Dogati Electric Superleggera

[melodious]

Where do you get the honeycomb patterned aluminum from? And what is it used for in the real world

 

[j3tch1u]

ya, i will have clear acrylic side panels epoxied to the sides so i can watch the wax melt. for the real battery, i'll epoxy one side, fill the wax not quite to the top and epoxy the other side (to give the epoxy more honeycomb to grab). using an ir thermometer, i measured 48-60 degrees during the white soft transitioning to clear liquid phase. i'll have the pcm vacuum sealed as well to see if it is even possible to ignite. i'll do more tests to see what the flash point of this wax is.

in the end i will do a overcharge destruction test of one cell to evaluate the PCM's effectiveness against thermal runaway.

Be sure to have video

Interesting idea. The outer housing will need to be sealed to prevent the wax from dripping out. Do you know what the melting temp of the wax is? It would be nice if there was a non-flammable PCM.

 

[j3tch1u]

well melodious (love that nickname), i bought the honeycomb core direct from a factory here in taiwan. there are many applications, from aircraft, formula chassis, architectural panels, ie. anything that has to be strong and light. they come in aluminum or nomex/kevlar core with laminated panels using stainless steel, aluminum, fiberglass, ceramic or carbon fiber. the company had all kinds of combinations and gamut of applications. they had a sample honeycomb core encased in perforated stainless steel...when you dropped it, there was no sound..total dampening of resonance. there was also this big long piece of carbon fibre honeycomb core that looked exactly like a penis (i'm not kidding). it was for a cargo scooter. for laughs, i held it up to my buddy (he is black) and our other buddy was just about to take a picture when the company boss walked in (he didn't see). it's so funny cuz this dude was all business and total poker face, and here we are acting like a bunch of clowns. i really would've liked to have gotten that shot for you guys :lol:

Where do you get the honeycomb patterned aluminum from? And what is it used for in the real world

 

[fechter]

The wax may be a good solution to battery heating issues. Especially with Lipo pouches, they need to be protected against vibration and impact, but any kind of foam will insulate the cells and trap in the heat. Paraffin is flammable though. A sealed container that keeps oxygen out would keep it from burning (mabye). I'm thinking of those puffed Lipos that spontaneously explode. Paraffin will also expand quite a bit with temperature, so you need to accomodate that.

With your transparent battery covers, you could tell if one cell gets hotter than the rest and only melts the wax around that one.

OK, a little research shows a product that has about the right melting point but is non-flammable:
http://www.halocarbon.com/oils/halocarbon_40-600-1500_data_sheet.php

I bet it's not cheap though...

 

[j3tch1u]

thanks for the advice fechter. i've got the honeycomb encased in the wax now and boy does the wax add a lot of weight (this is something i forgot to consider). so i looked up the molecular weight of standard paraffin and it is around 450. plugged that figure into my solidworks model and it works out to about an addl 600g (after the holes are cut out) to the pack. i will still carry out some pcm tests, but i will probably go with a simpler solution using 18650 cell holders that provide an air gap and support (with fans blowing thru the cells). the weight of the wax and flammability are definite drawbacks (not to mention a lot of messy work).

btw, would you know how to rate polyswitch ptc fuses for the parallel and series connections (i am doing 7P14S). i am guessing the parallel connections each need a holding current ~ cell current (10a) and trip current higher than that (mainly to protect the parallel string if a cell shorts). am i correct to assume only 1 series fuse is needed @ pack voltage/peak current?

The wax may be a good solution to battery heating issues. Especially with Lipo pouches, they need to be protected against vibration and impact, but any kind of foam will insulate the cells and trap in the heat. Paraffin is flammable though. A sealed container that keeps oxygen out would keep it from burning (mabye). I'm thinking of those puffed Lipos that spontaneously explode. Paraffin will also expand quite a bit with temperature, so you need to accomodate that.

With your transparent battery covers, you could tell if one cell gets hotter than the rest and only melts the wax around that one.

