Kingsong S20/S22 (Confirmed)

[supercurio]

I found interesting that the motor has the same Kv rating as the 18XL, 16X, S18.

It means, for the same amount of phase current fed to the motor, at a low speed:

• 16X has a 16" tire: smaller wheel circumference: it's torquier
• 18XL has a 18" 2.5" tire, it's not as tall: a little more torque than S20
• S18 and S20 tires are interchangeable, although the S20 stock tire is on the larger side - larger than what would fit in a S18 mudguard: S20 and S18 have the same torque

In order to get more low-end torque than the S18 or 18XL, the S20 board and batteries must be capable of driving more current into the same motor.
Since the S20 has twice the cell count as the S18 tho, that should certainly be feasible.

Edited March 23, 2022 by supercurio

[techyiam]

Thanks to Ecodrift teardown, it looks like the S20's suspension sliders' operating life can be extended with relatively simple design changes. The T-rails are replaceable. To replace the sliders, you would have to replace the battery boxes. Moreover, the soft aluminium is going go wear prematurely, especially since it exposed to road debris. Not economically sensible. Instead, discard the four polymer bearing blocks, and slip in bearing inserts into the sliders (like how it is done on the V11). In this way, the slider/battery boxes do not have to be replaced due to wear, since only the bearing inserts actually make contact with the T-rails. Only the bearing inserts have to be replaced when they are worn. The T-rails can be made from steel or other materials with a surface finish similar to fork stanchions or better. The T-rails are relatively simple to replace, as are bearing inserts. And the T-rails can be designed and made to have lower stiction and longer service life. The same applys to the bearing inserts as well. Aftermarket opportunities are wide open.

Edited March 23, 2022 by techyiam

[RagingGrandpa]

3 hours ago, supercurio said:

Ecodrift just published their first S20 teardown !

https://ecodrift.ru/2022/03/23/kingsong-s20-razbiraem/
https://ecodrift-ru.translate.goog/2022/03/23/kingsong-s20-razbiraem/?_x_tr_sl=ru&_x_tr_tl=en&_x_tr_hl=en&_x_tr_pto=wapp (translated)

Ok, so at least the difficult axle fasteners don't need to be removed if you're just changing the tire... because the tire can fit over the side-frame once the fender and kickstand are removed. Good.

Quote

 

And cartridge seals separate from the bearings, a first for EUC's, nice!

Quote

 

 

2 hours ago, supercurio said:

I found interesting that the motor has the same Kv rating as the 18XL, 16X, S18.

• 16X has a 16" tire: smaller wheel circumference: it's torquier [accelerates faster]
• 18XL has a 18" 2.5" tire, it's not as tall: a little more torque [acceleration] than S20
• S18 and S20 tires are interchangeable, although the S20 stock tire is on the larger side - larger than what would fit in a S18 mudguard: S20 and S18 have the same torque

Beware the T-word...
Torque does not depend on tire diameter.
You had the right idea, but really you're talking about acceleration (resulting directly from a force at the tire, and not a torque). 

The S18 had a freespin speed of 67kph @ 82V. 
So if we've got the same motor at 120V now, voltage says we should expect freespin at 98kph. And it seems so. 

This is good: we could get EX.N speeds, and 16X torque (but only if the controller would allow 200A ). 
And like you said, due to tire diameter, 16X can still accelerate faster than S20, with the same motor, at the same current. Fine.

[Tawpie]

3 hours ago, supercurio said:

S18 and S20 tires are interchangeable

Not that I'm already needing a tire change, it's nice to know that the K66 will work (if you can find it of course) and even the H666 as well (if you're not a mud runner, want good tread wear and don't mind tram lining). Plus all of the other street tire options will be available as well!

Of course, a shorter tire will mean lower top speeds in exchange for (slightly) more force at the tire/earth connection.

[Brendan "nog3" Halliday]

12 hours ago, rolis said:

And we have got the mosfets! But they are mounted with the necessary info the other way so we still dont know 😂

Since it's not possible to safely unbend them and inspect, I asked our KS sales rep. They've advised they are JieJie JMSH1504AS 150V 4.0m N-Ch Power MOSFET.

