Replacement ignition coil insanity (2022 edition)

I could tell I was having some misfires and found them to be accumulating on cylinder #7. No misfires on any other cylinder. I swapped the coils from #5 to #7 and it results in the same issue on cylinder #5 (which should rule out anything specific to cylinder #7 as a potential cause). Plugs recently replaced with NGK OE spec. Plug #7 examined and it looks acceptable.

I order a replacement coil:
Started with CUF2890 / UF597 - about 5 minutes of driving and I get a half-power MIL (RMA in progress)

Autozone branded coil - ditto (returned)

NGK coils (from different sources) both immediately result in a "hard" MIL for reduced power about 2 seconds after the car starts (one distributor refunded me for the part and didn't even want it shipped back)

Napa Echlin - had it brought to local store, it looks like the others, didn't buy it to bother trying it

I have a UF597 from another source on the way Monday but I'm skeptical that it will work, either.

Are there any known ignition coil part numbers that someone has purchased recently that are known to work (aside from $220 or more from BMW)?

I was on the same boat a couple months ago. None of the aftermarket coils worked, including NGK's, had to buy 2 oem ones from FCP.

I had the same issue ended up being an exhaust gasket not sealed properly giving a miss fire going into limp mode. I changed coils and plugs all OEM still had the issue $2k later. I found the exhaust leak at the header and test pipe giving the wrong A/F reading compared to the other bank setting off a miss fire. Check for that a $20 donut gasket fixed my issue

Have you tried buying directly from NGK? All I hear is horror stories and returns of coils unless anything buying directly from BMW.

I haven't seen feedback on buying directly from NGK.

Trinh_7 wrote

I was on the same boat a couple months ago. None of the aftermarket coils worked, including NGK's, had to buy 2 oem ones from FCP.

Thanks for your reply. I am about at my wits end on this and thinking that is the only viable solution at present.

One of the comments in one of the online listings for a coil was a mechanic who purchased two, went through a saga to actually get them, only to have them do the same thing that I'm seeing. Mechanic said he had to acquire used coils to get his customer back on the road. It does not makes much sense to me to buy a used OEM BMW coil only to have to replace it again...although, I am now so good at it, it only takes like 3-5 minutes to pull the intake and swap coils on that side.

akamal92 wrote

I had the same issue ended up being an exhaust gasket not sealed properly giving a miss fire going into limp mode. I changed coils and plugs all OEM still had the issue $2k later. I found the exhaust leak at the header and test pipe giving the wrong A/F reading compared to the other bank setting off a miss fire. Check for that a $20 donut gasket fixed my issue

Did you have a misfires accumulating on only one cylinder with the exhaust leak? I would think that you would have misfire codes for all the cylinders upstream of that gasket in that scenario (e.g. 1-4 or 5-8 or both if both gaskets were faulty).

I only have a misfire on cylinder 7. The misfire follows the coil, which is the EXACT criteria spelled out in the BMW factory manual with regards to misfire diagnosis.

Ervin87 wrote

Have you tried buying directly from NGK? All I hear is horror stories and returns of coils unless anything buying directly from BMW.

I haven't seen feedback on buying directly from NGK.

I bought one from Autozone. No go. Immediately throws a pending P138B upon startup.

The first one did the same thing. I was utterly shocked. Never in my tenure of working on cars did I ever think NGK would ship out known defective merchandise. I don't know why they haven't recalled all of them to save their brand image.

The first one I bought was from NGK.com, which is not NGK direct but a large distributor of NGK parts who wrote this back to me:

Thank you for contacting us! We have processed a refund for the defective part and will not require it to be returned at this time. Seem's like NGK is having an issue with this particular part NGK 48729 U5118 Ignition Coil that we are looking into.

They didn't even want me to ship it back and credited me the full amount for the part and outbound shipping. It wasn't even worth it to them to get it back into inventory.

Today's update:
I ordered and got another one. It is branded as Intermotor UF597.

Surprise, surprise.

Same result as the other dual labeled CUF2890 / UF597. Amazon for this one, so relatively easy return at the local Kohls next time I'm out that way.

