GNG, 1000W 48V BB-drive, $400

flyinmonkie said:
skyungjae

I am betting that the ego uses standard isis bottom bracket and freewheel cranks. They may use a an eno flanged freewheel or an ACS, or just a cheap dicta one. But, they don't have a support bearing like the GNG freewheel crank. So it is like the cyclone freewheel crank just with better parts.

You could do the same on your GNG by just building your own freewheel crank with off the shelf parts and forgoing the support bearing. The eno has been shown to hole up fairly well, but isn't designed to be over run as it only has one row of bearings and is expensive. It is easily serviceable though. The ACS has 2 rows of bearings and is thought to be better suited to be over run, but I haven't seen many use it. It is much cheaper though.

Clay

I already did away with the GNG chain ring adapter. When I stripped the threads on my GNG crank arm, I ordered some beefier cranks. Unfortunately, the spacing was off quite a bit. That's when I decided to get a 48T outer chain ring from Sick Bike Parts with one of their freewheels and use my 34T Blackspire inner chain ring. It spins true under pedaling when the GNG chain rings, even with the support bearing, did not (it did spin true under motor power).

If I had the coin, I'd order some ISIS crank arms and start guess and testing some BB's.
 
Sorry belt fans, but USPS arrived with my chain rings a few hours before the UPS guy came with my larger belts, and I'm an impatient man:


11:80 gearing

I ground it down to ~11mm with an angle grinder and did the rest with the file. Drilled a hole through the motor shaft to use the stock mounting pin which you just hammer in. Can't wait to try it out :D

I'll still be mailing a few of those pulleys out to anyone that wants to experiment with the belt some more.
 
A couple of noob questions as I've got confused after following the various paths for too long.

Are there significant benefits other than closer pedal assist cadence that people are seeking more reduction for?
e.g. When riding overpowered I burst throttle to a speed and then pedal alone until an incline or speed drops below pedal cadence.

With the adjustable plates, will this be good enough for reliable belt life and reducing the wasted power?
 
bee said:
Sorry belt fans, but USPS arrived with my chain rings a few hours before the UPS guy came with my larger belts, and I'm an impatient man:


11:80 gearing
A little light reading, a little late for your setup. You are, of course, free to do it your own way if it makes you happy, and if you accept the consequences.

To be fair, my own first e-bike had 12:144 gearing (to avoid an extra reduction stage). But the wrap angles were more than adequate, because the sprockets were more than 2 feet apart. It still liked to dump its chain with irritating frequency, like once per ride. If there had been enough operating clearance around the big sprocket to install guide flanges, I would have done that, and that probably would have cured it.
 
Chalo said:
bee said:
Sorry belt fans, but USPS arrived with my chain rings a few hours before the UPS guy came with my larger belts, and I'm an impatient man:
11:80 gearing
A little light reading, a little late for your setup. You are, of course, free to do it your own way if it makes you happy, and if you accept the consequences.

To be fair, my own first e-bike had 12:144 gearing (to avoid an extra reduction stage). But the wrap angles were more than adequate, because the sprockets were more than 2 feet apart. It still liked to dump its chain with irritating frequency, like once per ride. If there had been enough operating clearance around the big sprocket to install guide flanges, I would have done that, and that probably would have cured it.
Thanks, I'll check that out when I get #25 half-link that I need in order to remove the idler on the secondary reduction. I still have some room in the adjustable plates to extend both chains. I'm just testing the reduction for now, I got a bunch of different sprockets to figure out what feels best.
 
Got my pulleys from SamD, thanks Sam!

Had one of them smoothed out with acetone vapour today. Worked great.
pulleys.JPG

Go laugh at my setup :lol: It's a glass jar filled with ss nuts with a little bit of acetone at the bottom. The jar is submersed in boiling water. The nuts are there to stop it from floating. The acetone was boiling nicely. The plastic got all soft for a while but hardened up later.
setup.JPG
 
I think the 3D printed LARGE pulleys with acetone smoothing are an absolute winner just as they stand right now (many teeth engaged), but...I have lingering concerns about the small pulley. Although, I definitely want the experiments to continue...I am watching with great interest.