They have also confirmed there is a new firmware coming at the end of the week with more aggressive acceleration by default.

[Tawpie]

18 hours ago, Brendan "nog3" Halliday said:

they are JieJie JMSH1504AS 150V 4.0m N-Ch Power MOSFET

Should be adequate! Thank you @Brendan "nog3" Halliday!!!

https://jjwdz.pro.tokeys.com/file/fileDownload?id=10224&&type=f 

 

Edited March 24, 2022 by RagingGrandpa
(fix url)

[Resko727]

2 hours ago, Tawpie said:

Should be adequate! Thank you @Brendan "nog3" Halliday!!!

 For the not so tech savvy of us. Can you break this down some any good overbuilt? Or underbuilt? 

[Tawpie]

36 minutes ago, Resko727 said:

Can you break this down some any good overbuilt? Or underbuilt? 

Underbuilt? Because of the V12 debacle, the thing we were concerned about was "which device did they choose given they've increased the operating voltage". What we're looking for is a MOSFET that can handle the battery voltage plus 'room' for voltage spikes and the regeneration cycle during braking. This part is rated for a maximum of 150V Vds, which is more than 10% higher than the nominal fully charged voltage of the battery pack (126V). The 10% figure is arbitrary because I don't know what the actual peak voltages will be, but a max rating of 150V should be enough to avoid the issues the V12 is experiencing. In the case of the V12, their MOSFETs were rated for a maximum Vds of 100V, and the fully charged battery is 100.8V so they're already over the max limit before any other factors eat into their design margin (temperature mostly, Vds max goes down when the part gets hot) come into play.

The second thing I glance at is the maximum current rating of the part. This is the continuous drain current on the spec sheet (Id). You see that max current depends on the temperature of the junction (internals) of the part and as the part gets hotter the current it can handle goes down. In this case, the part is rated for 128A at 100C. The motor will melt if it's asked to 'use' more than 3300W for a long period of time, and the MOSFET can deliver 126V * 128A (theoretically) or 16,128W. This means that the MOSFET should be able to handle enough current to handily melt the motor wires, and the current the MOSFET can handle far exceeds what the battery pack can deliver (unless the battery pack is protected by the BMS, it will blow up if you tried to pull that much current out of it). Consequently, the MOSFET shouldn't overheat IF and only IF it's properly cooled.

Next we look at the package of the part, in this case it's a TO-247 which as better heat transfer characteristics than the TO-220 that has been common in the past. "Better" meaning that heat flows more rapidly from the part's internals to the metal heat tab so the temperature of the part's junction can stay lower. From the part's heat tab, there's thermal grease to a heatsink and the heatsink is thermally connected to the frame via that squishy pad. Having the frame as part of the cooling is really nice as it provides a grundle of metal for the heat to go into and lots of exposed surface area so the heat can transfer to the surrounding air. All this means that the wheel should run cool and the MOSFETs shouldn't melt. It also means KS didn't need a fan. Or two fans. Don't remove the squishy pad, and treat it gently. VERY gently.

Overbuilt? Not really. You can spend quite a bit more money for beefier MOSFETs that can handle more Vds and more Id, but at some point the bean counters show up and kick over your desk. Also, we don't know anything about the drive circuit, and the drive circuit is the next thing that you need to have designed properly. But at least the MOSFET looks like it passes the "sensible choice, it should have enough design margin for stuff to happen" test on paper. Real life is always different. Here's another of my favorite sayings:

"In theory, theory and practice are the same.
                     But in practice…"
                                          I don't know the source

OPINION: looking at the packaging (how the board is put together and installed into the 'module'), they've been working on this for quite some time... I'd guess they started the electrical design in earnest in late 2020, probably mid 2020. It doesn't have the look of something that was thrown together at the last minute in response to market forces. I think they had a good deal of time to think about it, test different ideas, and arrive at this implementation.

