95 325is. Oscillating mpg gauge with hesitation to match.

[snow663]

So I got rid of my old beater 325i and purchased a 150k mile 325is coupe. Had a few small issues and one of them was the ews2 system. I swapped the DME with a red label 413 and it fired right up but it has the same problem my last car had. I though maybe a bad DME so I got another one, it's still there.

The car runs great for the first couple moves. Like pulling out of the driveway. Then once i hit the first stop sign it hesitates pulling off the line and then when I hit the highway it feels like it just chokes up and I feel like I'm pulling a car behind me. If I hit the gas the power comes back but accelerates in waves and the mpg gauge reflects these "waves" of power.

[ed323i]

Can be several things.. Easiest to try first is just disconnect the MAF connector. If it runs fine then, then you'll be sure it's the MAF. If it's the MAF you can try cleaning it, or, if that doesn't help, you can get another one. Be sure to get an original VDO/Siemens one (a used one is fine, new very expensive), as the cheap $30 Ebay MAFs are a complete waste of money (believe me)..

[snow663]

I may have a spare actually, I'll give it a try. My last BMW had the exact same engine. It was a 93 325i. On that car I had the same issue and I hunted it for a year. Finally one day a coolant fitting broke. After I fixed it the problem dissapeared. It appeared to be an old hairline crack that existed for a time before it broke.

So I decided to look and here's where I'm confused. The cooling system is full cold (after work). And after a 20 Mile drive doesn't appear to leak but it doesn't build pressure, hose can still be squished with the engine hot, heater works fine, and there's no air from the bleeder screw??? Perhase it's a simple issue. But where do I start?

Sent from my Moto G (5) Plus using Tapatalk

[snow663]

Okay so today I pressurised the cooling system and intake, sounds like I can hear a small leak at the radiator cap. And a small vac leak at the oil filler cap. I bent back the tabs on the cap to get a tighter seal, no more vacuum leaks. But looks like I need a rad cap. Is it likely that this could cause my hesitation? The o rings look fine idk why it leaks.

On a side note I do have a spare maf but I really want to eliminate one variable at a time. No codes in memory. So it's gotta be something that affects the engines base tune. It runs fine at high rpm so I doubt it's a fuel pressure/volume issue since most of my lag is down low. It's very consistent. And almost seems to keep rhythm regardless of all variables. It's basically wildly swinging between rich/lean

Sent from my Moto G (5) Plus using Tapatalk

[ed323i]

I can't think of a way how a leaking radiator cap would cause the symptoms.. It will prevent the cooling system from pressurizing, which means you can easily get the coolant to boil if the engine gets a bit hotter. It can also boil after you've turned off the engine (coolant being heated by the heat of the head, when the pump is no longer moving the coolant).

My first bet remains the MAF.. All you have to do is disconnect the MAF, and drive it. If everything else is good, the engine should run just fine without MAF.
Next one to check would be the oxygen sensors. You can also just disconnect it. Best to rest the DME the adaptation values in the DME first though (use INPA or some other tool, not sure if removing power from the DME for a couple of hours also does the trick). The the DME will use the pre-programmed injection times in the DME (just like the old 1980's pre-CAT Motronic BMWs ran).

The M50 only has one pre-CAT oxygen sensor (at least the European ones do).. If that's going bad, then it might run fine when the engine is still cold, and the DME is not using the oxygen sensor input, and then, if the oxygen sensors give false rich/lean values, it can cause the shifts in the mpg gauge (because the DME thinks it should inject much less or more than what the engine actually needs)..

Good luck!

[snow663]

I can't think of a way how a leaking radiator cap would cause the symptoms.. It will prevent the cooling system from pressurizing, which means you can easily get the coolant to boil if the engine gets a bit hotter. It can also boil after you've turned off the engine (coolant being heated by the heat of the head, when the pump is no longer moving the coolant).