The finish on that small pulley looks pretty good, how long was it in the vapor before removal to cool off?

In response to bee's post just below this one: I would sand/file/dremel the stock steel pulley (with its sharp edges) down just enough so that the printed pulley slides on snugly, and then epoxy it on to remove any slop. All of this information is much more importamt that just upgrading a GNG kit. There have been a LOT of creative and interesting developments from the builders here at ES, and adding cheap and easy 3D printing to the mix will make for some fun ideas sprouting from the ES fields...
 
It only takes a few minutes of acetone vapor for a nice polish if the jar is warm enough. The trick is to get the acetone vapors hot enough so that they start to condensate on anything that goes into the jar, this causes a very thin layer of acetone to form in every nook and crevice of the printed part which quickly starts to melt the outer surface. I like to pull my printed part out halfway through the polishing process and flip it upside down since gravity does cause the melted plastic to sag a bit and flipping it reverses the sag a bit.

Here's what a non-polished pulley looks like after about 10km at 14s with a stock controller (click through for high res):



As you can see, the high speed and lack of tensioner causes the belt to float on the pulley and only engage on the tip of the teeth! This is the 20t sprocket and there was no belt skipping with this config, the pulley fell off after 10km because I didn't glue it.

If your pulley fits too tight, file it down before hammering it on or you will start to crack/stretch the plastic at the thin parts:


(I'm referring to the lighter toned "stretch marks" running down the valleys of the pulleys in this pic)

Here's the latest code for my larger sprocket if anyone wants to give it a try:

Code:
//GNG LARGE up-sprocket 0.2 by kow, based on:

// Parametric Pulley with multiple belt profiles
// by droftarts January 2012

// Based on pulleys by:
// http://www.thingiverse.com/thing:11256 by me!
// https://github.com/prusajr/PrusaMendel by Josef Prusa
// http://www.thingiverse.com/thing:3104 by GilesBathgate
// http://www.thingiverse.com/thing:2079 by nophead

// dxf tooth data from http://oem.cadregister.com/asp/PPOW_Entry.asp?company=915217&elementID=07807803/METRIC/URETH/WV0025/F
// pulley diameter checked and modelled from data at http://www.sdp-si.com/D265/HTML/D265T016.html

//GNG consts
inner_teeth = 80;
outer_teeth = 119;

// tuneable constants
retainer_ht = 1.5;	// height of retainer flange over pulley, standard = 1.5
idler_ht = 1.0;		// height of idler flange over pulley, standard = 1.5

pulley_t_ht = 16.0;	// length of toothed part of pulley, standard = 12
pulley_b_ht = 8;		// pulley base height, standard = 8. Set to same as idler_ht if you want an idler but no pulley.
pulley_b_dia = 20;	// pulley base diameter, standard = 20


//	********************************
//	** Scaling tooth for good fit **
//	********************************
/*	To improve fit of belt to pulley, set the following constant. Decrease or increase by 0.1mm at a time. We are modelling the *BELT* tooth here, not the tooth on the pulley. Increasing the number will *decrease* the pulley tooth size. Increasing the tooth width will also scale proportionately the tooth depth, to maintain the shape of the tooth, and increase how far into the pulley the tooth is indented. Can be negative */

additional_tooth_width_inner = -0.2; //mm
additional_tooth_width_outer = 0.2; //mm

//	If you need more tooth depth than this provides, adjust the following constant. However, this will cause the shape of the tooth to change.

additional_tooth_depth = 0; //mm

// The following set the pulley diameter for a given number of teeth


HTD_5mm_pulley_dia1 = tooth_spacing (5,0.5715,inner_teeth);
HTD_5mm_pulley_dia2 = tooth_spacing (5,0.5715,outer_teeth);

// The following calls the pulley creation part, and passes the pulley diameter and tooth width to that module