Edited March 24, 2022 by Tawpie

[Tawpie]

29 minutes ago, Tawpie said:

the current the MOSFET can handle far exceeds what the battery pack can deliver (unless the battery pack is protected by the BMS, it will blow up if you tried to pull that much current out of it)

I tried and failed to edit that, the choice of "far exceeds" is a bit of hyperbole.

What I should have said was the battery pack would last about 8 minutes with a 128A continuous draw. The BMS will surely forbid this as the batteries would get quite hot. The battery system will be the limit on power delivery, not this MOSFET. The battery pack is 30s4p, so it absolutely could supply 128A continuous without blowing up, but such a plan would be unwise.

Even at 100% PWM (full max power, about to eat pavement, KS won't let you go there), each individual MOSFET handles at most 1/3 of the motor current so you can safely assume that no MOSFET will see the equivalent heat load of 128A continuous.

Edited March 24, 2022 by Tawpie

[Paul A]

Think Tawpie is saying:

Wires will melt, batteries will overheat, before the mosfets will fail.

But the better heat sink and BMS (battery management system) will not allow that failure.

Therefore, it will be ok.

[Tawpie]

1 minute ago, Paul A said:

Therefore, it will be ok.

🤞

At least it's not "they chose whaaaaat?". So chances are pretty good, better than the V12.

[Paul A]

King Song Factory Introduction video

3,485 views

Dec 9, 2019

King Song International

3.56K subscribers

 

___________________________________________________

New King Song Factory Brief Tour

1,110 views

May 27, 2018

electricunicycle.org

536 subscribers

 

 

[Blunzn]

S20 pad test in french. Grizzly, Clark v2 & standard pads.

he also switched the pedals to the forward position. 

 

[Patrick Robert]

32 minutes ago, Blunzn said:

S20 pad test in french. Grizzly, Clark v2 & standard pads.

He liked the Grizzlas the best, the only ones which allowed him to brake adequately. 

[RagingGrandpa]

On 3/23/2022 at 8:54 PM, Brendan "nog3" Halliday said:

I asked our KS sales rep. They've advised they are JieJie JMSH1504AS

(And thanks @Tawpie for the datasheet!)

But, it's not very impressive  

That is: if we were expecting "Gotway performance," I look for a maximum output condition of 100A per FET @ 40us on-time, at full battery voltage. 
But as shown above, this JJM 150V FET is only meant to supply about 50A for that condition. 
Compare it to our often-seen Hooyi 125V FET, which is rated past a whopping 600A for the same pulse.

The same story is told by the power dissipation ratings: It's just a lower-power FET.
  JJM 150V:          167W @ 100°C
  HY5012 125V:   250W @ 100°C

My conclusion: this JJM FET can be safe and durable, but Kingsong won't be able to set the firmware to command as much current with it, as they could have done if a stronger FET were used. And so S20 will never be capable of the low-speed motor torque that we have with 84V and 100V Gotways at >100A pulses per-FET. 
(This also seems to match the demo rider feedback, which often mentioned weak performance from a standstill.)

For us torque-lovers, perhaps in a future wheel they'll find room for a 3rd parallel FET, or change to a more expensive part with higher current ratings.  

 

On 3/24/2022 at 12:36 AM, Tawpie said:

So chances are pretty good, better than the V12.

Yes, better than V12.
Thankfully, this part is not fundamentally "outside the specification," like V12 had with its strange choice of a 100V-rated part in a 100.8V system voltage.

 

Edited August 17, 2023 by RagingGrandpa
(img url)

[Chriull]

26 minutes ago, RagingGrandpa said:

That is: if we were expecting "Gotway performance," I look for a maximum output condition of 100A per FET @ 40us on-time, at full battery voltage. 
But as shown above, this JJM 150V FET is only meant to supply about 50A for that condition. 

The operating condition for these fets is either conducting or switched off.

Just while switching there is some, very short time while the mosfet is in the "linear region". 40μs would be very slow switching.

So the points Vds from 1-2V with max 100-200A, just imited by rdson are of interest.

