HXT, Staton gearbox, NuVinci build.

[dontsendbubbamail]

anybody here used these: http://www.bmsbattery.com/product_info.php?cPath=34&products_id=172 on a rc motor? they say it can run sensorless

I have not used a controller from this vendor. The size of the controller in the picture looks more like a 12 fet then an 18 fet. My reference point is the e-crazyman controllers that use the same case.

Bubba

 

[johnrobholmes]

looks like a 9 fet pic to me :lol:


Your very long battery and motor leads do not help the situation at all, per the popping caps. The controllers really need about 7 times the capacitance of normal. You could easily make a sensor board for the motor and run an Ebike controller. They aren't as high performance, but that seems to be why they last longer.

 

[recumpence]

The issue is not the components, but the way they are being used.

Believe it or not, the worse thing you can do is run an RC controller at low throttle for too long. They like being run at high throttle (50% or above) most of the time, or they blow.

I am very sorry you have had such problems. I went through this at the beginning as well.

Matt

 

[Kepler]

Bad luck on the failed components. For it to fail while not under load suggests to me some sort of short and most probably from the extra capacitor bank as you suggested. My thoughts are that you have over complicated the design and tried to include too many features. I am not sure what the terminal strips are doing but I am not too keen on using this type of low cost connector. My advise is either hard solder any connections or use a high quality connectors like Deans, Andersons, or gold plated bullets. I would also ditch the fuse, fan and pre charge resistor as I think they are only providing un necesary compication and added weakness to system.

I like to keep my RC setups as simple as possible. The CC controller and Throttleizer are a great combo and normally very robust espsecially if setup conservativily. I noticed you are using the temp probe to measure the air temp in the box. I think you would be much better off fitting it to the heat sink on the ESC. Also the Throttleizer already has built in caps and a BEC so there is nothing else that really needs adding. CC dont recommend pre charge resistors and for 36V, the spark isnt that big anyway.

My thoughts are, get the ESC out in the breeze, keep all cables as short as practable and reduce your parts count down to only what is necesary. Your drive looks great and I think a few mods to the electical side will make all the difference. Good luck with the repairs.

 

[Hardcarve1]

I agree Kepler, thats what I will do next and get it back to bare bones.

Thanks for the advice.

 

[12p3phPMDC]

IT seems the the RC controllers are just too sensitive for the application....and that you have to treat them with kid gloves.

They are overrated in terms of current handling. They need to be derated IMO. Otherwise you end up with Castle Cremations or Burnigy.

They have proven to be too much for the big HXT time and time again.

 

[Kepler]

Thats the good thing about the throttleizer as you can effectivilly protect the ESC by setting the current limit down to a safe level and at the same time introduce extra thermal protection. This setup will have very good performance if the current limit is set to around 60A and at this setting, the ESC wont even be breaking a sweat. I think RC setups can be quite robust as long as a few well documented rules are followed.

 

[gwhy!]

I have played around with 'cheap' rc controllers with success without current limiting and do work for your run of the mill no frills e-bike speed controller as long as you don't want to explore the full potential of the motor its driving ( they are not designed for this sort of dead load). There were 2 reasons why I went down the sensored controller route was I needed 100% instant start ( without fear of blowing the controller up ) which I dont think anyone has achived using a rc controller and the cost. Cost and reliability are 2 very good reasons to go sensored I think :wink: . Sensored controllers are programmable, configurable, repairable and upgradable. No one can use the size/weight of a rc controller thing as a argument because of all the other additions you need to add to make rc controller 'hopefully' reliable.

 

[AussieJester]

MY hv110 has stood up fine pulling upward of 137amp hasn't got even slightly warm. I run a fechter
throttle interface with current limit OFF...I think the extra caps make a big difference in keeping
them alive, i wonder how many that have blew them had 5 additional caps in front of the ESC?

Size argument is definitely valid still GWhy, the hv110 and throttle interface are still less than half the size of a 12fet controller. If it wasn't for the instant start issue i would be happy as a pig in shit uisng the HV110 and was actually getting very used to the start up just before the batteries went...The modifications to fechters box and the magura throttle made it very useable, i think if i had had the eno on the rear rather than the sloppy Sturmey archer cracking and complaining when giving it stick from standstill i wouldnt have worried about the sensored conversion.

KiM

 

[Grinhill]

My Turnigy HV100 still puts up with occasional abuse of this type:


(can you top that Kim )

This was through a Standalone CA with a 100A recommended peak as well

Of course, I wouldn't attempt to go up the hill in front of my house on 3/4 throttle for more than about 5 seconds. The heat just can't get out of the little bugger quick enough.

