Motorcycle startup help with LiPo

[brainwash]

I'm on my 3rd battery right now and getting annoyed by this topic. I have a large-ish 400c scooter that is a beast to start up and eats batteries. I've connected an XT60 plug and use a 2200mAh 3S LiPo in parallel to the main battery to help start it.
Before you ask: connections are fine, wires are fine, charging is fine, valves are set up correctly. Starter motor might be old, but whatever.
I'm using the 2-wheeler seldom so the other solution is to keep the battery in-house and keep it topped up, but it's a pain to reconnect each time.

So my procedure is: prime the fuel pump, rotate the engine a few times, connect the LiPo, start the engine, keep the LiPo connected for a few seconds, disconnect it. A few seconds is enough to keep it fully charged. I wonder if there is a way to automate the process, so here are my questions:

- 3S or 4S? My charging circuit pumps ~14V but resting voltage is ~12.4V. 3S would have a range of 9-12.4V, 4S would have 12-16.8V.
- mosfets or relay to connect/disconnect the battery? I don't know what parts can take >70A.
- what a reliable and simple way to detect that the motor is running or the battery is fully charged and disconnect it? I don't mind sticking an IC there with a voltage divider, but perhaps a TL431 is more reliable.

My idea is to stick everything on the contact voltage input, not on the starter input. This is because the main battery can still be recharged (recover) between startup attempts.

So: read contact voltage, if above 13V (engine running) disconnect LiPo, if between 9-13V then connect LiPo. However this is 0.6V above the 12.4V limit of a 3S. If I use 12.4V as a threshold then the resting voltage of the battery might be as high as 12.7V (which is fine for starting) but might drop during cranking. So the circuit would then connect the LiPo, voltage would go back up, it disconnects the lipo, voltages comes down again -> bouncing.

On the parts side, I have quite a few SMD automotive high-side smart MOSFETs (relay replacements) in my parts bin:  BTS6163D, BTS5012SDA, BTS6143, BTS4443, VN750, BTS247Z, ... I also have a couple of those with current sensing (AUIPS7145R, VND5E025MK-E, VN7140, ...).

Perhaps I'm overthinking this.

[KL27x]

So the circuit would then connect the LiPo, voltage would go back up,

This is an incorrect assumption. Starter motor is still, presumably, cranking. And connecting 3S LiPo of that capacity is not going to raise voltage back to over 12.4... Not even close. Heck it tops out at this 12.4 at rest. It'll go down to around 9 under heavy load - which, presumably, you have. 12.4V won't be reached on the power rail until after the stator and voltage regulator are up and running AND after you let go the button to stop cranking the starter because the engine has started.

Sound like your main battery does not provide enough juice to start the scooter. The voltage must be sagging very low. If it could sustain even 9V, it would be fine. The starter would be cranking just fine. The voltage must be much lower than that while cranking, without the auxiliary pack. And probably much closer to 9 than 12.4V when you add the pack into the circuit. I'd bet much closer to 9V than 10V, even. Maybe closer to 8. At any rate, if this were ever a problem you come across in the future, you might be interested to look up "hysteresis." And read datasheet of a common comparator to find app notes on how to do this with a comparator.

3S is better matched than 4S. Lead acid battery will droop to around 10 (if the load wasn't too much for it, anyway). And 3S will also have sustained load voltage of close to that (both will probably drop to 7 - 9ish under short, unsustainable higher loads ilke this). If you used 4S, the 4S would get tapped as the vastly greater provider of current, at about 11-12V, while the lead acid does nothing... But unless your LiPo can do that alone, what actually happens is output of your lipo would drop like rock down to 9V to where lead acid is also doing its part, but most of the voltage drop between 12 and 9V  would be dissipated inside the LiPo battery itself. Wasted energy and extra heat, on top of that.

Some small comment, here, on the safety of your setup. Whether or not you're exceeding max current draw and/or consequences of such are on you. Let alone your charging method. Should really use a separate regulator, and maybe power resistor if not an active 3S Li ion charging IC, and a separate switch to cut in to charge, to ensure that battery is not charged too fast or too high. Fact that the stator/regulator can sustain 14V while a semi-drained lead acid battery is connected to it should tell you it's capable of a lot of output to the LiPo. Also, in this kind of high drain use, your LiPo may very well need frequent balancing.

