Is there any definitive way to test them for a shorted out one? Bought all new ones to eliminate a shorted out and blown #28 fuse. So far it seems to have been the right decision. I just hate the thought of throwing them all out if only one was the culprit. c good
Hate to throw all my old coil packs away.
- Thread starter c good
- Start date
You bought six new coil packs to fix one shorted one? I thought it was possible with a cheap OBD II and some process of elimination to figure out which one was bad so you'd only have to replace one.
That's water under the bridge and I could be wrong about that anyway. Do you own a DMM? Someone more knowledgeable will chime in but I was thinking he could just put it in ohms mode and see which one's the odd man out (should have lower resistance than all the others if it's a short)?
That's water under the bridge and I could be wrong about that anyway. Do you own a DMM? Someone more knowledgeable will chime in but I was thinking he could just put it in ohms mode and see which one's the odd man out (should have lower resistance than all the others if it's a short)?
X2 ... you got 6 units to compare your measurements to (assuming that you measurement technique and equipment is good)... those are pretty good odds PLUS... you have 6 others. Should be a lot of resistance values to compare against. Although, if I recall the story correctly, there was some "wiggling" involved so that might make things a bit more rigorous but still doable if you want to "save" a few units.
RedEnvoyDenali
Member
- Dec 4, 2011
- 520
I think I read his problem was intermittent, therefore it may not be an odd man out scenario. Hope I am wrong. 
Well, kinda. It was an intermittent short. It would blow fuse #28 a few times, then it would run fine for a couple of days. Never could just get it to stop working altogether. It never did throw a code, so couldn't isolate which coil was the problem. They're all old so just went for it and replaced them all!! I have to say it's running and starting better than it has in years so something was going wrong.
I would like to figure out a way to test them though just for fun then cut the bad one up to do some forensics on it to see how and where it failed. c good
I would like to figure out a way to test them though just for fun then cut the bad one up to do some forensics on it to see how and where it failed. c good
I do own a Digital Mulitimeter. Just don't know what connections to probe. Each coil has three post connection.
so then without any schematic "help" you have 3 possible measurements... so do three. It probably takes more time to post than it takes to do the measurement....c good said:
c good said:
I dont see why you couldnt use a plain old spark plug tester. That is unless you want to delve futher using an oscope and take a look at burn time and peaks...etc.
c good said:
Id probe resistance across all 3 contacts. 1-2, 1-3, 2-3. Whichever coils have the lower resistance are likely the bad ones.
meerschm
Member
- Aug 26, 2012
- 1,079
toss em. if you have anywhere close to 100k on the plugs, change them out with the proper plugs, and drive on
the coil pack has solid state drivers as well as the coil.
three terminals are for power, control, and ground.
when the pcm commands the coil on, it gets ready, when transitioned to off, it fires the coil.
to find an intermittent short would be difficult. could be activated by vibration, certain temperature, or a crack which makes it sensitive to humidity. if you are really interested, perform as close a visual inspection as you can and see what you can see.
if you find one that acts different, what would you do with it?
or with the others?
the coil pack has solid state drivers as well as the coil.
three terminals are for power, control, and ground.
when the pcm commands the coil on, it gets ready, when transitioned to off, it fires the coil.
to find an intermittent short would be difficult. could be activated by vibration, certain temperature, or a crack which makes it sensitive to humidity. if you are really interested, perform as close a visual inspection as you can and see what you can see.
if you find one that acts different, what would you do with it?
or with the others?
Send em my way. I am actually using what were believed to be two bad coils from a memeber who replaced all of theirs.
I can only think of three ways to discover which one is intermittently bad. (And I design a LOT of experiments to discover root causes of subtle engineering problems)
1) Wire a sub-fuse to the power wire of each coil, of a value lower than that of the main fuse. That ensures the sub-fuse will blow first. This may not be a good use of your time or materials.
2) Run a sequential test - putting in one suspect coil at a time. This can take as many as six experiments to find the bad one, and the time each experiment takes can also be variable depending on what your experience was with the time it took to demonstrate the problem.