 

[recumpence]

IT seems the the RC controllers are just too sensitive for the application....and that you have to treat them with kid gloves.

They are overrated in terms of current handling. They need to be derated IMO. Otherwise you end up with Castle Cremations or Burnigy.

They have proven to be too much for the big HXT time and time again.

I 100% disagree. The issue is not that they need to be handled with kid gloves, actually, the problem is they are normally being babied too much. What I mean by that is, partial throttle for long periods of time is the single biggest problem. I beat the friggin CRAP out of my Castle controllers and they survive just fine. My first twin motored trike using old style CC HV controllers is still cranking and that thing is constantly seeing 350 amp peaks (two HV140s) without even getting warm. The 3220s and HV140s run pretty much cool all the time.

I agree that more durable solutions are possible. However, the issue is more the specific setup and incorrect use than anything else. Also, part of this is, the cheapo HXT motors are problematic. I am getting more than a little annoyed that everyone is so addicted to buying freakin cheapo garbage motors and are bewildered when they have problems. Those motors SUCK, plain and simple. If you want to reglue magnets, replace bearings, rewind the stator, and mount sensors just to save a bit of money, fine. However, I find that to be a total waste of time, unless you are looking at it as a fun hobby (rebuilding the motor). However, if a decent motor is used, these problems seem to not be a problem.

I am sorry if I sound a bit short tempered here, however, I find it odd that everyone is so determined to use these HXT motors. I understand that better choices cost more, however, nearly every problematic RC build I have heard of use that crap HXT motor.

I do not know about you guys, but there seems to be some consistancy here......... :?

Matt

 

[AussieJester]

So Matt...if i switch to an Astro i'm no going to have sync issues then? We both know we will. I am more than happy with my HXTs performance aside from the sync at start up, i did add additional glue but on inspection it really wasn't needed there was heaps of excess around the magnets, version 2 turnigys have larger skirted bearings so no need to change those either, the windings on my version are good to no idea why people are rewinding them? I also see no reason to change the shaft, setup on your reduction drive it works perfectly. Not everyone has what ~$US500 to outlay on an Astro Matt, then wait for it too be made when some aren't even certain their rc setups will work. HXT are 100 bucks...for those wishing to try an RC setup i think they are a good choice... each to their own though no disrespect meant Matt If i can afford one maybe ill try an astro one day, if they ever get hall sensors on them as they are, they are no better than a HXT for my needs

KiM

EDIT: Grinhill no i couldnt the Fechter interface as an 140amp cap apparently haha with me riding the bike i am yet to get the HV110 even warm its stone cold after every ride i have been on and all i ride is hills haha When Matt.P rode it for the first time it was in 3rd gear and he could only ride around under half throttle it got stinkin hot then...They love the revs though keep them on 3/4 to full throttle and they buzz happily away heat free i have found. The HXT also, i have yet to heat it up to the point i can't place the palm of my hand on it indefinately after a ride, it has been faultless thus far, if it last 12 months and has to be replaced i would consider this money well spent.

 

[12p3phPMDC]

Boy, it looks like I hit a nerve.

It seems we have more controller failures than motor failures. Am I wrong?????
I understand that the usage at low throttle will hurt the controller due to switching losses, etc..

Cost is an issue as AJ points out. From my vantage point, it's like the Astros are the Twin Overhead Cam 32 Valve 400 hp aluminum block V8
that cost $10k. But, I can buy a pushrod small block chevy motor for much less, and have just as much fun. 8)

Castle Setup with 3220. $800 to $900 bones
Turnigy Setup with HXT. $200 to 250.
6 monster fet setup with HXT. $250

I think that GWhy has many excellent points in regards repairablilty, upgradability, driveability, etc. That's a goal I'm shooting for with
the 6 fet I'm putting together, all with monster fets that all told will cost less than a Castle and you'll be able
to run higher voltages. Try modding a Turnigy or Castle to do that. Eventually, there will be something available
without mods. Even the RC controllers are being modded to some degree, although caps are easy.

The size of the controller matters as well. The more thermal mass a controller has, the more heat it can take before it saturates the sink and then
you get thermal runaway on the fets and POW!!!! toast. Miniturization and extreme power density is awesome, but the small stuff saturates faster.

 

[gwhy!]

Size argument is definitely valid still GWhy, the hv110 and throttle interface are still less than half the size of a 12fet controller.

KiM

We will have to agree to dis-agree on this one KiM. Im using a 6FET controller and its the same size as all the bits needed to make a rc controller work and make it reliable, My 6fet can output upto 120A and upto 80v ( I havent tried this yet at the moment its limited to 100A @48v and it just works, for a quarter of the cost of a CC+ other bits).

