We provide support for a number of systems installed in 24V vehicles. When the vehicles are not in use, the users like to leave the system's main power switch and circuit breaker both in the on position. If the vehicle then needs to be jump started, the users often do not first turn off the system's circuit breaker or power switch, despite continuing attempts at training them. Jump starting under those conditions tends to destroy a very expensive electronic module in the system.
Modifying the systems themselves is not a viable option. And we do not yet understand the exact failure mechanism that is occurring. We do know that there is an internal SMPS upstream from the at-risk module, but don't have schematics for the system. The system draws about 35 Amps continuously, when operating.
Knowing only that, is there something that could be strapped across the system's DC power terminals that would prevent damage from jump starting? Something available off-the-shelf would be ideal. If nothing already exists, what might work?
I recall a similar problem. It turned out an inductor in series with the power was causing overshoots on the power supply rail big enough to kill an opamp. A big capacitor solved the problem.
OTOH: in 12V automotive systems people usually design stuff to withstand 60V (IIRC) spikes on the power rails.
--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------
I'll have to check what voltage the alternator produces. It's probably 27-28v. Maybe trying to limit the voltage to 30v would be good.
So, a number of hefty zener diodes in series to get their conduction point above the normal operating voltage; How much current could flow through them, then, at say 48v? No resistance in series with the zeners? Would like them to work more than once. Or would it not be a problem?
I bet it's frying because a completely dead lead-acid battery becomes a non-conducting bucket of water. There could easily be 20VAC power noise until the battery starts conducting again. The downward cycle is usually the fatal one because it draws power out of onboard capacitors in a path that the semiconductors weren't designed to handle.
--
I won't see Google Groups replies because I must filter them as spam
It's the load dump that you need to absorb. Probably fails when the jumper cables are connected. I'd put say 0.01ohms in series with the 24v then the 30v Zener to gnd. That turns out to be a 15w resistor, so you can choose something lower.
What about using TVS (silicon avalanche diode) type of device?
I was just looking at these at mouser.com . For example, for the Littlefuse 15KP28 (Mouser 576-15KP28), they list Operating Voltage
28v, Clamping Voltage 47.5v, Breakdown Voltage 31.1v, Peak Surge Current 316 A. (probably not for long!). I'm not sure what they mean by Operating Voltage. I think I would need a Clamping Voltage around
30V. But those all have Operating Voltage specs around 17-18v.
OK, the datasheet calls 28v the "Reverse Breakdown Voltage", which I shouldn't have to think about unless the jumper cables are accidentally reversed, right? And it calls the 47.5v the "Maximum Clamping Voltage at Ipp" (i.e. @ the 316 Amps max). So it sounds like when it starts to conduct, the actual voltage across it will be higher if the current is large, which sounds reasonable. If that's the case, would it work to parallel two or more of them, to keep that max voltage down for the very high currents that would probably ensue?
So, among the 15KPxx devices at least, I guess I would want one with a breakdown voltage that is about where I want to clamp.
Should I think about using a bidirectional one, or just try to get one that also has a reverse breakdown voltage that would be a bit below the negative of the normal alternator voltage, in case they hook the jumper cables up backwards?
I guess similar considerations could be applied to Zener diode specs.
It looks like those Littlefuse 15KPxx devices might be obsolete. But I guess there are others that are similar.
Yeah, that type of circuit would be nice. Jim Thompson also has one that's similar in effect I think, but all discretes. And Maxim has a discrete version schematic of something similar in some literature about a similar IC they sell.
Does anyone know if there any COTS products (already available) that use something like that?
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html
Well, we probably have 20 feet of heavy wire between the system and where it connects directly to the vehicle's batterries. But our system is not drawing any current when the problem occurs.
That would be nice. But how to know what kind and what rating for the fuse? The TVS diode devices can dissipate 15000 Watts for a very short time. So we'd want to let them do that. I guess I would need to know what sort of voltage vs time profile might be involved, during jump starting, and go from there, or else go for massive overkill in the zeners
I also found the Littlefuse 30KPA28CA TVS diodes, which are good for
606 Amps peak pulse current and 30 kW peak pulse power dissipation. Their breakdown voltage is 31.26 volts.
From the datasheet at
formatting link
use_TVS-Diodes_30KPA.pdf , it looks like they would have up to 50 Volts across them if they were conducting up to 606 Amps (which I guess they could only do for some smallish number of microseconds).
I think I read somewhere that a load dump transient can last for a good portion of a second. But I guess I'll have to do some jump start testing, to find out what a typical jump start transient voltage waveform looks like at the system's power supply input (using a suitable dummy load in place of the system). Even then, I probably won't have enough information about the TVS parts to know if I need to use mroe than one in parallel. Maybe the best way is to just look at ir with no TVS diode, and then look at it with one, then two in parallel, etc, and see what that does to the voltage transient (or find the point at which the TVS's stop exploding).