[calexanian]

I see you are in Europe, but I would go with a different solution and get a LiPo starting battery. I have met the guys from this company https://shop.antigravitybatteries.com/productline/starter-batteries/small-case/ and their products appear to be quite good. Ignore the silly name. They have internal charge controllers so they may be used in place of lead acid batteries with no modification and they have sizes suitable for bikes. At first I thought they were just Chinese junk but on closer inspection they are quite good and very light. They have very very large cranking capacities, but relativley low reserve capacity. One down side of LiPo starting batteries. Hope this helps.

[brainwash]

I've seen the LiPo batteries and was thinking of fitting one but it would almost double the scooter's value  This was just a fun project that I was thinking of, I anyway use my bicycles more than the motorbike.

I omitted a 'small' requirement from the project, doing it with the ignition signal is easy but was thinking of not using that, just a direct 12V connection to the battery. Otherwise I would have to drill some holes. That extra requirement makes things harder. I could monitor the voltage and act accordingly:
- 10-14V - battery is ok, motor might be running, disconnect the circuit
- 5-10V - engine is cranking (or battery is rapidly self-discharging), connect the circuit (?)
- 0-5V - battery is disconnected, do not connect the circuit

Again, this circuit would be just for fun, it's not a huge deal to: use the key to open the battery cover, connect the XT60 plug, take the key out of the panel and use it to start the engine, disconnect the LiPo.
Just brainstorming, it's winter anyway here.

[calexanian]

Does this scooter have a clutch or is it an automatic? We just used to push start manual ones. Who needs a battery? Hahahaha if it's an automatic one then it's time to evaluate your life choices. Perhaps a Honda or some such thing with a real transmission should be in your future, if you intend to continue riding.

[rstofer]

One of the best uses for an old fashioned V-O-M (analog voltmeter) is to connect it at the battery and watch how it sags under cranking.  It is also useful to measure at the starter or even across the cable connections (stud to stud, not wire terminal to wire terminal).

A long time ago, the starting system presumably worked.

Lead acid batteries work well for a while.  Then they build up a sulfate coating and their ampere capability goes away.  A V-O-M will show this far better than a DMM.

[NiHaoMike]

I remember one of the multimeter reviewers here does a frequency response test with a signal generator and a bargraph display is usually faster than a moving pointer.

As for the solution to the problem, simplest I can think of is a slow blow fuse, relay, and a large diode. Have the relay only close on the start position and then use a linear regulator set for 12V or so as a charger, plus some sort of balancing circuit.

An alternative solution is to add some supercaps.

[brainwash]

I've used a DSO to characterize the voltage drop, it's going near 5-6V at cold cranking. I can assume it's because of the starter motor but don't want to play the game of 'swap and find'. It's not loss through the wires or connections or anything else. It just needs a boost, complemented nicely by a LiPo. The idea was to think of a plug-and-play solution that can react on the main 12V input and decide whether to connect the LiPo or not.

Again, the simple fix is to get a new battery, keep it maintained and it will still last 1yr, maybe 2 (if lucky). I always take the batteries in the house during the winter or keep them above 12V when below 0C outside. I'm looking for a decent solution that provides the boost when needed and perhaps allows the LiPo to be charged a bit by the alternator. But I don't mind taking a 300g battery inside the house to charge it, it's just a bit fiddly.

The other solution is to install another 12V battery in parallel with the existing one, a route some people have already taken, but I would be giving up storage space.

[brainwash]

Reviving this topic as I might get some suggestions.
I've attached an image with what I'm working now, basically I want to parallel a few AUIPS1201 N-MOS switsches, as I have plenty of those and they are cheap 45A MOSFETs with inductive protection.
I basically want the MOSFET to kick in during cranking cycles, so the yellow line. The dotted line is what I expect to happen, but don't wish it would. Basically, during normal deep discharge, the LiPo will take some of the drain, but I want a low-peak detector.
So let's say, the voltage drops 1-3V within a 0.1s, the circuit should kick in. But it shouldn't during slow discharge.

I want to avoid a microcontroller, even though it would be much easier with one. I'm thinking either TL431, LM324 or MC3423, +555 , whichever results in a lower part count. At this point I might be better off trying to optimize power-saving modes in an uC.