FLYVMO

I am looking for some detailed information regarding these electric starter data points for a typical 928 electric starter motor:

Nominal and maximum current
Inrush current
Inrush current duration
Nominal power

Anyone out there that would know?

Cheers!
Carl

FLYVMO

Bumbiediebump..

Does anyone know how to measure the above and what device would one use?

Cheers!
Carl

dr bob

Carl, I use a power analyzer for most tasks like this, but mine doesn't read low-voltage or instant DC current very well. Meanwhile, there are some interesting clamp-on ammeters (Fluke and some chi-clones) that include peak-hold features so you could see the maximum current numbers. But not the duration, as that's a function of how quickly the engine starts. Inrush and max current will be the same, FWIW. I'd use my antique storage o'scope to plot the inrush-current duration if that became important.

For those playing at home, specifying conductor sizes needs the info that Carl asks for above. The basic current-carrying capacity of a cable includes factors for type of insulation and how the cable is installed, so that cable heating doesn't exceed the capability of the insulation and voltage drop is not going to severely cramp the ability of the starter to start the car. Using basic electrical cable-sizing methods, the battery cables are too small. It's only because the use duration is so short that we get away with using the cables we do. The cost of the copper and the cost of dragging the weight around are certainly considered.

FLYVMO

Thank you Dr. Bob!

Great information, which generates some additional questions. Is it possible to mathematically calculate approximate data based on knowing the compression ratio, battery voltage, starter cable gauge and cranking time?

I have found some general data on 4 cylinder engines, peaking at 110A and then settling at around 101A for the cranking period, battery at 12.45V in the example. I have heard statements that the 928 is peaking around 250A but no solid data to confirm those statements. I assume the inrush duration is for a time period of less than 1 second for a typical car starter and nominal and duration/amps are for a typical cranking time frame, let's say less than 5 seconds?

Cheers!
Carl

ramcram

Carl, I think 12.45V cranking voltage is a bit optimistic. Remember there will be no charging at that rpm, so the cranking voltage will be much less than static battery voltage.

FLYVMO

Correct, I should have clarified that it was the static battery voltage prior to cranking in the mentioned case, a not completely charged battery.

Cheers!
Carl

dr bob

Carl, your numbers are probably OK, but... wondering why you are concerned. As mentioned, the factory cable sizes are a good compromise between what's "needed" for continuous duty and what "will work" thanks to the relatively short duty cycle of the starter current.

---

I was un-summerizing my snowblower a couple weeks ago. It was carefully put up with tank and carburetor drained, and a couple squirts of ATF in the cylinder prior to storage. Plug was finger-tight. So the plug comes out, pull the cord to rotate the engine and clear any remaining ATF. Plug back in and torqued. Fresh fuel, a couple pokes at the primer pump. Then pull the cord slowly to bring the piston up on the compression stroke. Usually this is followed by a hard pull on the cord, but this time it started on the slow pull to compression. So how easily will your engine start? A lot depends on where it stops. The ignition and injection both need to see the index gap pass the CPS before they know enough to make a spark and spray some fuel into the intake. How many cylinders will go through compression stroke before that happens? Might be one, might be four. Big difference in the starter current demand profile and duration.

Thank goodness there's no need for APU boost for normal starting.

FLYVMO

Hi Dr. Bob,

The questions were not related to my car, but for the 928 in general, where I am trying to figure out what the inrush and nominal starter currents are for a solid state project aimed at the 928. Those data points are needed in order to program the threshold for opening the circuit to prevent overload.

Would a typical volt/amp meter connected to the battery terminals give an accurate measurement of those while cranking?

Cheers!
Carl

Alan

The voltage at the battery alone won't tell you anything meaningful about the current - since you don't know the battery internal resistance or the rest of the loop resistance (wiring & starter).

You really do need a DC clamp meter and DC clamp meters are expensive (AC clamp meters are much less so but you cannot use them for this purpose).

You could also use a high current shunt and voltmeter to measure current - but for this high of a current application the shunt would likely affect the results significantly, so not ideal.

