I've read this entire thread on lithium cells sold by Tumich.
I was wondering if anybody could answer few questions based on their previous experience from purchasing Tumich's cells :

1 - Was the
shipping fees expensive

I don't need to know the exact amount, just the approximate cost in orders of magnitudes (example : 20$, 50$, 100$, 200 $ ?). I understand shipping might vary a bit according to country to delivery, but it might not be so different between USA and Canada (same continent).

2 - Did you have to pay
custom fees when the parcel entered your country (I myself am in Canada)

If so, was it expensive/how much

3 - Anybody had
problems because of UN/DOT 38.3 transportation regulations

(ex : parcel held, parcel seized, parcel lost etc. ?)
I ask because I'm trying to figure the
whole real cost for the unwelded cells or never welded cells.
I'm maybe looking to order around 200 x Sanyo NCR18650GA cells (or higher load-rated cells) from Tumich next april 2017 (when the climate is not so cold that it will degrade the cells during the 2-3 months sea trip).
If I could get my hands on high-amp rated cells (30A or 20A) with an average medium to good capacity (2000 mAh or higher), that would be even better ! My prefered choices, in decreasing order of preference would be : 1st -
Sony VTC4 (30A - 2100 mAh) > 2nd -
LG HG2 (20A - 3000 mAh) > 3rd -
Sony VTC5 (20A - 2600 mAh) > 4th -
Samsung INR-18650-30Q (15A - 3000 mAh) >> 5th -
Sanyo NCR18650GA (10A - 3500 mAh) >> 6th - Panasonic NCR18650PF (10A - 2900 mAh).
Thanks in advance for your input,
Matador
PS : For those of you who wonder, I ride a 30A BBSHD on a 28inch hybrid bike (bike + motor kit + battery = 70 lbs. Me 150 lbs) with front chainring 46T and rear gears 14T to 34T. I toped off at 1600 watts and a max 62km/h on flat while testing a homemade fully charged 58.8V 14S9P used rainbow laptop cells pack (
calculated 1075 Wh at 4.5A load (0.22C) from individual cell testing : average 2300 mAh/cell tested [with Foxnovo 4S at -0.5A/cell from 4.2 to 2.8V]... worst cell 1756 mAh, best cell 2690 mAh, average DC-IR 143 mOhm/cell, best DC-IR/cell 79mOhm, worst DC-IR/cell 249 mOhm, worst voltage drift (from 4.20V) was 0.20 V over 6 month, best 0.02 V over 6 months. 1S9P subpacks were equilibrated so that each 1S9P subpack had 20.7Ah and each 1S9P subpack had a calculated DCIR between 14.0 and 15.6 mOhms).
View attachment Li-Ion.xlsx
After 20 miles ride in 2h on flat,
the battery pack temp rose to 70°Celcius. Used up 550 Whr and rest voltage was 52V.
I'm concerned about number of cycles, knowing I like to push the throttle all the time very hard. At one point, at 31-32 amps load (1.6C), I saw voltage sag of 10 volts ! (53 to 43 volts under full load... almost hit LVC and I'm just below 45%SOC!). After 10 cycles, I lost a significant amount of my battery capacity (maybe 10-15%) from the cells overheating. I want to push my BBSHD to the maximum limit (maybe even upgrade the mosfets in the future) and not have the pack heat up more than 10°C over ambiant. I want to have good life cycles and plan to use battery in 80% SOC to 20% SOC.
Hence my interest in 30A rated cells...
I want very low IR for very low heat loss.... Less temps rise mean better cycle life to me. To me a 14S4P pack made of 10A-rated 18650 cells is out of the question for a 30A BBSHD, I'll be pushing full throttle all the time. I'm not interested in a pack that will only last me 50-100 cycles. If I go the 10A-rated cell route, the pack will have to be at least 8P if not 10P. I want my full throttle habit to correspond to the cells average continous amp rating and with a conservative margin...
I've gained experience from testing around 220 used rainbow laptop cells (test for mAh, DCIR, voltage drift, heat rise) in the past last year. I learned from my mistakes. Now that I gained experience, I think I can avoid mistakes and build a pack with new and more precious (as in more expensive) 18650 cells. My ambition is to build the best battery pack that I can (to me, that means minimizing total pack resistance and favoring thick copper connections rather than nickel strips). I'm serious about this. I think I like battery pack engenering even more than building the ebike itself.