//belt
//translate ([0,15, 13])
//	cube([140,1,15]);
//handle

//translate ([15,0, 7.5])
//	cube([25,5,1.2]);

difference()
{	 
	pulley ( "HTD 5mm" , HTD_5mm_pulley_dia2 , 2.199 , 3.781,outer_teeth,0,1,additional_tooth_width_outer);
translate ([0,0, 1])
	pulley ( "HTD 5mm" , HTD_5mm_pulley_dia1 , 2.199 , 3.781,inner_teeth,0,0,additional_tooth_width_inner);
	translate ([0,0, pulley_b_ht + pulley_t_ht + retainer_ht -40])
		cylinder(h = 27, r1 = inner_teeth -1, r2 = inner_teeth -20);
	difference()
	 {	 
		//cylinder(h = 15, r1 = 110, r2 = 110);
		//cylinder(h = 15, r1 = 94, r2 = 95);
	} 

translate( [0,0, 40] ) 
	rotate(90, [0, 1, 0])
		for ( i = [0:17] ) {
  			rotate( i*360/18, [1, 0, 0])
   				translate( [0,78, 0] ) 
  					rotate( 90, [0, 1, 0]) cylinder(r = 11, h=50);
 		}

}

// Functions

function tooth_spacing(tooth_pitch,pitch_line_offset,teeth)
	= (2*((teeth*tooth_pitch)/(3.14159265*2)-pitch_line_offset)) ;

// Main Module

module pulley( belt_type , pulley_OD , tooth_depth , tooth_width,teeth,retainer,idler,additional_tooth_width)
{
	echo (str("Belt type = ",belt_type,"; Number of teeth = ",teeth,"; Pulley Outside Diameter = ",pulley_OD,"mm "));
	tooth_distance_from_centre = sqrt( pow(pulley_OD/2,2) - pow((tooth_width+additional_tooth_width)/2,2));
	tooth_width_scale = (tooth_width + additional_tooth_width ) / tooth_width;
	tooth_depth_scale = ((tooth_depth + additional_tooth_depth ) / tooth_depth) ;


//	************************************************************************
//	*** uncomment the following line if pulley is wider than puller base ***
//	************************************************************************

//	translate ([0,0, pulley_b_ht + pulley_t_ht + retainer_ht ]) rotate ([0,180,0])

	difference()
	 {	 
		union()
		{
			//base
	
	
		difference()
		{
			//shaft - diameter is outside diameter of pulley
			
			translate([0,0,pulley_b_ht]) 
			rotate ([0,0,360/(teeth*4)]) 
			cylinder(r=pulley_OD/2,h=pulley_t_ht, $fn=teeth*4);
	
			//teeth - cut out of shaft
		
			for(i=[1:teeth]) 
			rotate([0,0,i*(360/teeth)])
			translate([0,-tooth_distance_from_centre,pulley_b_ht -1]) 
			scale ([ tooth_width_scale , tooth_depth_scale , 1 ]) 
			{
				HTD_5mm();
			}

		}
			
		//belt retainer / idler
		if ( retainer > 0 ) {translate ([0,0, pulley_b_ht + pulley_t_ht ]) 
		rotate_extrude($fn=teeth*4)  
		polygon([[0,0],[pulley_OD/2,0],[pulley_OD/2 + retainer_ht , retainer_ht],[0 , retainer_ht],[0,0]]);}
		
		if ( idler > 0 ) {translate ([0,0, pulley_b_ht - idler_ht]) 
		rotate_extrude($fn=teeth*4)  
		polygon([[0,0],[pulley_OD/2 + idler_ht,0],[pulley_OD/2 , idler_ht],[0 , idler_ht],[0,0]]);}
	
