Adding capacitors to improve performance

[Planemo]

I know a few who have added big caps, and they claim the difference is noticeable. But if the wheel isn't dipping/cutting out I don't see how it's possible to tell the difference.

Personally, if I had the spare space, I would just look to fill it with more cells to get additional range/headroom/reduce sag. But thats just me.

The Sherman doesn't have any spare room so I never looked any further at it 

[Chriull]

17 minutes ago, Planemo said:

I know a few who have added big caps, and they claim the difference is noticeable. But if the wheel isn't dipping/cutting out I don't see how it's possible to tell the difference.

This should not really help against cutting out - maybe some dipping.

The caps just provide (minimal) power but this very fast by them beeing nearer to the mosfets. So they provide the short power spikes. Once the capacitor legs in my wheel broke it drove like in a very strange super soft mode... By the wire inductance the needed short power demand cannot be provided by the battery.

Another big difference could be the temperature of the two big buffer (?overburdened?) capacitors on the motherboard. 

17 minutes ago, Planemo said:

Personally, if I had the spare space, I would just look to fill it with more cells to get additional range/headroom/reduce sag. But thats just me.

The Sherman doesn't have any spare room so I never looked any further at it 

Why even more batteries for a sherman? 

[Bizra6ot]

What's about plug&play additional caps from EucYou is it worth it has anyone tested it?

https://www.instagram.com/p/CWIxGHhs_7b/?utm_source=ig_web_copy_link

[Planemo]

2 hours ago, Chriull said:

Why even more batteries for a sherman? 

As you suggest, it's not needed on the Sherman, but it worked well on my MSX 1860wh conversion - much rather the additional pack than a couple of caps. Talking of which, the ones in the pic seem to have pretty thin wires given the job they are expected to assist with

 

 

[Tawpie]

Not wanting to get too nit picky or anything, but "more is always better" when applied to "buffer caps" isn't as simple as it sounds. The energy stored in a not-supercapacitor is pitifully small, there's no way it'll save you from an over lean event. Accelerating to get out of the way of a car is going to mean you're leaning hard for a second or two minimum—not the few hundred milliseconds of power that may be in your buffer caps. And unless you add additional circuitry, once the cap has discharged the battery is going to dump some of its (limited) current into the caps to get them charged back up again—leaving that much less for the motor.

But. As long as their presence doesn't mess up the various control loops keeping you upright, and you can stand the additional failure opportunities introduced by parts whose 'values' are hand written in sharpie on shrink-wrap... nah, not for me.

Edited December 14, 2021 by Tawpie

[Chriull]

2 minutes ago, Planemo said:

Like how much power can be stored in the caps, and for how long it can deliver it for, Vs the amount of power available from the packs?

As i see/understand this one has to think more or less in pwm cycles.

The buffer capacitors at the mainboard are necessary for the mosfets to perform the pwm switching - without them by the wire inductance from battery to the mosfets the switching edge would be very slow and burdening the mosfet. 

Without buffer capacitors the battery only delivers current during the duty cycle on phase - so it has to deliver pulsed current, for which li ion cells are not (officially) specified. And, for example 50A average motor current delivered with an duty cycle of 50% would mean 100A pulses from the battery! As dissipated power is proportional to the square of the current that leads to extrem burdens at (internal/contact) resistences.

The buffer caps on the mainboard straighten this out, presumably just a tiny, little bit and could easily be overburdened by this. The 560uF kxj capacitors used in the RS are specified for ~2A ripple current...

So some capacitor banks inbetween nicely reduce the buffer caps burden on the motherboard and the pulse burden for the battery.

And as "side effect" increase a bit the driving characteristics. But quite sure this does not safe one from a cut out.

[Tawpie]

Back in the day when I did this stuff for a living, we had a magic component we dubbed a 'resductor' that we used for high power current limiting. You use a series big giant (yes, physically big and giant) wirewound resistor... during the inrush phase, the wound wires made it act like an inductor but as the current became more steady state the resistive characteristic took over to continue to limit the current draw. The thing would make an audible SNAP when you applied power because the wound wire tried to contract inside its composite potting (magnetic field). It was all very exciting to power up a new design for the first time.

We used these to discharge our caps as well, it's always better to deenergize in a controlled fashion than to just zot things to ground.