 

[recumpence]

Ahh, yes, that is right, the new HXTs have better quality. I forgot about that.

There is a correlation with type of motor used, however, when it comes to controller survival. In fact, though I hate to say it, my $1,100 Plettenberg on my recumbent is hard on the ESC. If I ride above 50% throttle, it is fine. But, below that, my ESC heats up. WIth my AXI motor, that does not happen. So, something with the pole count, or steel laminations, or something else is at play there. In the end, yes, I like Astro motors. I do understand the cost factor. However, this is kind of like entering a race, you can skimp on the engine, but racing is about performance a bit more than cost. RC stuff is that way. It is expensive because the weight is low and performance is very high. That being said, there are ways to cut cost. However, if the right blend of components (motor, ESC, batteries, charger) are used, and run properly, there are very few problems.

I remember my first two stroke dirtbike (79 YZ80). I beat the crap out of that thing. But, a friend of mine had a KX 80 (supposedly better bike), yet he had problems with his. But he never let the bike get on the pipe. He bogged it everywhere, afraid he would hurt it if he revved it. Turns out, he was hurting it by boggin it more than I was by revving mine out. Same here, RC stuff needs to be setup right, and ridden right. It is not bad to run them slowly here and there. But, they really like to be pushed, as long as you push them above 50% of their RPM. At that point, hit em hard! Below 50% of their RPM, you gotta be nice to them.

Matt

 

[Jeremy Harris]

My own view, having taken apart a few controllers and looked to see how they failed, is this.

RC controllers seem designed to operate as Matt says, at high power settings. If you look at the gate drive signals on them you'll find they aren't that good, as they use a large number of paralleled, small package, surface mount, FETs to get the high current rating in a tiny, light package. The snag with this approach is that all those paralleled FET gates creates a massive gate charge requirement. Even the hefty gate drivers used on the better controllers struggle to provide really good transitions, which means that the FET switching losses at low throttle settings are going to get pretty high. The problem reduces the harder you drive the controller, just as Matt says.

If you want a controller that will tolerate pootling along at part-throttle for long periods, then either get an RC one that is massively over-rated or go for a sensored ebike controller that uses a smaller number of bigger FETs. The bigger FET packages are capable of handling high current, with better heat conductivity to a heatsink, and will generally have a much lower gate drive charge requirement, so will have lower transition losses.

When it comes to motors then frankly I think it's just a matter of where you want to spend resource. If you're cash rich and time poor, then buy an expensive motor that's just plug-and-play. If you're cash poor and time rich then get an HXT and spend a couple of hours and a few dollars tweaking things like the bearings and magnet bonding. As I know that I'd just have to take any motor I bought apart, if only from curiosity, then I'd always opt to buy at the cheaper end and put the time in to make it better. I'd also always advocate fitting sensors and going to a sensored controller, based on the massive difference it makes to start up and low speed, high torque, control that I've experienced.

Jeremy

 

[recumpence]

Thanks for the tech info Jeremy. I can give analogies, but I am not an engineer.

I appreciate guys like you who can articulate these things!

Matt

 

[Hardcarve1]

Looking back now on why the controller blew Matt is right, I was just feathering the throttle while peddling to try and maximise my range. I was aware that you should not load the RC controller up at low speed like on start off, but was actually loading it up because I was using a taller ratio during pedalling. I should have just reduced the ration on the NuVinci and held much higher throttle speeds which in some way is not what I really want to do because its not conserving battery. Maybe sensors will be better suited to my style.

 

[AussieJester]

It is not bad to run them slowly here and there. But, they really like to be pushed, as long as you push them above 50% of their RPM. At that point, hit em hard! Below 50% of their RPM, you gotta be nice to them.

Matt

This would make a good argument FOR a two speed transmission Matt, a very very low range to allow the motor to rev high whilst
still alowing you to potter along at very slow speed without overloading the ESC! It could possibly by argued mulitple gear ratios would be a good thing, i had changed my opinion of using gears to that of if you use a single speed geared to your meet your needs multiple gears on higher powered electric motors are not needed...MIght have to re-think that...(i still stand by low powered 750watt and below electric motored bikes could benefit from multiple gearing) I still MUST HAVE a Thud dog box transmission but that wasn't for performance just a case 'pose' value hehee...that and gear changes would sound so sweet on the lecky motor at full speed ...

KiM

EDIT: Just had a thought ..Matt...anychance we might see an automatic 2 speed transmission like the 1/8th onroad nitro cars had/have?