OK, a little research shows a product that has about the right melting point but is non-flammable:
http://www.halocarbon.com/oils/halocarbon_40-600-1500_data_sheet.php

I bet it's not cheap though...

 

[fechter]

btw, would you know how to rate polyswitch ptc fuses for the parallel and series connections (i am doing 7P14S). i am guessing the parallel connections each need a holding current ~ cell current (10a) and trip current higher than that (mainly to protect the parallel string if a cell shorts). am i correct to assume only 1 series fuse is needed @ pack voltage/peak current?

Sounds right to me. I would suggest PTCs for the parallel strings but a standard fuse or breaker for the main pack. PTCs trip sort of slow in the event of a short.

 

[korpin]

thank you john. i find building an ebike more of a challenge than a software only project. it encompasses a lot of different engineering disciplines--mechanical, electrical, software, physics and some chemical. coming from an industrial engineering background, everything is about automation. forces you to see every piece in production terms--low-cost, hi-quality, reproducibility. the challenge is to minimize the materials and actions (processes) from stock to finished product. take for instance the lowly hex cap screw--i will try to use the same screw in many different places just so it can be sourced in higher quantity, use the same tooling, take up less organization space, minimize human error and require less brain cycles to use. i apologize for digressing into a lecture on operations research

that's what I like about my Toyota tundra....almost all the "user accessible" bolts on the engine are 11 mm and many have the same thread so if I loose one I can borrow one from a non critical area

 

[korpin]

couple of questions:

you say 24" wheel...you mean 24" total or just the rim size...(I know bicycle and motorcycle wheels are sized a bit different not sure on moped wheels)

and amount of swingarm travel possible?

 

[korpin]

wow spent the entire weekend going over this thread and all the links came away with a couple thoughts:

1)Regarding the geared hub...Rolhoff not strong enough? they are used on 2000 watt optibikes hope that 300 nm sturney archer works out seems like EVERYONE is looking for a non cassett solution to mid drive motors! (love to know final solution u come up with thinking of a revolution frame from HPC with mid drive want to ditch chain tensioner!

2)the clutch mechanism...am I to understand the main purpose is to get brake regen?..seems like a lot of weight/trouble but perhaps it also helps with no assist pedaling?...but freewheels should take care of that?

I am not as technical as the other guys anybody that can explain in plain English thanks in advance!

 

[trevc2]

Epic thread indeed. I've been following this saga for a while. I love the creativity and style you've shown with your build.

Looks like you've upgraded to two piece handlebars, how are they? Which model did you choose?

Cheers

 

[kfong]

Ben, the graphite PCM I saw seemed fairly light. It appeared to be mostly graphite with wax as the binding agent. I assume the graphite conducts most of the heat and the wax does its PC magic. It doesn't go liquid, but goes soft from my understanding, being less messy but still getting the benefits of PC. Graphite can be purchased for composite model making. I use it to add UV protection.

 

[liveforphysics]

PCM material is more fussy to setup than you may assume. The trick isn't just getting the right blend of waxes for your desired melting point. It's about getting the right melting point of waxes in a solid structure that both thermally conducts well enough to help, and still capable of holding it's shape and not separating liquids/solids when it's melted.

I would see if you can simply buy unpopulated machined 18650 holding PCM blocks from Allcell. Likely cheapest/fastest/best result path.

 

[hillzofvalp]

OR say you're a student and get them for free like I did (in raw form, however)

 

[j3tch1u]

Dogati V2 coming soon.
high res

 

[j3tch1u]

 

[Byte]

Beautiful. Amazing. How long did it took you to make those CAD files? This whole build amazes me! Astonishing things you're building!