Car starts and runs fine (doing this on a totally cold car with the ECU reset so there are no misfires in the count). Drives fine for about 5-10 minutes. No misfires counted, so I am pretty sure that the coil is the problem, since I can hear/feel it miss on the OE coil and it seems to be better on the new coil (same behavior as the other one, improved BUT fails a different ECU test parameter).

Unfortunately, while it seems to operate better, it goes to reduced power on almost the exact same point in my OBD drive cycle loop that the other one did...when you are working to clear the 40-60 mph portion (6 of the 9 indicators are ready/have passed and it's working to pass 7-9).

Just about when you think you are in the clear...reduced power....proceed with caution.

I can reset the ECU and swap back the OE coil and the car seems to go past that point of the drive cycle and the misfires aren't enough to fail the OBD drive cycle but it _IS_ missing and it will eventually accumulate enough misfires to result in a MIL given enough time.

Here is what I think might be the issue with the "bad coils":
The coil has to be failing something the ECU tests in regards to coil operation being within normal parameters. The BMW tech manual mentions that the ECU expects 2-6 coil oscillations after complete discharge and if it doesn't see any, it may lead the ECU to think that the coil is out of spec. Is it possible they are "too good" and since they don't oscillate, the ECU throws up the flag and doesn't like them?

The NGK coils fail in such a manner that they are out of spec within the first coil cycles right after the engine starts. ~1000rpm idle and about 2-5 seconds...that's like 30 rotations on the low end, each coil fires every what, 2 rotations? So, somewhere on the order of 15-50 ignition events causes the ECU to deem the NGK faulty.

Makes no sense to me how these coil can drive better but fail the drive cycle. I could try to swap coils without the drive cycle and ECU reset and it might work if the test that is failing is only performed during the initial OBD drive cycle but somehow, I doubt that would work long term and then if I ever have to reset the ECU, I have to keep the "dying" OEM coil and swap back and forth every time. That definitely does not seem like a viable solution.

It looks like I have no choice but to order a $230 OEM coil at this point. Continuing to drive it trying to wait for a non-OEM cheaper coil with a misfire could foul up the catalysts and I don't want to spend even more money on one of those.

You could gamble on a used one off ebay.

I just got off the phone with Standard Motor Parts tech support.
They claim they get the UF-597 from BMW.
Uhh. That bodes rather poorly if that information is correct.
Made that guy aware that I've now tried two UF-597 and they both have the same problem. Hopefully someone passes the message up the chain and they actually do something about this instead of assuming they are stocking/shipping a working SKU.

I guess I will be ordering an OEM BMW coil from Amazon to try.

I also called NGK corporate tech support (not to be confused with NGK.com, which they acknowledge sells authentic NGK product). NGK seems to be aware of the issue.

I was transferred to "WVE" (the division they acquired that makes them - I believe that this is Wells Vehicle Electronics - https://wellsve.com/) and that support staff gave the most complete answer as of yet which aligns with my research (which would also mean that Wells 5C1699 and anyone who rebrands that part are also affected):
These coils are "complicated" and have the ion sensing system and send data back to the ECU after firing. The ECU doesn't see the data and in its absence, shuts down the cylinder. They believe that the microchip in the coil is not right/in spec or otherwise not providing this data back to the ECU which is why the NGK immediately fail.

My overall stance was if they know about it, why have they not:
1) Recalled all of them (they don't want to until they determine if it's just one batch or all of them)
2) Put a stop sale on them until further notice so that mechanics are not having to eat billable time themselves or charge it to customers dealing with a known defective product (close to my chest having owned and run an automotive service business)
3) Do #1 or #2 ASAP to save the brand reputation. I have been almost exclusively selecting NGK plugs/O2 sensors/etc. for 25+ years and was shocked to know they have a bad product out and they haven't done the "right thing"

I contacted NGK about a month ago. They stopped making ignition coils for the S65 some time ago.

AFAIK, the only company now making coils are Eldor and they only supply BMW. They won't sell to anyone else. BMW now charge 4 times the amount they were charging a couple of years back. The situation stinks.

bond_007 wrote

I just got off the phone with Standard Motor Parts tech support.
They claim they get the UF-597 from BMW.
Uhh. That bodes rather poorly if that information is correct.
Made that guy aware that I've now tried two UF-597 and they both have the same problem. Hopefully someone passes the message up the chain and they actually do something about this instead of assuming they are stocking/shipping a working SKU.

