On-off switch requirements
- Thread starter Garrick_s
- Start date
lifepo4ever
10 kW
There's a little less than 1 mF worth of caps per phase on the battery lines, so you can use a 10K 1W resistor on each line for up to about (probably further than) 100V of batteries, and it'll take about 15 sec after connection w/o any switch to get pretty much up to charge, which for me would be less time than I'd take getting around the bike and on my saddle. That isn't a heck of a lot of current, so it's a fairly small resistor (wouldn't be this big honking thing weighing the batt lead down), and costs about 25 cents apiece (0.16 BP, in packs of 10), but if you get 25K from DigiKey, that'll bring the price down to about 5 cents each (Group Buy, anyone? :lol: ).
Correct me if I'm wrong.
Cameron[/quote]
i have regular resistor here , there no ceramic around it can i use it ?
Correct me if I'm wrong.
Cameron[/quote]
i have regular resistor here , there no ceramic around it can i use it ?
oldpiper
1 kW
auraslip said:
'Cause that's where I plug my batteries in and I get a good spark. That point of contact is obviously not what he was talking about.
Cameron
John in CR
100 TW
The spark will mess up your connectors, or switch, or circuit breaker, or whatever it is making that contact. Contactors are about the only things designed to withstand the abuse. The rapid inrush of current also wears on the caps, reducing their life, and as I understand from DrBass's result can also damage other stuff on the controller's circuit board. Maybe on 36v or 48v rigs it's less of an issue, but the darn spark at 80v always make me jump a bit even though I know it's coming, so I precharge the caps now whenever I need to connect a pack, and if I did it frequently I'd put the resistor in a circuit and do it the right way.
SoMoney
10 mW
- Joined
- Dec 6, 2010
- Messages
- 30
Trying to find an actual picture without any luck so I wiped this up in Google ScetchUP... How about this idea for a Spark Killer switch? The BLUE connector would be one of those white block like 20W Ceramic resistors. Since the positive pole will ALWAYS be HOT thanks to the 20W resistor, would this stop the spark too?
EDIT: Now that I think about it. I doubt it will help much and probably just move the spark over to the negative poll. In order for this to work the negative poll must touch 1ST a split second before the positive connects. THEN it would diminish/remove the spark. Easy enough to ROCK a deans connector in NEG then POS, but I'm not aware of any switches that do that.
Attachments
lifepo4ever
10 kW
nice!
Alan B
100 GW
Here's another variant:
A wavecrest Tidalforce battery uses two 40amp fuses and four mosfets for on/off functions. these http://www.datasheetcatalog.org/datasheet/vishay/71704.pdf .
Two for the positive side and two for the negative side, all having their source leads towards the cells. the mosfet driver then makes them conduct. have used this method
for a long time with no problems, a small tiny switch turns on the controller, no sparks no issues at all.
Four ideal diodes can be modified to do this.
Two for the positive side and two for the negative side, all having their source leads towards the cells. the mosfet driver then makes them conduct. have used this method
for a long time with no problems, a small tiny switch turns on the controller, no sparks no issues at all.
Four ideal diodes can be modified to do this.
Alan B
100 GW
Do they turn the FETs on somewhat slowly, or hard? How are the gates driven??
Alan B
100 GW
If they are using a driver that would likely indicate they are driving it pretty hard. Which would still make a large current surge, just no spark since the connectors are already fully mated.
Interesting.
I would think that turning the FET on a little slower would be okay too. Not sure why they need FETs in both directions.
Any other FET based on/off switches out there?
I made a simple sketch of one, maybe will be able to post it tonite. Might have to do some simulation or testing to see if a simple design will hold up. The more complex design I did had a separate FET and resistor for precharge, but that may not be necessary.
In essence most BMS's have this type of switch for discharge onboard. It is a fairly straightforward FET design. The charging of the caps makes it a little harder to predict the current and heating. After that, the controller drain loss and heating is easy to predict.