If you are using a connector to connect the jumper cables (like a military slave connector) you could rig a relay/contactor that would not allow power unless the switch was off.
Yes, you can do that, and the advantage is that you only need one.
Right. As with zeners, the devices are primarily rated by power (maximum watts that they can handle), rather than current. And TVS specs are oriented to transients, not continuous currents, because that is what they are intended for.
Look at the datasheet. Littelfuse calls it "standoff voltage", and so does Vishay. Match this to the highest voltage you want to pass, and keep in mind that the device will flow a very small amount of current at that voltage (which for the 15KP series can easily be 5 mA or more). You may want to use a 26V or higher device rather than 24V to allow for a voltage from the battery of greater than 24V, and whatever else might be considered nominal.
After standoff voltage, look at the breakdown voltage, and to be safe, use the maximum specified in the datasheet. This is the avalanche breakdown of the device, and it can flow a lot of current without the voltage going much above that. This is usually what limits the peak of the surge, and the actual peak of the suppressed voltage will depend on how much current is flowing.
To be very clear about this, DO NOT select a device with a breakdown voltage anywhere near the rated 24V of your battery!
As others suggested, your problem is probably from some inductance causing a high-voltage spike, maybe hundreds or thousands of volts. It's unknown what maximum voltage the sensitive electronic circuits can handle, but it's possible that it's thousands of volts that are killing them, and not tens of volts.
Be careful not to set your standoff/breakdown voltages too low, because if you do, it will cause trouble during normal operation.
Yes, that's it.
I suggest you try to determine what kind of surge you have. Can you put a 'scope on the circuit? Maybe take one of the vehicles, and follow the recommended jumpstarting procedure and watch what happens? Doing this will probably give you a much better idea of what you're dealing with, so you can be more sure you have a good solution. And if you get a few TVS or other devices, you can add them to the circuit and see if they do what they need to.
Putting surge suppressors in parallel is not always a good idea, since in some cases it can reduce the response time. (I got this advice with respect to MOVs, and I don't know if it applies to zeners and TVS devices.)
Yes.
You can, and it simplifies and failsafes the installation because it can't be connected backwards.
More likely they just don't sell enough of them to stock them. And if that is the case, it might be a hint that you don't need one that big.
You might do fine with a smaller device (e.g., the 1.5KE series, which are 1500 watts). Do a more general search at Mouser for TVS devices and see what they have less than 15 KW. You will see increasing availability with the smaller devices.
In Vishay's 1.5KE line, there are -A devices that have tighter breakdown voltage specs (5% rather than the usual 10%). This is helpful because a lower-voltage device can be used to keep the let-through voltage (maximum peak of the surge) lower.
In general, read the datasheets carefully and notice that for each voltage spec, there is a matching current that the voltage is measured at. When you understand this, you will be able to handle standoff vs. breakdown vs. clamping voltage specifications easily, and it should all make sense to you.
Also, lower power handling devices may have a smaller difference between standoff and breakdown voltages, for better protection if you don't need to handle more power. I don't remember this well enough from when I was studying it, so check me to see if I'm right on that.
BTW, you also might consider putting additional TVS devices at the power inputs of the sensitive electronics in the vehicle. This is a little different from paralleling them at one point in the circuit because it takes time (a few ns, but that is enough to matter) for the surge to travel around the circuit.
There you go, a _proper_ solution... avoid the problem in the first place ;-) ...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Obama: A reincarnation of Nixon, narcissistically posing in
politically-correct black-face, but with fewer scruples.
Thanks, Jim. I wish it were that easy. We could never get the users to go for that and we basically have no control over them. And remember I said they won't even follow the instructions to turn off the breaker before jump starting.
But I did briefly entertain an idea for a sort of automatic switch at the system's 24V input, i.e. a relay that would disconnect the 24V line when it went below a threshold, and then have a fairly long time delay before closing it again when the voltage came back up. That would be a rather "jump start specific" form of protection. But the system would never be drawing much (if any) current when it was dis- connecting or re-connecting, which shouldn't happen a whole lot, anyway. So the relay contacts should stay good. But it seems like a TVS or Zener clamping circuit would be a lot cheaper and smaller, and maybe more reliable and robust, and should work well-enough. And they are available and ready to use as is. There are too many of these systems out there to spend a lot of money on each one.
Just for everyone's information, here a link to a paper that looks like it has a pretty-good summary of the properties of the various types of automotive transients:
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.