3D Printed NT carb filter adapters and various testing results - In progress thread!

[FNTPuck]

Was bored yesterday and saw the 30mm cone filter from the Keihin clone I purchased lying around and wanted to use it on the stock NT carb, but didn't want to spend the $20 and wait for the velocity stack adapters. I fired up sketchup instead, and started making some prototype adapters 3d printed out of PETG, which is a hard plastic with high heat resistance that doesn't soften much until ~140-150c, which is almost 300*F so should be fine for the intake side of a bike (print temp is 235c). It is a bit brittle though compared to something like bridge nylon, so I printed with 100% infill but may need to eventually add some gussets and reinforcements - my guess is the weak point will be right where the base meets the tube for the filter attachment due to the weight of filter and vibrations, but that will be easy to reinforce if needed.

First print was just to test the dimensions of the mounting area, but I thought it may make a good low profile filter so I threw it together and tested it. It uses the stock hardware, filter, and retention screen but I wasn't expecting to actually use it so it doesn't have the notches for the retention screen. To make it fit and prevent the filter from getting sucked against the carb I just used an xacto to cut them off and make it round. It is 40% thinner, and doesn't have the ~1/2" protrusions coming off the bottom so tons of room even with the stock non-offset intake.

Test bike setup is a stock 66cc with CNC head, supercharge CDI, NGK plug @ .025, two ~1/4" holes drilled across muffler baffle tube (one on top one on bottom of the center tube, plates untouched) and a 39t rear sprocket on a heavy Huffy Fairmont cruiser. Motor is brand new with less than 5 miles on it, so nowhere near broken in, and using 16:1 break-in ratio. Carb is all stock with needle on 2nd from top notch.

Note that much of this testing is subjective and not super scientific - I am just using a cell phone GPS app and doing 2x identical laps around my neighborhood and comparing top speed but I'm not using decimals so its not super accurate. Route has one short downhill, one long uphill, and one long straight so its a great test lap. Also, things like idle quality and smoothness are hard to quantify but I am trying to be as accurate as possible.

Stock filter with "NT" circle logo drilled out to help it breathe(+trimmed braces that block intake on retention screen, as pictured):
31mph, idle was properly dialed in after warm up. Noticably soft low end but stock jetting is pretty close, not lean or rich. (New motor, 39t sprocket, heavy bike, break-in fuel, etc is likely why low end is weak. I'm used to 41t or 44t sprockets and this is my first 39t).

Low Profile(click for big pics):




Same 31mph, idle is higher without touching anything so it definitely flows better at least down low. Mid range is cleaner, seems to rev better. Top end felt better and I got excited but GPS said otherwise...same exact top speed. I'm guessing top speed is the same since the stock filter media pushed up against the carb intake is probably the restriction now instead of the housing itself...or maybe drilling out the NT logo on the stock housing increased airflow over stock enough to limit the difference this much larger intake hole makes. Dont have a stock sealed one to test but may go back to the stocker with that hole taped up to add to results for a true bone stock filter comparison.

Next test will be the full design print with the cone filter attached and the intake extended ~2" and to the side to clear. I'm gonna make the design a bit smaller diameter with a slight overhang to not have to use stock retaining screen. Will hopefully update later today depending on how much time I have to work on it.

Unfinished model for the cone filter adapter:

 

[FNTPuck]

**Warning, boring techy numbers ahead***

Intake area comparison, plus some requirements math:

Stock Punisher/Cthulu intake area is 324mm^2. Drilling out the NT logo increases it ~400mm^2. (Boring math: 4x 9mm diameter tubes and a ~70mm^2 drainage hole on bottom that I am counting even though it is on the side of the filter and not in front of it. Drainage slot is a 5mmx10mm rectangle with addt'l 5mm dia half circles on each side=69.625mm^2). That is 324mm^2 of total intake area, of which all of it needs to make a sharp 90* turn immediately into the carb. The one I tested also has a 10mm hole I drilled right through the center, which adds 79mm of straight air path making total intake area of the drilled one ~403mm^2.)



Low profile and cone adapter version are both straight 30mm intakes, total area ~707mm^2, 218% more area. Also, all the air is in a direct path with no turns to make. If the stock intake is any restriction what so ever, even the low profile w/ stock filter sheet should be an improvement. With it removed and just a high flow cone filter, increase should be even more dramatic. Note that in preperation for the cone filter+longer adapter, I did a cheap and dirty offset setup that also coincidentally extends the intake path by ~2in.


Now, all of these numbers are wwwaayyyy more than these little motors seem to need at first glance. At 6500rpm and 55% VE(2strokes are 55% VE according to HPWizard.com) the CFM required from a 66cc 2stroke @ 6500RPM is only 4.16cfm. At 6k, the CFM required drops to 3.85cfm. The problem is, the crappy stock filter is not only in a terrible housing with only 324mm^2 of intake area, but at these small sizes foam is very restrictive and needs to be spaced farther from the small ~13mm carb inlet to use more of the available filter space.

Supporting story from one of my failures: I once tested the stock filter sheet right up against the carb when a mountain bike frame didnt have room for the housing and was surprised when the bike would idle fine but once you gave it gas it would barely run under its own power.