Edited December 14, 2021 by Tawpie

[RagingGrandpa]

FYI, Chriull did an interesting simulation that suggest that having more capacitance near the board may slightly improve efficiency (e.g.: range):

https://forum.electricunicycle.org/topic/18815-capacitors-on-motherboardsany-tech-gurus/?do=findComment&comment=397921

Quote

Edited March 1, 2022 by RagingGrandpa

[RagingGrandpa]

9 hours ago, GoGeorgeGo said:

I think the best you can hope for is super capacitors with the ability to maintain a rapid tiltback for 200 feet with a 200+ pound rider. That way in the event of a [overlean], you just get quick but gradual tiltback to a stop. Wont prevent all accidents and could create different issues like stopping unexpectedly  the middle of busy traffic. But in those situations you where going to just faceplant anyways

An interesting suggestion... 

I think in order to produce more torque at a high speed (to remediate overlean), the capacitors would need to provide a source of voltage higher than the pack voltage. And ok, if we're suggesting a brand-new type of controller designed specifically for it, that's practical.

Just curious: How did you envision the "reserve capacitor" being connected to the circuit? 

[GoGeorgeGo]

1 hour ago, RagingGrandpa said:

An interesting suggestion... 

I think in order to produce more torque at a high speed (to remediate overlean), the capacitors would need to provide a source of voltage higher than the pack voltage. And ok, if we're suggesting a brand-new type of controller designed specifically for it, that's practical.

Just curious: How did you envision the "reserve capacitor" being connected to the circuit? 

Not an electrical engineer, i have no clue how it would work or if its even possible. Just an idea i had that sounds plausible and could at least make cutouts slightly less scary. Instead of 100% chance of a catastrophic crash maybe its down to a 25-40% chance with the majority of incidents just ending in riders soiling their pants 🤣

[RagingGrandpa]

(I'll avoid burying the capacitor topic deep in the EX20S megathread...)  

1 hour ago, Chriull said:

Energy stored in Ws in a capacitor is calculated as 1/2*C*U². With common maximum Voltage of supercapacitors of about 3V on would need a ~1600F supercondensator bank...

So an reserve li ion battery bank for emergency would be more efficient.

I was thinking the boost voltage source would be in series with the existing pack... so that a sagging-voltage pack would still be providing the bulk of the energy, and the reserve (when activated) just adds a little extra to raise the voltage. But it requires a new control board design, so it's a from-scratch dream, not an add-on.

The 'switch' (e.g.: FET) would need to be triggered smartly, like when the ESC nears 100% duty, and must also switch back to the pack if the voltage becomes too high for the controller or reserve to tolerate (e.g.: during braking). All sounds practical... 

When the switch connects the reserve it will be tortured by inrush (instantly charging filter caps in the ESC), which may cause a switch durability issue. An inductive 'choke' in series might help? But we don't want to slow it down very much, since the whole point of the reserve is to be fast-acting.

And how to make the reserve- batteries, or capacitors?
Batteries are tons more efficient in terms of energy-per-size... but not voltage-per-size.
We're talking about a peak-power high-current situation, expecting as high as 100A pack current... short pulses at 100A might be reasonable with high-drain cells (e.g. LFP, P42A, etc). But it's at least 6 extra cells, to get to the ~25V extra that we're interested in. 
Capacitors could perhaps be 25V in a single can... but deplete 10x more quickly. A tradeoff.

Edited March 2, 2022 by RagingGrandpa

[Tawpie]

I was going to bring that up in the EX20S thread too, but then it dawned on me... how does this really improve things? This is a finite power source that will run out, it might be good for a bump but if you keep the pedal to the metal or hit one too many bumps before you can recharge your reserve, you'll still go down. There are already mechanisms (except in the MPro and evidently the current S20s) to keep you in an operating range that has room for balancing power and we either disable or ignore them. If you don't ride the beeps, you don't really need super caps/topping reserve. But if you must ride the beeps, this is just another 'beep' to ignore isn't it?

[Tawpie]

BTW, by putting your supercaps in series with your battery, the V term in 1/2CV^2 gets to a much more manageable level and the C necessary to maintain power for a few seconds drops into a more practical range. I think your emergency switch almost has to be a BJT, you'll want to carefully dial the power in and out so the motor doesn't jerk too much. Entirely doable BTW, but not simple.

[RagingGrandpa]

4 hours ago, Tawpie said:

this is just another 'beep' to ignore isn't it?

Hehe right, if we increase the limits, people will always find them... 

For us non-daredevils, I was envisioning major pedal tiltback being commanded whenever the reserve gets triggered. Something powerful and noticeable. (Today, EUC's can't tiltback to respond to overlean, since there was not enough torque to start with.)