[mike_bike_kite]

2 hours ago, RagingGrandpa said:

That is: if we were expecting "Gotway performance," I look for a maximum output condition of 100A per FET @ 40us on-time, at full battery voltage. 
But as shown above, this JJM 150V FET is only meant to supply about 50A for that condition. 

I know nothing about electricity other than not to lick loose wires but doesn't Gotway/Begode report current differently to other manufacturers in that they show 3 phase current which reports much higher.

[Rawnei]

126v x 50A = 6300W doesn't seem that bad though? Or am I totally missing something. 😁

[RagingGrandpa]

20 minutes ago, Rawnei said:

126v x 50A [is a lot of watts]

Wrong idea, sorry...

Gotway permits about 220A peak current in the motor. Aka "phase current." And this 220A pulse is only happening at the peak-current position of the motor as it spins; then current drops again in a sinewave-shape.

Voltage is applied to the motor in a regulated fashion (PWM), so the average voltage is much much less than the battery voltage. At low speeds, a relatively low voltage is needed to motivate those 220 amps.

Calculating average power from FET pulse current is not practical with the limited information we have.
But power really isn't the topic here. Low-end acceleration is a current-limited problem, not power problem.

[Planemo]

It frustrates me why people try to reinvent the wheel (no pun). The HY5012 are tried and trusted. They just work. Given this knowledge, if we do end up with wheels that need more current capability we could just add more 5012's together (within limits). So all bases are covered yet we see these ones from JieJie (anyone even heard of them before?) being fitted.

I guess the only excuse I could understand is that the 5012's are unavailable.

Truth be told I'd like to see massively over-specced FET's fitted to ALL euc's. But then I could say that about capacitors too. At least XT90's have started to be used so I should be thankful for that I guess..

[RagingGrandpa]

3 minutes ago, Planemo said:

HY5012 are tried and trusted
we could just add more 5012's together

They're great for current; but can't handle the S20's system voltage @ 126V. 
Adding more together in series would nearly double the physical size of the controller... not a great option. 

So I think HY5012 can't be the answer for our new generation of >120V EUC's; but there are certainly more attractive chips out there, than these JJM.
I suspect the choice had a lot to do with availability, with semiconductors being a strained commodity right now...

[Paul A]

The constraint of weak performance from a standstill seems quite minor.

Might not be much of a concern except for extreme riders like Chance.

[Rollin-on-1]

21 minutes ago, Paul A said:

The constraint of weak performance from a standstill seems quite minor.

Might not be much of a concern except for extreme riders like Chance.

Not to mention the report that new FW will be out this week which will improve the accelleration.  Although it remains to be seen how much improvement we will actually see.

[Tawpie]

40 minutes ago, Paul A said:

Might not be much of a concern except for extreme riders like Chance.

That makes me glad I'm not as good as Chance! And that I don't ride in traffic.

[supercurio]

45 minutes ago, Paul A said:

The constraint of weak performance from a standstill seems quite minor.

Might not be much of a concern except for extreme riders like Chance.

Depends, for technical off-road - like going over various obstacles - stopping for an instant to change direction then lurching yourself forward again to jump or bonk another obstacle, especially when combined with the suspension, getting as much power as possible at any speed is really important.
It's the kind of stuff I hope the S20 to be good at also. It would be a letdown for me if it the low end torque is not the best.

And in essence it would also mean that the idea of rising the voltage to 126V with the idea to get high torque and high speed, but then not getting the high torque capability due to the limited MOSFET choice would be kind of pointless.
Making all the 100V fast chargers incompatible in the process, etc.

Let's see what KingSong can get out of this configuration.
While it's a very problematic example, Inmotion got very good performance out of dramatically undersized MOSFET on the V12 for instance, reputed for its zippiness, impressive acceleration capabilities despite all wrong components paired with a 100.8V system and a HS motor.

So the S20 should at least perform better than the OG V12 HS: more torque oriented motor and beefier MOSFET pulse capacity.
Compared to the V12 HT tho, hmm probably the V12 HT will feel significantly more powerful and responsive within its narrower safe speed range.