EMF destroyed voltmeters!

[Electricdirtbiker]

Hi all.

I'm having a problem where the spark from switching the battery power on to my speed controller (alltrax), destroys the voltmeter that monitors the battery pack. The digital volt meter's back light comes on but the screen is blank. I believe the spark causes a big energy field that somehow overloads the ic's in the volt meter. 2 dead meters now.
I've tried using a bulb to soft start the controller, this works so far, but I'm worried that any spark anywhere else = nuked meters.
Any ideas would be nice.
Thanks,
Wes

 

[swbluto]

I think I would add capacitors parallel to the battery. Also, possibly using a voltage divider to the voltmeter may help. Also, starting from the highest voltage rated setting on the volt-meter when switching the battery might help. There's probably voltage spikes on the measuring points due to the stray inductance and possibly the main inductance, so any method of reducing that voltage spike will probably help.

I really doubt the EMF is strong enough to induce a chip-destroying voltage surge.

 

[Electricdirtbiker]

Thanks Swbluto,
Those are all good ideas. I'll start with a cap, but where is the best place to put it? Close to the battery or at the meter? What would be a starting value? Not knowing the polarity of the inductive pulse, ceramic over electrolitic cap? The meter reads through a double pole selector switch, that selects each battery in a string, so there is only one meter range is needed.
Wes

 

[swbluto]

I would opt for an electrolytic since it's cheaper to get a larger capacitance, and I think you're probably looking for a "large capacitance" like higher than 100 uF. The higher the capacitance, the better for the voltmeter, but it may cause the switch to erode faster. Polarity shouldn't be of a concern, assuming you're hooking the voltmeter up correctly, as it's very likely you're seeing a voltage spike which usually skyrockets upwards. One way to prevent this erosion is to put a resistor in series with one of the voltmeters leads (I prefer the positive lead), and then place the capacitor across the voltmeter so the resistor is "in front" of it (Any current going to the capacitor has to pass through the resistor first) - I'd choose a low resistance power resistor, like a 5W+ 10-100 ohm resistor. The capacitor will smooth the voltage that the voltmeter reads, so it won't be useful for reading transient voltages and it may slow down the voltage reading of other voltage signals, but voltmeters usually aren't useful for transient voltages anyways.

If it's just being used to see what the output voltage of a battery is, like for LVC purposes, it should be ok.

 

[TylerDurden]

I rekon the bulb is the right direction. It's acting as a precharge resistor.

Caps are likely the root of the inrush problem.

You can add a small circuit:

 

[Icewrench]

Cap close to battery connection, ev global had them inside battery box. A 24 volt bike used a 4700 mf 35 volt cap.

 

[Electricdirtbiker]

The cap and resistor in series sounds like the solution, will try it this week.
Do you know of a non micro processor circuit that will automatically pick and display the lowest voltage cell in a battery string?
Thanks,

Wes

 

[amberwolf]

I can imagine one using comparators and opamps; even one using transistors only, but that's a lot of wiring, or a large PCB. It'd be a lot easier to use an MCU.

Well, I guess you could actually use a pair of analog multiplexer chip with enough inputs to cover all the batteries (one chip for the +s and one chip for the -s, unless you can get one with pairs of inputs simultaneously activated), along with something that counts up and increments thru the inputs. Then use a comparator on the output of those chips to look for a specific threshold voltage, triggering a light whenever it comes on. But it won't do it except at the preset threshold.

To check all the voltages and compare them to find out which is lowest, you'd have to sample every voltage and compare them one at a time or simultaneously, and that would get complex again, to the point an MCU would be the best way to do it, unless you just really like analog components.


Using only a few transistors (one or two per cell), you could setup something to display all the voltages as single LED outputs, and simply go by brightness--the dimmest LED is the lowest cell. If you prefer looking for the brightest light, add another transistor to invert the signal and then the brightest LED is the lowest cell.


As for the other problem, you could try putting a ferrite core around the meter leads. If you have any old computer power supplies, there's probably at least one large inductor in it you can take the wires off of to leave you a 1" or greater diameter ferrite core. Just pass your meter lead thru that and wrap it around and thru a few times, near the meter end, and it'll help. Alternately, if you have any old monitor cables around from PC monitors, they probably have a ferrite shell that's molded onto them. you can split the plastic around that and dig out the ferrite halves, then zip-tie them or tape them around the meter leads as a pair. Again, it'll help.

Also in those PC power supplies are some blue (occasionally mustard yellow) capacitor-looking discs, usually right at the input from the wall cord. You can take them out of there and solder across the meter input leads. That'll help "short out" really high voltage spikes.

Without knowing how high the surge goes or how long it lasts, I don't know how well any of the solutions will work. Just have to experiment.


@Icewrench--that first pic has one of the most thoroughly blown caps I've seen in a while.