3) Run a binary search, which can reduce the number of experiments to two or three, at the cost of more time to change the coils to run the experiment. Try putting in suspect coils A, B, and C. If the problem happens, you know it's one of the three. So take out coil C and see if the problem happens. If it does, then it's A or B, and one last experiment will determine which (3 tries). If it doesn't happen, then the problem coil is C. (2 tries). If the problem doesn't happen with coils A, B, and C, then you know it's D, E, or F, and your second experiment can be on coils D, and E. If it happens, then it's D or E and one more experiment will determine the culprit. (3 tries). If it doesn't happen with D and E, then you know it's F without doing another experiment (2 tries).
Binary searches are particularly useful when troubleshooting or scientific experiments take a long time to run or to set up, or cost a lot in terms of consumables. They are also almost necessary when you have a LOT of variables to consider. I'm in the middle of a problem at work with an opto-electronic system trying to track down about a 1% variation between units that is not calibrating out to a known good standard. I have at least 40 variables, each test can take 8 hours to set up and 4 hours to run, and the customer is screaming at us every day because it's late. Even though the published spec is 3%, they concluded after they bought it that they were wrong and they needed 1% correlation between units. It's a business discussion going on above my pay grade to see how much more they're willing to pay for better units, but it's my job to make sure we can even MAKE better units. Normally I deal with electronic measurements that are 0.05% accurate or so, and the variability of optical systems is pretty frustrating.
1) Wire a sub-fuse to the power wire of each coil, of a value lower than that of the main fuse. That ensures the sub-fuse will blow first. This may not be a good use of your time or materials.
2) Run a sequential test - putting in one suspect coil at a time. This can take as many as six experiments to find the bad one, and the time each experiment takes can also be variable depending on what your experience was with the time it took to demonstrate the problem.
3) Run a binary search, which can reduce the number of experiments to two or three, at the cost of more time to change the coils to run the experiment. Try putting in suspect coils A, B, and C. If the problem happens, you know it's one of the three. So take out coil C and see if the problem happens. If it does, then it's A or B, and one last experiment will determine which (3 tries). If it doesn't happen, then the problem coil is C. (2 tries). If the problem doesn't happen with coils A, B, and C, then you know it's D, E, or F, and your second experiment can be on coils D, and E. If it happens, then it's D or E and one more experiment will determine the culprit. (3 tries). If it doesn't happen with D and E, then you know it's F without doing another experiment (2 tries).
Binary searches are particularly useful when troubleshooting or scientific experiments take a long time to run or to set up, or cost a lot in terms of consumables. They are also almost necessary when you have a LOT of variables to consider. I'm in the middle of a problem at work with an opto-electronic system trying to track down about a 1% variation between units that is not calibrating out to a known good standard. I have at least 40 variables, each test can take 8 hours to set up and 4 hours to run, and the customer is screaming at us every day because it's late. Even though the published spec is 3%, they concluded after they bought it that they were wrong and they needed 1% correlation between units. It's a business discussion going on above my pay grade to see how much more they're willing to pay for better units, but it's my job to make sure we can even MAKE better units. Normally I deal with electronic measurements that are 0.05% accurate or so, and the variability of optical systems is pretty frustrating.
budwich said:so then without any schematic "help" you have 3 possible measurements... so do three. It probably takes more time to post than it takes to do the measurement....
As suggested by others, there are probably "better" tests that include significant effort but since you are doing this more as an "experiment" (you have already replaced them), I would add a "blast of heat" from a hair dryer or equivalent while doing these simple measurements.... this may help with finding some "expansion" type issues that may have caused your intermittent issue.
Further, the "heat method" might also be useful on the connectors / harness on the intake area to further check your setup although if things are "happy", you may not want to disturb them any more.
the roadie said:
Being educated as a computer scientist I can appreciate the usefulness of using a binary search tree. But in this case its a bit overkill and the description of how such searches are performed can be a bit confusing to everyday folk. But nevertheless its a good idea.
But as Sheldon Cooper said when his companions told him it was useless to try to educate the DMV clerk that their test was wrong: "How else are they going to LEARN?!?"CaptainXL said:
SheldonQuotes.com » Arguing with the DMV
View attachment 31237
Attachments
Hey, I learned a lot from those 3 suggestions. I don't have the OP's problem, but now I've learned some ways to go about troubleshooting an electrical component problem in the future.