I guess I will be ordering an OEM BMW coil from Amazon to try.

I also called NGK corporate tech support (not to be confused with NGK.com, which they acknowledge sells authentic NGK product).

NGK seems to be aware of the issue. I was transferred to "WVE" (the division they acquired that makes them - I believe that this is Wells Vehicle Electronics - https://wellsve.com/) and that support staff gave the most complete answer as of yet which aligns with my research:
These coils are "complicated" and have the ion sensing system and send data back to the ECU after firing. The ECU doesn't see the data and in its absence, shuts down the cylinder. They believe that the microchip in the coil is not right/in spec or otherwise not providing this data back to the ECU which is why they immediately fail.

That would also mean that Wells 5C1699 and anyone who rebrands that part are also affected.

My overall stance was if they know about it, why have they not:
1) Recalled all of them (they don't want to until they determine if it's just one batch or all of them)
2) Put a stop sale on them until further notice so that mechanics are not having to eat billable time themselves or charge it to customers dealing with a known defective product (close to my chest having owned and run an automotive service business)
3) Do #1 or #2 ASAP to save the brand reputation. I have been almost exclusively selecting NGK plugs/O2 sensors/etc. for 25+ years and was shocked to know they have a bad product out and they haven't done the "right thing"

I have 4 coil packs I ordered from NGK.com back in May. Haven't installed as their part of my track spares kit but sounds like I should just send them back.

Regarding selling with known issues, AIM is also doing this. They're knowingly selling their data loggers and dash systems without the various accelerometer chips that provide the acceleration data (long, Lat, vertical G's). I can appreciate that their supply chain is distrusted, but nothing has changed in their marketing to reflect the absence. They're selling these at full price and make no mention until you reach out o inquire about the missing channels in your data set. I told their support guys it amounts to fraud, but their president and sales VP don't seem concerned. They'll eventually have a solution to have impacted folks send their units for upgrade at some unknown point on the future. Argh…

From FCP Euro's Website wrote  

Note: The ignition coils used in the S65 engine are far from standard ignition coils. In order for the S65 engine to make maximum power at close to 8400 RPM's continuous ignition monitoring is required more than what traditional acoustic knock sensors can provide. The S65 engine uses an ionic current monitoring system derived from the S85 V10 which actively and continuously monitors combustion quality. Unlike the S85 engine which uses an inline control unit for each cylinder bank, the simplified system utilized on the S65 engine has incorporated the system in the individual ignition coils themselves. During ignition cycles, the measurement current is stored in a capacitor integrated in the ignition coil itself, and after ignition, is available at the spark plug electrode. In the S65, the ion current measurement and evaluation is also performed exclusively by the MSS60 DME.

That's also probably why the others are failing about 5-10 mins later...they likely work for a while but stop reporting data once they warm up enough...

DRLane wrote

I have 4 coil packs I ordered from NGK.com back in May. Haven't installed as their part of my track spares kit but sounds like I should just send them back.

Regarding selling with known issues, AIM is also doing this. They're knowingly selling their data loggers and dash systems without the various accelerometer chips that provide the acceleration data (long, Lat, vertical G's). I can appreciate that their supply chain is distrusted, but nothing has changed in their marketing to reflect the absence. They're selling these at full price and make no mention until you reach out o inquire about the missing channels in your data set. I told their support guys it amounts to fraud, but their president and sales VP don't seem concerned. They'll eventually have a solution to have impacted folks send their units for upgrade at some unknown point on the future. Argh…

You can swap all four out on the driver's side easily and see what the ECU does. On the NGK, the ECU immediately went into half power mode at startup. If they don't do that, I would complete the OBD drive cycle and if they keep working...you somehow got a good batch...

As a HPDE instructor (haven't been on track in a while, though) I am shocked to hear that AIM would do that. It's one thing to ship hardware with half-baked firmware or needing firmware update but to ship a datalogger without GPS?