Interesting.
I would think that turning the FET on a little slower would be okay too. Not sure why they need FETs in both directions.
Any other FET based on/off switches out there?
I made a simple sketch of one, maybe will be able to post it tonite. Might have to do some simulation or testing to see if a simple design will hold up. The more complex design I did had a separate FET and resistor for precharge, but that may not be necessary.
In essence most BMS's have this type of switch for discharge onboard. It is a fairly straightforward FET design. The charging of the caps makes it a little harder to predict the current and heating. After that, the controller drain loss and heating is easy to predict.
Alan B
100 GW
Here is a FET design with pre-charging. Not tested.
The keyswitch turns on Q1 which charges via the precharge resistor.
Q2 keeps Q3 and Q4 off until the precharge is within a few volts of full battery voltage.
Q3 and Q4 then come on and provide a low impedance path for full current load.
When off the leakage is approximately 200 uA.
Using an FET switch without the precharge reduces the parts count even further, but makes it hard to predict the charging currents which might be quite high. The above design limits charging currents to 10 amps at 100 volts.
The keyswitch turns on Q1 which charges via the precharge resistor.
Q2 keeps Q3 and Q4 off until the precharge is within a few volts of full battery voltage.
Q3 and Q4 then come on and provide a low impedance path for full current load.
When off the leakage is approximately 200 uA.
Using an FET switch without the precharge reduces the parts count even further, but makes it hard to predict the charging currents which might be quite high. The above design limits charging currents to 10 amps at 100 volts.
Alan B
100 GW
I made a simpler design also. But it doesn't really limit precharge current very much. Might work, but the current surge will be high which is hard on the FETs and hard on the capacitors.
Be interesting to see what you find.
Be interesting to see what you find.
Erogo
100 W
I have a question about the resistor that you charge the caps through to avoid The Spark. Peak currents through this resistor are a lot lower than the spark current would otherwise be. It's a short pulse of this high current though. Not the instantaneous flash of a spark, but a lot shorter than "continuous rating" would apply to. So,... does the resistor actually need to be real grunty? How long would it take to fry a resistor if you were running above its rating? Has anyone ever blown a charging resistor? It wouldnt be a major deal anway hey, they're cents not dollars.
Then also, a second question:
Rather than FETs, could you use relays, like in this circuit?
EDIT: pologies, just saw the many other threads on this topic...
Im no engineer, so just tried to figure this out in a circuit simulator http://www.falstad.com/circuit/ which was good fun.
I plan to have all the high current down the back in the trailer, and run signal/control current only at the handlebar. There's two relays at the moment, mostly cause I was excited about using relays. Probably the first one you can do without?
The first switches voltage to the ESC caps through a resistor, operated by a non-latching switch.
The second switches voltage straight to the ESC without a resistor. I'm going to control that with a non-latching switch too, as my drive is only a momentary-on affair, just for our steep driveway. I think a more normal implementation would latch this switch.
I think you could put a failsafe in there somehow, so that only once the capacitors were charged up by the slow charging circuit could the main relay switch.
Chip chip,
Eric
Then also, a second question:
Rather than FETs, could you use relays, like in this circuit?
EDIT: pologies, just saw the many other threads on this topic...
Im no engineer, so just tried to figure this out in a circuit simulator http://www.falstad.com/circuit/ which was good fun.
I plan to have all the high current down the back in the trailer, and run signal/control current only at the handlebar. There's two relays at the moment, mostly cause I was excited about using relays. Probably the first one you can do without?
The first switches voltage to the ESC caps through a resistor, operated by a non-latching switch.
The second switches voltage straight to the ESC without a resistor. I'm going to control that with a non-latching switch too, as my drive is only a momentary-on affair, just for our steep driveway. I think a more normal implementation would latch this switch.
I think you could put a failsafe in there somehow, so that only once the capacitors were charged up by the slow charging circuit could the main relay switch.
Chip chip,
Eric
Similar threads