My first bet remains the MAF.. All you have to do is disconnect the MAF, and drive it. If everything else is good, the engine should run just fine without MAF.
Next one to check would be the oxygen sensors. You can also just disconnect it. Best to rest the DME the adaptation values in the DME first though (use INPA or some other tool, not sure if removing power from the DME for a couple of hours also does the trick). The the DME will use the pre-programmed injection times in the DME (just like the old 1980's pre-CAT Motronic BMWs ran).

The M50 only has one pre-CAT oxygen sensor (at least the European ones do).. If that's going bad, then it might run fine when the engine is still cold, and the DME is not using the oxygen sensor input, and then, if the oxygen sensors give false rich/lean values, it can cause the shifts in the mpg gauge (because the DME thinks it should inject much less or more than what the engine actually needs)..

Good luck!

Okay so I ran without the MAF and it just ran a lil richer, smoother sounding but the random hesitaion throughout the rpm range was still there and the mpg gauge still had the same oscillation. I shut it down, reconnected the MAF, started the car drove it up on ramps, shut it off, cleared the code and disconnected the O2 sensor. Per the manual the battery must be disconnected for several minutes to reset adaptations. So tomorrow morning I will reconnect the battery and give it some time to idle before I drive it to work. Hopefully it is the O2 but it looks like I'll need to do a fuel pressure and volume test. Other less likely options on the list are TPS, or loose ground. I've already ruled out the DME. D&G are my most closely matching columns. CO content too low adds the iacv to the list but it holds a great idle. I highly doubt that. Very well could be the filter. This car is unmolested. It still has the cover over the fuel filter, it's probably original. Also sat for 4 yrs before I got it, coulda gummed up the pump. It sounded wierd the first few seconds when I bypassed the relay. but it's completely silent now.

Sent from my Moto G (5) Plus using Tapatalk

[snow663]

Unplugging the O2 sensor made the oscillating mpg gauge and most of my hesitation to dissipate. It still has a flat spot coming off from a dead stop around 2k. Does this indicate a low fuel pressure/volume situation? The O2 sensor obviously works if it is creating a consistent 1hz oscillation.

Sent from my Moto G (5) Plus using Tapatalk

[ed323i]

The most logical explanation would be that the oxygen sensor needs replacing.
With the O2 sensor disconnected, and assuming that the adaptation values were succesfully reset, it runs with the default fuel map in the DME. The fact that it runs mostly fine now almost certainly excludes the fuel pressure from being a problem.

If it still has the original fuel filter, I would change that anyway.. The fuel filter would cause problems when the engine asks the most, i.e. between 4500-6500 rpm at WOT.

Perhaps there is also problem with the Vanos. You could disconnect the vanos solenoid and see how it runs then, w.r.t the flat spot at 2000 rpm.

I know the M52 oxygen sensors are very expensive (+/- 100 euro out here). I think the M50 ones are cheaper, but you'd have to look it up, using realoem.com .

Good luck!

[snow663]

It's hard to explain since you can't physically feel how my car drives but the "flat spot" may actually just be my tranny mounts. It moves a little between shifts. But after disconnecting the O2 my mpg got worse! By about 4 mpg. The only thing that improved is the car doesn't have an erratic accrlleration through the Rev range like before. Now it does not have any "hiccups" per say but it does still feel like I'm missing a good 30-40 horse. All the way through the Rev range. Struggles in 4th and 5th going uphill. It's weak. And honestly I think the O2 sensor is working. I have an oscilloscope so maybe I'll test it later after work. I've seen bad o2s before and this just don't quite fit. I didn't have a bouncy idle untill I unplugged it.

Sent from my Moto G (5) Plus using Tapatalk

[ed323i]

Connecting your BMW to INPA would be great. With O2 sensor connected you can monitor the adaptation values, and this can help you diagnose the problem.

Assuming the O2 sensors are OK (which I really don't know, and have a feeling they are not), the big swings in the mpg-gauge mean that the O2 sensors at some point say: "Wow, this is WAY too rich", and it keeps saying that for quite some DME cycles, and then it reverses to "Wow, this is WAY too lean", and also keeps saying that for many cycles. And then it repeats.