		}
	}  
}


// Tooth profile modules

module HTD_5mm()
{
	linear_extrude(height=pulley_t_ht+2) polygon([[-1.89036,-0.75],[-1.89036,0],[-1.741168,0.02669],[-1.61387,0.100806],[-1.518984,0.21342],[-1.467026,0.3556],[-1.427162,0.960967],[-1.398568,1.089602],[-1.359437,1.213531],[-1.310296,1.332296],[-1.251672,1.445441],[-1.184092,1.552509],[-1.108081,1.653042],[-1.024167,1.746585],[-0.932877,1.832681],[-0.834736,1.910872],[-0.730271,1.980701],[-0.62001,2.041713],[-0.504478,2.09345],[-0.384202,2.135455],[-0.259708,2.167271],[-0.131524,2.188443],[-0.000176,2.198511],[0.131296,2.188504],[0.259588,2.167387],[0.384174,2.135616],[0.504527,2.093648],[0.620123,2.04194],[0.730433,1.980949],[0.834934,1.911132],[0.933097,1.832945],[1.024398,1.746846],[1.108311,1.653291],[1.184308,1.552736],[1.251865,1.445639],[1.310455,1.332457],[1.359552,1.213647],[1.39863,1.089664],[1.427162,0.960967],[1.467026,0.3556],[1.518984,0.21342],[1.61387,0.100806],[1.741168,0.02669],[1.89036,0],[1.89036,-0.75]]);
}

My printer currently isn't tuned properly for prints this large and I'm too busy at work right now to fix it.
 
bzhwindtalker said:
Even if I have given up on the GNG as a kit, I like what you guys are doing with it and all this 3D printing stuff is amazing!


Bee,
I really like your signature :)
Are you a moto-fetishist too? :p

You're not giving up on riding that sweet setup you have, are you? Your vids are one of the main reasons I bought this kit and have been modding it for more torque :)

The price of the GNG kit makes me want to design a copy with an extra $100-200 invested into some stronger parts. This kit changed my opinion of how simple/cheap a bolt-on mid-drive can be and I think we are only a few mods away from having something very reliable, fun, and powerful.

In other news, I tried to shift under full throttle from 1st to 2nd gear after my 11:80 chain conversion and now my Deore XT rear derailleur and chain are in multiple pieces :p (chain fell off)

Hopefully the replacement gets here soon. The snow is receding from the mountain in my backyard quickly this year!

 
Short cage!

Anything with a single ring up front is better off using short cage rd. You wouldn't find any medium cage on DH bikes.

Get a clutched one if you're running 10 sp.
 
I'm running 11-34 9 speed on the rear. I'm looking at Deore XT so I think I'm limited to medium and long cage (bike came with long), I can get a medium XT for about $75 shipped while the shorter clutched XTR stuff is $150+. I figure I should go with the cheaper one in case I manage to do the same thing again, I would feel bad ruining a nice XTR. I just installed a shimano grip shift so I can't go with the SRAM stuff unless I replace that too.

Gonna dig through the GNG threads to find some more info on chain guides, looks like I'll be needing one with this setup.
 
Regarding sourcing the motor for about $35. Is that through Jon at GNG, or were y'all able to find an alternative supplier?

I found it on Mister Tao. http://www.mistertao.com/taobao-products/taobao-item-18586248326.html The selling price is great but the shipping comes to around $70! I wonder if we could do a group buy of several of these and get some other type of freight other than airmail?

Are there significant benefits other than closer pedal assist cadence that people are seeking more reduction for?
e.g. When riding overpowered I burst throttle to a speed and then pedal alone until an incline or speed drops below pedal cadence.

With the adjustable plates, will this be good enough for reliable belt life and reducing the wasted power?

Matching pedal cadence to peak motor power is the main thing. When you're trying to get up that really nasty hill it's good to have motor and human power working together. With an under-reduced system like the GNG you have a choice of partial power in too high a gear or pinning the throttle and not pedaling. The stock reduction is way off. Around 40:1 would be ideal with a 3,200 rpm motor. With 80/14 and 44/13 the GNG comes in under 20:1.

The adjustable plates will eliminate the noise and drag of the tensioner and allow you to get better tension on the belt and chain than the poorly designed GNG tensioners. With the addition of one of Bee's slip-on pulleys, belt life should be hugely extended. The adjustable sheets also allow you to install a larger driven pulley or sprocket.

Sorry belt fans, but USPS arrived with my chain rings a few hours before the UPS guy came with my larger belts, and I'm an impatient man:
11:80 gearing
I'll still be mailing a few of those pulleys out to anyone that wants to experiment with the belt some more
.