I get everyone is "just trying to get by" but stuff like this is inexcusable. The whole world seems to be resorting to this nonsense. We need to hold these companies accountable for their practices and vote with the wallet by purchasing from competitors so they prosper and cheaters go out of business.

Just get it from your Local BMW dealer. I've never ever had issues with new OE BMW

Trinh_7 wrote

I was on the same boat a couple months ago. None of the aftermarket coils worked, including NGK's, had to buy 2 oem ones from FCP.

I’m in the same boat with the Autozone coil. Any issues with the code coming back after the FCP coil?

This is an all too common issue with aftermarket parts for our cars. I’d think there would be a tune to remedy this issue.
Has anyone heard of any tune out there that would override the ECM on the coil issue?
I know misfires are bad, but it doesn’t happen until the ECM decides it t doesn’t like the coil.

bond_007 wrote

DRLane wrote

I have 4 coil packs I ordered from NGK.com back in May. Haven't installed as their part of my track spares kit but sounds like I should just send them back.

Regarding selling with known issues, AIM is also doing this. They're knowingly selling their data loggers and dash systems without the various accelerometer chips that provide the acceleration data (long, Lat, vertical G's). I can appreciate that their supply chain is distrusted, but nothing has changed in their marketing to reflect the absence. They're selling these at full price and make no mention until you reach out o inquire about the missing channels in your data set. I told their support guys it amounts to fraud, but their president and sales VP don't seem concerned. They'll eventually have a solution to have impacted folks send their units for upgrade at some unknown point on the future. Argh…

You can swap all four out on the driver's side easily and see what the ECU does. On the NGK, the ECU immediately went into half power mode at startup. If they don't do that, I would complete the OBD drive cycle and if they keep working...you somehow got a good batch...

As a HPDE instructor (haven't been on track in a while, though) I am shocked to hear that AIM would do that. It's one thing to ship hardware with half-baked firmware or needing firmware update but to ship a datalogger without GPS?

I get everyone is "just trying to get by" but stuff like this is inexcusable. The whole world seems to be resorting to this nonsense. We need to hold these companies accountable for their practices and vote with the wallet by purchasing from competitors so they prosper and cheaters go out of business.

It's not as bad as excluding the GPS, but still the accelerometer/gyro/mag are critical and thus knowingly misleading the consumer is egregious.

TonysM32011 wrote

This is an all too common issue with aftermarket parts for our cars. I'd think there would be a tune to remedy this issue.
Has anyone heard of any tune out there that would override the ECM on the coil issue?
I know misfires are bad, but it doesn't happen until the ECM decides it t doesn't like the coil.

Isn't the plug/coil used for knock detection?

TonysM32011 wrote

I’m in the same boat with the Autozone coil. Any issues with the code coming back after the FCP coil?

They were genuine BMW so no problems. Car has been running perfect since then. The only problem was the price.

Here's my post about NGK coils. Also tried to buy 2 coils directly from NGK and they were junk. https://www.m3post.com/forums/showthread.php?p=28883913

Trinh_7 wrote

They were genuine BMW so no problems. Car has been running perfect since then. The only problem was the price.

Here's my post about NGK coils. Also tried to buy 2 coils directly from NGK and they were junk. https://www.m3post.com/forums/showthread.php?p=28883913

I hear you. FCP it is.

TonysM32011 wrote

This is an all too common issue with aftermarket parts for our cars. I’d think there would be a tune to remedy this issue.
Has anyone heard of any tune out there that would override the ECM on the coil issue?
I know misfires are bad, but it doesn’t happen until the ECM decides it t doesn’t like the coil.

Unlikely to override an essential ECM function of knock sensing at high engine RPM through the ignition coil. The fancy coils are why there is no knock sensor. The ECM determines knock after the ignition coil fires vis the ion-sensing function. It’s actually pretty smart where you get knock monitoring for each cylinder.

The aftermarket suppliers have the wrong chip in the coil, most likely. Like everything else these days, short on supplies or upstream supplier sent them garbage. The aftermarket stuff did work until this most recent batch. I just wish they would admit to their fault and pull stuff off the shelves.

Uro did this, from my conversation with ECS if they had any insight on when the back orders would come in (no idea and they basically said get OEM at this point in time).