Sparks when connecting battery

[Leslie]

Thanks mate, I'll pick up one of those for sure
And you're right, you would need about 5K for a 5v @10Ma sink. Do you know if your LM317 draws any current?

I have a circuit in mind, driven by a PIC that I can update as necessary. What is will do is basically switch output pins on a delay, and input pins will check for voltages present etc. ensuring there are no mishaps.

Not trying to go crazy, but the setup will include 2 x 6S1P25C/5Ah (long LiPo packs I got the idea from Alans mention of hobbyking in another post).

Sniffing around I found a reverse polarity protect circuit and a precharge circuit using a relay to switch. I basically just joined the 2 together after a delay switched relay circuit I came up with (that could be driven easily by NE555 and 2 push buttons but I'll use a PIC with GUI)

Although there looks to be alot of relays here, they can be changed to a handful of fets but I chose this because I have some decent 30A relays that I feel won't have a problem with the current passing through them at full power. Relay rating is for a switching current. As the bike will be stationary, or zero throttle/load I'm tipping this will be okay. I won't be switching them with any substantial current passing through them.

There is already a latching relay built into the GM battery, so the code will just check there is not voltage present on the GM battery output before attempting to switch on the backup packs. This will eliminate any worry of both the packs switching on at once, and stop any problem should R6 fail for some reason, and both battery sets arch and connect.

Anyways, what do you think of this? Still needs a few things added...

far left -> series/parallel switched relay circuit (relay coil connections removed to read easier)
middle -> reverse polarity protect
far right -> precharge auto switch on

Why?
-Switches between backup packs and gives option to series/parallel connect matching packs
-Can use 3 completely different packs altogether, and switch between them individually
-Parallel mode for matching packs for greater range
-Can use battery monitors to auto switch off depleted pack and switch to fresh pack without having to touch anything
-No sparks lol my Anderson connectors are covered in ox

Estimated current draw from added components (relays switch with around 10mA)
-> GM battery = 0
-> Backup battery series/single mode = < 50mA
-> Backup battery parallel mode = <100mA

Parts I need:
3 fets I have all the rest gathering dust
The battery from hobbyking (around $100 or so for both)

Compared to buying a frog pack for backup, I put my hand scales in the air and $270 for a 10Ah 24V I can guarantee will have a far less discharge rating than 25C. My thought is to use the backup pack for takeoffs/hills and switch to the GM pack everywhere else. I assume I can near double my range

Not tested at all, and I'm sure I'll find some discrepancies in my circuit.... It's a fun experiment!

MM

No the LM317 will only drop in volts over the output and requires about 2 or 3 volts (some watts) to keep it hard regulating the current.

But the relay is good like this.  It can take 32v on the control lugs so if you wanted to use a resistor you don't need make it too efficient.  As its pretty efficient anyway compared to non latching relays.

And that latching relay you put me onto is interesting.

I hook it up to an amp meter and it draws like 90ma or what ever on the sound of the switch the then drops to 14ma if you were to hold it there.  It like they go into low power mode.

They wont burn out like one may think they do if you don't have a pulse and use straight DC..
The LM317 chip is set to deliver no more of no less than 10ma regardless of the voltage but has a 37v limit to the v drop.

I use the resistor to drop to 22v @ 10ma and from there the LM317 will always give a healthy 10ma to the relay control lugs.

If the volts are high or low, still 10 ma.  This can save power as using just a resistor one may need to consider any rail drop that may occure during discharge and have to use more current to the control lugs so slumps do not risk shutting down the relay, like LVC.  You could set an LVC on these if you really wanted to.



I had this down working to 5ma but the SSR indicator led wasnt shinning too brightly.  I just upped the current until the SSR on led looked an acceptable brightness.  10ma has done the trick.

[Leslie]

You always pass lots of current just when that spark hits.  Caps are terible at hitting on the reed in relay switches.

I think inrush to caps can be very high to the cap postshunt and before the shunt worse.

Usually engineers will use a low ohm wire wound resistor or thermister to limit in rush from damaging to caps like this or a similar ramp up to precharge.   In rush Limitation on the main rails is not practical in ebike applications.


The SSR is a good alternative and precharge is better and easier to implement.  It cant fail.


Caps are rated at how many high energy shots they can take when discharged and also how many shots they can take being charged like this, With the big crack spark really gives them a shock that is no better than shorting them outright.

IMO!

Cap failure can occure anytime over a year or two and your controller should follow. With the precharge, all the big capacitors are going to be treated much more gentle.  

Without precharge the better the current pathway (switch amp rating) the more shock the caps take.

I also think that without precharge SSR fets wouldnt take kind to the energy zap that the controllers take up when connected either.

My longest living controllers had precharge.  I think it made a difference over the long term.

[MonkeyMagic]

Yup you're right but it's just the caps that suck the juice, it's inevitable unless you bring the voltage and trickle the current through first.

I should record a spark on my car when I connect the car stereo power cable linked to 3Farad of caps. That makes a BANG noise and lights up my boot lol

I'm thinking of linking them in series for the voltage and putting them on a seat post rack or something for quick boost lol they won't work well though...

Okay so LM317 can do some kinky stuff,

Les --- what do you think of my kinky circuit? Did you follow it? I was after some feedback man! lol

[Leslie]

Hmmm the series caps is a bit of a stretch. as they will block any DC when discharged and behave as a flat battery would.  But they would drain till theyre flat but realy fast ..

But the backup pack switch is cool enough.  I still feel if you just shunt each pack with the correct resistance there should be no need to have the switches and you can use both packs at the same time.


Another option one can look at is using brushed controllers to controll you batteries outputs current.  They are excellent for this as they limit the PWM duty cycle which is current based.

Set LVC individually or wow use a CA with the throttle hook up to do the hard work for you, the concept of a CA's connected to the brushed controllers throttle logic, make ideas start to flourish in my head.
 
If you could get a PIC to utilise the throttle interfaces of 2 brushed controllers you certainly could prolly figure a way to use both pack at once and reap the rewards..

But I would go for a identical pack voltages between them so they can be shared..

A lot of work in your plans and I do hope it works for you.  But I think you could come up with something better again.