I received a packet from Bee today full of a variety of his printed pulleys. Since he's bailed on the pulley side of the equation and joined the chain gang, I'll keep fussing with the pulleys. I'm excited to finish my build and get out on the road. Thanks Bee for all of your hard work and generosity. We'll let you know how the pulleys work out. I have my eye on the 20mm 18T as a a choice solution.
 
bee said:
Here's the latest code for my larger sprocket if anyone wants to give it a try:

***fixed*** couldnt get the file to render in openscad - now it's working - 26 minutes - wow!

Also, for the others reading this - the acetone not only smooths but improves the adhesion between layers - it may have been stated above not sure. An important point to stop the pulley snapping mid-cross section if the layers weren't bonded properly, it can happen.

Love your jar with nuts F-T. There are slightly different internal infill percentages in the three pulleys I sent - sorry I should have noted them.
 
I pulled the sheet metal flange off the end of the GNG driver pulley in preparation for playing with pulleys. I decided to take a shot of the stock profile just to show how messed up it is. Looks like one quick pass with a ball nose to cut the groove and then rotate. No attempt to radius the edges of the teeth like a properly made pulley.

GNG_Driver_Scan1.jpg
 
full-throttle said:
Both XT and XTR don't come in short cage version. Since you have 9sp drivetrain and Shimano-compatible shifter your options are medium cage XT or short cage Saint.
the shimano saint deraileur is pretty beefy. I'm putting one on the GNG bike I'm building for a friend. It looks twice as durable as a standard MTB deraileur. Main drawback is the cost.
 
My printer is limited to 150mm square bed size. There's some "circa 120 tooth options" (scroll right down) for a 19mm wide version of the large pulley here, but the pitch is 5.08mm - any thoughts on if there might be a candidate?
http://www.smallparts.com.au/store/partslist/beltstiming5080mmpitchxl19mmwide/beltstiming5080mmpitchxl19mmwide/all/1/
 
Samd said:
My printer is limited to 150mm square bed size. There's some "circa 120 tooth options" (scroll right down) for a 19mm wide version of the large pulley here, but the pitch is 5.08mm - any thoughts on if there might be a candidate?
http://www.smallparts.com.au/store/partslist/beltstiming5080mmpitchxl19mmwide/beltstiming5080mmpitchxl19mmwide/all/1/

That's an XL. You need HTD5.
 
There's been a lot of back and forth about having a quality freewheel on the primary side like this one from Mel at BikeMotive:
http://bikemotive.myshopify.com/collections/bmx/products/custom-acs-crossfire-flanged-freewheel

If I made up some cast resin 23mm wide 120T driven primary pulleys would people want them with freewheels or a simple steel flange that has a keyed 12mm hub?

Again, the advantages of the freewheel on the primary side are: larger, better quality freewheel and a smaller drive sprocket on the secondary side for better reduction.
Disadvantages: You have to turn the jackshaft while pedaling. The jackshaft turns easily but you do have the secondary reduction working backwards against you to increase the effort.

It seems to me that people who pedal a lot with little to no assist would like the little GNG freewheel. People who use the motor a lot, and want as much reduction as possible, may prefer the primary freewheel.
 
For anyone looking for a top quality replacement bottom bracket for the GNG, or a solid platform to build a custom front freewheel crank set around, Phi Wood offers square taper 17mm spindle set in lengths up to 155mm. They are available in both stainless steel and titanium. They also offer beautiful alloy track arms.

http://www.philwood.com/products/bbpages/ssbb.php

I wrote to Phil Wood customer service and sent them my dimensioned drawing of the GNG BB. Here is their reply:

"Provided that the taper is a JIS taper, our standard 155mm JIS bottom bracket should work for this application. Typically we offer it in a 100mm bottom bracket shell, but we can custom build it into a 68/73 shell at no additional charge.

The bottom bracket and threaded cups are sold separately. If you need deeper cups, we do offer them. The extended cup set that we make is only wider on the drive side, and only the drive side includes a lockring.

The bottom bracket is $213, the extended cups are $64. If you're doing the installation yourself, you will need at least one of our bottom bracket installation tools. They are $16 each."

Phil's products are not cheap. But they're made in the US and the quality is tops. If the GNG kit has taught us anything, it's that we get what we pay for.
 
Back
Top