This 'reserve' concept begs for a clear problem statement... I think it's like:

" With today's EUC's, with a full battery, riding at high speeds 
(speeds when maximum possible motor current is limited by motor EMF, and not the controller's self-protection limits),
if a rider leans forward and the controller reaches 100% duty,
motor current (torque) cannot be increased,
because the voltage applied to the motor by the controller reaches the maximum voltage available,
and therefore self-balancing corrections are no longer possible,
and therefore the rider crashes."

To improve this overlean crash problem, there are many options available for consideration:

1. Decrease EMF.
   a. ... by changing the motor construction (It's why we install a C30 motor instead of C38, when maximum speed is desired), or
   b. ... by using field-weakening circuits (changes to the controller; higher heat dissipation and energy-inefficient to sustain; causes additional voltage sag when it activates).
    
2. Increase the input voltage. (It's why 100V EUC's are faster than 84V EUC's using identical motors.)
   a. ... by installing a higher-voltage pack, or
   b. ... by reducing voltage sag of the pack during load 
        1. ... by installing more cells in parallel (a much larger pack), or
        2. ... by using lower-resistance (high-drain) cells, or
   c. ... by installing a reserve energy supply (extra batteries or capacitors, as discussed in the posts above).
    
3. Increase the output voltage. 
   ... by using a boost converter at the input of the controller (large components added to the controller; higher heat dissipation and energy-inefficient to sustain).
    
4. (and perhaps additional ways you can suggest?)

The bold ones are techniques the EUC manufacturers have already used... the rest come with new challenges. 
And 2.b.2. is just now becoming an option, with eWheels, AlienRides, etc offering high-drain cell options in new EUC's. Those should be safer to ride fast (when fully charged, at least) than the standard cells.

These are the same answers to "how to increase the maximum speed of an EUC"... and are missing emphasis on avoiding the crash.
For avoiding crashes, we also need to answer "how best to alert the rider that an overlean crash is imminent?" 
Alarms that activate at the moment that control is lost, don't help. And alarms that attempt to predict a future event (overlean) are often perceived as a nuisance.
Tiltback at the moment of overlean is only possible by using 1.b., 2.c., and 3. (the methods nobody uses yet).

So, room for improvement  

Edited March 2, 2022 by RagingGrandpa

[redsnapper]

 @RagingGrandpa for President! can I get a HUZZAH?

[InfiniteWheelie]

I've thought about this backup supercapacitor idea before, and I think it's probably possible. Look at Skeleton supercapacitors. Their 3200 farad cell is 3.6 wh and only 60 x 138 mm. Just one may be enough, but you could probably fit up to four inside an upper motherboard compartment.

This would be incredible for a few reasons. Firstly you could actually find/practice your wheel's acceleration and top speed limits without overpowering the wheel and crashing. This makes it both an excellent training tool and safety feature. You'd always be able to accelerate as hard as you want without fear. You could also cruise near top speed without the worry of a bump/pothole/hill causing an overpower situation. Plus in the case of a total battery failure you'd have enough backup energy to come to a stop.

There'd be an extremely loud beeper that sounds any time the supercapacitor is active. For momentary overpowers (leaning too hard, going too fast, hitting a bump), you simply back off causing the beep to quickly stop. If the beep doesn't stop you know the battery failed and you must pull over right away.

After typing this out and thinking about it more, it seems like an amazing idea. The ability to safely learn you wheel's limit while never fearing overpowers or battery failures would take these wheels to the next level.

 

Edited March 3, 2022 by InfiniteWheelie

[InfiniteWheelie]

My above post got moved here as it was off topic, but I'm happy to discuss it here. Something I missed is that you could have two sets of supercapacitors, one for overpowers and one for braking. The braking capacitor would be kept empty and immediately fed back into the wheel after braking. In theory this means having a cutout from braking should be impossible, no matter how hard you attempt to stop. 

In terms of overpower situations there seems to be two ways to handle it. One suggestion in this thread is simply using it as backup energy to provide tilt-back once the battery is overpowered. Maybe it's because I'm not the best rider but it seems to me tilt-back is either weak (meaning you may not notice it), or when made stronger can be dangerous especially at high speed.

Another idea I like better is simply having the supercapacitor produce significantly higher voltage, allowing for higher acceleration and top speed at 100% PWM.  Perhaps something like 25% higher to allow the rider plenty of time to hear the alarm and react. Unlike other beepers it should be extremely loud, perhaps changing from a steady tone to pulsed when the supercapacitors drains to 50%, giving an idea of when it'll cut out. Implementing this is trivial because the supercapacitor already needs a voltage converter by nature of having a linear discharge curve.

