbatterypacks.com/index.asp?menu=996203GEB9&sid=211318

now into the gray area-
this is a link with photos of Methods handy work modding a 6fet controller. glean what you can
http://endless-sphere.com/forums/viewtopic.php?f=2&t=14488

I have no problem reading & re-reading these posts & slowly learning what seems easy to the more electricly inclined. I am just aprecitive they took the time to post this stuff for me to mine.
The mechanical act of soldering & making brackets to mount hall sensors are the easy part of these things.(for me anyway)

Again the reduction & mechanical side of the build come easyer to some. Now that Gwhy showed me how to post a pdf, i will post all the brackets I construct.
Recumbence offers very high end reduction units but you still need the chops to mount one to your bike.
I have Not blown an RCcontroller yet (I am running CC Ice100 & a turnigy HV200AMP) but I am not satisfied with the slow speed performance, throttle reponce (kinda notchy)
 
Thanks for all your responses.
Looking for links and specific info regarding how to identify a suitable infinion type controller.
Looking for high volt and amp units or suitable units that I can mod with Methods fets and more.

Is this a good controller to get:
72V 1500W Brushless Speed Controller for Electric Bikes
http://cgi.ebay.com/ebaymotors/72V-...4193044QQptZMotorcyclesQ5fPartsQ5fAccessories
It's about $50 with shipping.
 
If you want an ifineon controller pm Lyen
he has 6 9 12 and 18 fet Infineons and will fit 4110fets as well as supplying additional cables for programming etc...BRILLIANT service I have bought 2 myself.

KiM
 
RWP said:
Thanks AJ.

Looks like the Lyen controller is not rated for more than 2200rpm if I am understanding correctly:
Turnigy (at up to 2200rpm only), Astro (at up to 2200rpm only).

found just above the photos here:
http://endless-sphere.com/forums/viewtopic.php?f=31&t=17191&p=251196&hilit=leyn#p251196

Is this correct?
I want 6000 to 8000 rpm

I believe that should be his sensorless only - with sensors you shouldn't have an issue, on 846 style (older) 9FET infineons I have driven my Tower Pro on 10S above 8K with a very light load prop on it... no damage and controller was dead on.

My internally reduced geared hub motor has something like 3800 RPM + before reduction and it runs perfect with the Infineon too (even at higher RPM than my current 15S pack can provide sufficient voltage for).

-Mike
 
Back
Top