You also need to use devices that can measure accurately Max/Min numbers and hold them (Volts and Amps). The event is short and you cannot realistically get a min voltage or max current reading looking at a normal meter that takes time to settle.

You need this Max current and Min voltage to understand the peak conditions. But you'd need to take data over at least 10-15 starts then you will have ~worst case peak data from the Max/Min reading (for the conditions you are testing in e.g. primarily temps) but it won't tell you anything much about duration or off peak profile which is also a real factor for wire sizing.

For that you'd need to use a trace method to really have a good feel for what is going on. In this case you need a really big (low impedance) shunt in series and a way to gather dual measurements of voltage and current over time - say on a digital storage scope. You'd then also need to post process the bulk results across many different starts (10-20 - with the battery similarly charged each time) to get meaningful data. You can then ~correlate the max/min data to the trace to create an approx adjustment factor for the shunt impact on the peak. It's actually all quite complicated to do it right.

Alan

FLYVMO

Thanks Alan for that excellent explanation. It gives me a road map of what to do next.

So the best place to measure primarily would be at the starter itself, then the battery?

Cheers!
Carl

Alan

For this application I'd measure voltage across the battery terminals and current in the battery ground strap - this is probably easiest and without adding a longer loop into the starter connection you will get the same current results.

As long as you only need total system values for a generic 928 this is likely good. If you wanted to characterise the starter alone then you might want to measure the voltage across the starter (terminal to case) more directly. This is possible but a little awkward - however a reasonable proxy for this can be measured at the front jump post to an engine ground (not a chassis ground), this will be quite close to the direct measurement.

Alan

FLYVMO

Great, thank you Alan!

Cheers!
Carl

dr bob

Late reply, as Alan answered some of this already.

A standard in-circuit amp meter needs to have 1000-amp capability, about 100 times what they come with, to read starting current. The battery has somewhere around 750 cranking amp capability, minimum.

Because of the very high current demand, and the fact that the current demand varies a lot based on ambient and more specifically engine temps, the starter circuit is typically left out of "normal" protection schemes.

There are inductive clamp-on devices for DC circuits that might be interesting. These introduce a bit of AC field around the DC conductor, and measure the shift in the AC to estimate the DC current passing through the cable. A device with peak-hold capability will help but won't give you the total inrush or peak-current times.

Can you expand some on how your SS protection device is going to measure starting current? Time-overcurrent protection on large AC motors is pretty common, using current transformers ("CT's") inserted in each feeder phase. No such option exists for the DC starter motor in the 928. You could add a high-current shunt in series with the battery as Alan suggests, then look at voltage drop across the precision shunt to calculate current. To be even loosely accurate though, you need to compensate for temperature and temperature rise during testing. Suddenly a simple voltage-drop measurement needs some real-time integration/correction for temperature.

FLYVMO

Hi Dr. Bob,

Thank you for the technical insight and suggestions. While I am not ready to divulge too much detail yet, as the project is in early infancy, basically the data is needed to give the "device" a baseline of what to "expect" when triggering the starter (or any other load). It can be programmed for a wide range of starter setups, and other "power users", but basically it is a part of a bigger whole, so let's call it a solid state PCM (power control module).

I will post the measured data here for posterity once a measuring device has been acquired and a suitable 928 has been measured

Cheers!
Carl

jcorenman

Carl,

I've got a Fluke clamp-on DC ammeter which I believe is good for 600A, and a recording digital scope. I've done the measurement you want, DC amps on the ground strap while cranking, but didn't save the data. If you can give me a few days I can do that again and post the data.

Offshore sailboats have serious energy-management needs and are usually fitted with high-current shunts to keep track of things, including the starting bank.
A typical one is here, 50mV output for a 500A load: http://www.westmarine.com/buy/blue-s...shunt--3827151

A "calibrated" piece of wire would also do for a shunt-- all you need is the resistance per foot for whatever wire you use for a ground strap. Or, for a factory strap (in good condition) just measure the voltage drop while doing the current test above, which tells you the resistance.