 

 

Edited March 3, 2022 by InfiniteWheelie

[Tawpie]

14 minutes ago, InfiniteWheelie said:

you could have two sets of supercapacitors, one for overpowers and one for braking

Me, being a fan of KISS, would use a dumb resistor to deal with braking. When the battery won't absorb the braking energy, dump it into a big power resistor. Even that adds complexity, but not a whole lot. It seems straightforward enough to make me think that it was deemed unnecessary for 'typical' operation—now that we like to operate our wheels beyond their designed limits though, such a decision could/should be revisited. Or not. You don't want to lose traction because the brakes were able to lock up.

Edited March 3, 2022 by Tawpie

[InfiniteWheelie]

26 minutes ago, Tawpie said:

Me, being a fan of KISS, would use a dumb resistor to deal with braking. When the battery won't absorb the braking energy, dump it into a big power resistor. Even that adds complexity, but not a whole lot. It seems straightforward enough to make me think that it was deemed unnecessary for 'typical' operation—now that we like to operate our wheels beyond their designed limits though, such a decision could/should be revisited. Or not. You don't want to lose traction because the brakes were able to lock up.

I have no electronics expertise, so let me ask a few question about this method.

So a resistor in this case just turns motion into heat, just like a brake disc correct? Can resistor braking be precisely controlled, because if not it's useless in this application since we need to balance. In what scenario could the wheel entirely lock up?

Edited March 3, 2022 by InfiniteWheelie

[Tawpie]

Correct. A resistor simply converts energy (in our case, the energy that results from converting motion into electricity) into heat, exactly like a disc brake. When braking, our wheels are actually generators.

Resistive breaking can be easily controlled, it can be brute force max on or totally off, or precise and applied in small and changeable increments.

The regen braking system is limited because the control board only allows a certain amount of regen, after that amount is exceeded you can't add additional braking. I'd want to do something similar with resistive braking. As a starting point I'd investigate tying the amount of power shunted to the resistor to the motor spin rate and ease off resistive braking as the wheel slows down (without a computer in the loop)... at some point regen braking can take over and bring you to a skidless stop.

Most of the time you'd just use regen because it's adequate and does return a small amount of energy for later use. The resistive brakes would only apply in cases where regen maximums are being exceeded. At least in my first sketches! The whole system would need careful thought, there are always unintended consequences.

Edited March 3, 2022 by Tawpie

[RagingGrandpa]

Do we really need more braking force?

I think overpowering the EUC during braking is a very very rare type of crash...

[InfiniteWheelie]

If precise resistive braking can be achieved as you explained it may be a better solution for absolute braking power, assuming you can dissipate the heat which probably wouldn't be an issue. It would certainly be simpler than having two separate supercapacitor arrangements. The amount of times/duration when exceeding battery braking performance would be very small, so only a tiny amount of regen would ever be lost.

On 3/3/2022 at 3:48 PM, RagingGrandpa said:

Do we really need more braking force?

I think overpowering the EUC during braking is a very very rare type of crash...

Yes I think so. It is certainly possible to overpower basically any wheel on emergency braking if you're a good rider. That doesn't stop us from riding and it's rarely an issue, but having unlimited braking power would be a huge plus for safety and piece of mind when riding.

 

Edited March 6, 2022 by InfiniteWheelie

[Tawpie]

56 minutes ago, RagingGrandpa said:

I think overpowering the EUC during braking is a very very rare type of crash...

I agree, it should be very very rare. The handles added by Leaper Kim and now KS give you a way to do this pretty easily though, so I can see it becoming less rare.

[kolmog]

On 3/3/2022 at 7:21 PM, Tawpie said:

Me, being a fan of KISS, would use a dumb resistor to deal with braking. When the battery won't absorb the braking energy, dump it into a big power resistor. Even that adds complexity, but not a whole lot. It seems straightforward enough to make me think that it was deemed unnecessary for 'typical' operation—now that we like to operate our wheels beyond their designed limits though, such a decision could/should be revisited. Or not. You don't want to lose traction because the brakes were able to lock up.

This would be awesome.... And, to add a slightly tongue in cheek suggestion, combine that with amplifying the hall sensor input (at least during the braking side) and you could get some actually half decent braking! Why don't we have an open source controller to play with again?

[Number1stunner]

Ok, so has anyone actually tried this on the Sherman?.....

 

I personally have converted my begode Rs (black board version) to 1845wh 18650cells and I included the super capacitor.

 

For me I love the capacitor instant difference the pedal is solid I wouldn't revert the rs back..... 

 

However, I haven't tried this on my Sherman as there is limited space and I just don't know if it would actually make a difference....

 

Who is willing to be the guinea pig?