Why? When you do the math, a 13mm carb inlet is only .2in^2! As an empty pipe that would be over 20cfm without any restriction, which is tons of flow for these tiny things. According to bobsmuffler.com, an old website full of automotive tech, the "best" foam filters flow 4.38cfm per sq in. I had to double check my math since it sounded so crazy, but that means with the filter stuck up against the intake you are looking at only .9cfm!! That explains why it would barely run, since .9cfm starts choking the motor at only 1500rpms. More pertinent to this discussion, you only get 2.2cfm total from a stock filter setup (324mm^2=.502in^2, times 4.38cfm per sq in = 2.2cfm total). Based on the engine size x rpm, these motors need at least 3.85cfm to fully breathe at 6000rpm. This means the stock intake, at BEST, starts to become a restriction at 3500rpm. Drilling out the NT logo 1/2" for the extra 80mm intake area raises that to ~4200rpm before it starts to be a restriction, which explains why when I first did it I noticed a crisper mid range but zero increase in speed - all it did was move the restriction point higher in the RPM range but still quite a bit away from WOT. Note that obviously doesn't mean the motor just wont work past 3500 or else none of the stock motors would run very well...it just means that after that RPM, the stock setup is a restriction and any intake upgrades will be beneficial from that RPM onwards. This also does not factor the gas reversion that wets the filter, further reducing its flow rate.

---------------------------------------
Haven't even made the V1 w/ adapter yet but for V2, I plan to straighten out the air and reduce turbulence by designing a velocity stack insert that screws in place of the stock retention screen, and for the low profile ones it will have a recessed area to make sure the filter sheet is not smashed against the carb outlet so it can breathe. This will allow the full 707mm^2 of exposed filter to be used on the low profile setup, theoretically flowing 4.7cfm and being enough for a stock motor at up to 7300RPM. Above that, due to the bolt locations only allowing a max of 30mm of the stock foam filter to show, the adapter style that converts to a foam filter will be the next step for gains unless I space the filter from the cover AND the carb...which won't make it low profile anymore, lol.

Please feel free to critique, brainstorm, or test my math...I tend to "boredom drink" while designing things so sometimes I miss simple stuff! .

 

[FNTPuck]

Still on the Low Profile setup, I spaced out the filter media from the carb more and am finally seeing some changes. Bike now has a lean bog, so its definitely doing something. Moved needle jet down one and the mid range cleaned up, moving the bog to just the last bit of throttle so it looks like it needs to go one up on the main jet which I will do when I switch to the cone filter. Needing to go up one jet size is a good sign that even with the stock filter, the new housing flows much more.

________________________________________________________

Moving on to the adapter one, I updated the model before I try the first full run print. Tapered the tube at the base so it is thicker on the bottom since I was worried about that being the weak point. Also added a small flange on the end to help keep the filter on by grabbing the ledge inside the filter boot. It is oversized by ~1.5mm so the rubber filter boot will slightly stretch over it so hoping it will grab tight enough to not have to use a clamp.

Grabbing something to eat then gonna set it to print - will take roughly 2hrs15mins at 100% infill so may not have any updates today besides maybe some completed print pics.


All the edges are because I am using the standard "circle" polygon settings of 24 sides since it is fast and makes it easier to quickly line up stuff by eye since there are points to reference. A finished model will be an actual circle, not whatever you call a 24 sided polygon lol.

 

[FNTPuck]

Got the prototype installed and it was pretty close first try. I edited the model to enlarge the outlet by ~2mm for a tighter fit on the filter, make the flared end deeper, and recessed the bolt holes so the filter can fit all the way flush...you can see in the next pic that the bolt heads keep it from fitting to the end. It's not loose enough to fall off or anything, but I want to be confident no air would come in from around the filter so need it nice and snug. For now I just tossed a ziptie around it and it seems perfect.


Calling it done for today, tomorrow I will test it and jump up one jet size to see how it likes it. I think the SBC valve cover filters from autozone are similar diameter, so might be a cheap local source for a better quality filter for the same price or cheaper than these flimsy garbage ones. Need a proper offset intake to put the carb back to facing straight, for now I just hacked the end off a spare intake and used it for a ~45* angle with as short of a piece of hose as I can to make sure its nice and firm. Too long of a hose will make the carb flop all over the place and foam up the fuel.

 

[gary55]

Air Filter - Standard - Sick Bike Parts

 

[FNTPuck]

Air Filter - Standard - Sick Bike Parts

Yep, those are one of the ones I checked out since NT sized filters are hard to find that dont use an adapter/velocity stack and it is a proper washable filter unlike the junk one I have that came with a gasbike carb. I almost ordered that exact one weeks ago but SBP wanted to charge insane shipping, making it like $20.

I had this filter lying around already, so decided to make my own adapter to use it . Coincidentally I ended up with a low profile housing as well which I can use on the stock motor mountainbikes with limited room for the filter and save the $10 on an offset intake and get a bit more air flow at the same time. Cost me a whopping ~1.50 in filament per housing lol.

 

[gary55]

Half the fun is in the tinkering. The shipping is a pain. I usually buy 4 or 5 at a time to ease the shipping pain a little. I think it goes down to about 16.50 ea. on a order of 4. A lot to be said for having parts on your bike that you made yourself.