 

[briangv99]

after riding a zero-s (thanks again luke for that memorable night ride) i decided dogati NEEDS to go sensored. so..i picked up one of these capacitive commutation encoders: http://www.cui.com/product/components/encoders/incremental-encoders/modular/amt303-v for a quick drop in solution. here are the steps i took to make it work (btw, these things are pretty tiny):

1. chucked the 3220 rotor onto my hobby lathe
2. drilled and tapped an M4 thread in the shaft (rear)
3. used a countersunk M4 screw to attach the encoder (the countersink helps secure the plastic encoder bushing firmly to the shaft axially and radially)
4. used some hot glue to temporarily secure the encoder housing.
5. connected the encoder to the programming board and set: direction (CCW), output (QAQB), resolution (1024), # poles (8) and left the commutation offset at 0.

this board has a nifty auto-zero function that gets stored into eeprom so you don't need to fiddle around with the alignment. however this only works if you know the correct commutation offset (for wye, offset is 0). i wasn't sure what the offset should be for my delta winded 3220 so i still had to adjust the encoder. running sensorless it draws 1.5A at WOT so i turned the encoder until it matched that. hooked it up to my kelly kbs controller and..the start-up is as smooth as margerine, niiiice

one thing to be aware of, you need to short GND_C with GND (pins 14 and 4) when hooking up to your motor to put it into normal operating mode (vs spi comm mode). it's written in really small print on the manual and easy to miss (this had me puzzled for a few hours).

so there you have it, a sensored 3220 for $35 bucks from digikey (http://www.digikey.com/product-detail/en/AMT303-V/102-2051-ND/2278847?cur=USD) and about 40 mins to set it up.

View attachment 1


chucked back up onto the lathe to check commutation signals against BEMF on the scope. lookin good [thumbs up]
[youtube]-FmfJFFl9y0[/youtube]

Hey Ben, how are you finding the performance of the Astro/Encoder/Kelly KBS. It sounds like the ideal setup for buttery smooth throttle control. Any issues? noticed you were enquiring recently on Burtie's optical encoder thread about sourcing some of his optical encoders, so interested if you had any difficulties with the encoder.

cheers

Brian

 

[j3tch1u]

Hey Ben, how are you finding the performance of the Astro/Encoder/Kelly KBS. It sounds like the ideal setup for buttery smooth throttle control. Any issues? noticed you were enquiring recently on Burtie's optical encoder thread about sourcing some of his optical encoders, so interested if you had any difficulties with the encoder.

atm i've only bench tested the astro w/ cui encoder and kelly (ie. i did not have the chance to mount it on the bike). i don't see any issues w/ using capacitive or optical encoders (maybe halls due to the interference). 1st batch of dogati will have optical encoder similar to burties. i'll also be running a sine controller and not the kelly.

 

[j3tch1u]

Beautiful. Amazing. How long did it took you to make those CAD files? This whole build amazes me! Astonishing things you're building!

it's been an incremental process over 5 years (but i wasn't always working on the bike). not sure of the total cad-hours. there have been many design changes and endless tweaks. in each iteration i try to seek a balance between aesthetics, cost, ind standards, serviceability, reliability, upgradeablity, durability and ease of manufacture/assembly. the cad design is the most enjoyable part for me...i consider it as play and not work. ride-testing would be #2. going to the factories to do QA checks would be last.

 

[kudos]

Drool. So are you actually going to be making these for sale??

Kudos

 

[j3tch1u]

i'm only going to be making 10 and most of them are already spoken for. however, once the first batch is sold and if there are no issues we'll do a 2nd run.

Drool. So are you actually going to be making these for sale??

Kudos

 

[ghettoracer]

[youtube]tEPWz54E0pU[/youtube]

 

[ghettoracer]

Ben sent drawing to China and in less than two weeks and less than $20 later we have 10 prototype PCB's!! gotta love the speed and efficiency of China! imagine how much this would've cost in USA.

modifying the motor from dogati 1.0...

black and white checker board is simply printed and pasted...

120 degree separation, there is also a temperature sensor...

ready to assemble and test!

 

[fechter]

Nice!

It looks like there's a little timing shift between the rotor and the encoder disk. I guess it doesn't matter as long as the sensors line up at the right spot, which could be anywhere.

What holds the circuit board on? Just the glue?

 

[liveforphysics]

Looks fantastic! Im very excited to see how it runs.

Great to see you Frank.