Perhaps poor fuel delivery can cause that, but you'd expect it to stay at "This is way too lean", of if the fuel pressure is too high, at "This is way too rich". It shouldn't cause swings.. Unless the pump is sometimes running fine and other moments it's not?

So, O2 sensor remains a good candidate, especially if it's old. They tend to become slow at reacting to a lean/rich mixture at old age. But, indeed, a scope can show if it's working more or less OK. It should go rich/lean all the time, several times per second. It should react very quickly to changes in the AFR.

Update: Correction.. The oxygen sensor should shift from rich to lean once per second, so 1Hz.. Not several times per second..

[E36328Coupe]

Easiest to try first is just disconnect the MAF connector. If it runs fine then, then you'll be sure it's the MAF.

No you won't because it could also be an air leak and not the MAF. Then you would be buying a MAF you do not need.


First port of call should be diagnostics, read fault codes or flash codes if available, though sometimes you have to take them with a pinch of salt

[snow663]

I'm gonna have to order the com cable for INPA I have an old desktop PC with serial port and the 20 pin connector but I'm missing the chip that actually communicates with the car. I can actually scope the sensor while it's plugged in and watch it. And then unplug it and "bench" test it. It's a zirconia sensor. Generates 0-1volt 1hz signal when new and in a correctly tuned engine. You may be right but I went through this with my last car and replaced everything, it never did get fixed. I'm pretty cautious of buying parts that I don't absolutely need. And now I know better. I can get an o2 pretty cheap, This particular Bosch sensor is extremely common. I broke like 4 of em on my other car over 6 years. But before I condemn it I would like to see what it's telling the DME. Maybe I'll test the new one too and put the data side by side so everyone else can see too. I'm sure if the sensor is botched the scope will show a slower than 1hz cycle.

Sent from my Moto G (5) Plus using Tapatalk

[E36328Coupe]

If you can solder build the "almost free ads" interface
And look for mistreku.zip if it's motronic. This works with the single vanos M50

[ed323i]

No you won't because it could also be an air leak and not the MAF. Then you would be buying a MAF you do not need.

True.. "you'll be sure it's the MAF" is a bit too strong.. But in this case, with the MPG gauge swinging a lot, and the car idling OK, I doubt it's a vacuum leak. There still may be a vacuum leak, but I don't see how it would change injection times that much, like 10-20% or more, and swinging hence and forth, and at higher loads (where vacuum leaks have less of an impact because of low manifold vacuum compared to low loads = high manifold vacuum).

I agree that INPA would be the best next step, especially if you want to make sure you don't buy parts that in the end didn't need replacing.

[snow663]

I'll order a tinyADS interface this week and we will find out for sure. I don't have the resources to build one our radioshacks are long gone!

I had another odd glitch today that threw me off. I was accellerating from coming out of a sharp turn and at about 4000 rpm it felt like I had a single dead miss. Just a tiny jerk and I couldn't replicate it. Possible symptom of running without an o2?

Woah.. update... At idle it's kinda shaking. The rpms aren't changing but you can clearly hear a pattern of ups and downs in the tone of the engine. I thought it was hunting till I opened the door and seen the tach. Mind you I still have the O2 UNPLUGGED

Sent from my Moto G (5) Plus using Tapatalk

[E36328Coupe]

Is yours the motronic 3.3 single vanos 325? If so you may get stomp test flash codes while waiting (even if you need to use a test lamp)

INPA won't get you anything on a 325 with motronic without the zip I mentioned

[snow663]

Is yours the motronic 3.3 single vanos 325? If so you may get stomp test flash codes while waiting (even if you need to use a test lamp)

INPA won't get you anything on a 325 with motronic without the zip I mentioned

Okay I ordered the TinyADS from openlabs. 3 day priority but it will probably be 5 days cause of the weekend. This .zip file, what does it do and why do I need it for DME 3.3.1? Is it a patch or something? I'll have to boot up the old guy and see what version of inpa I have installed. Maybe do some searching on the forums about INPA I was unsuccessful at getting it to work with my last car. But it coulda just been me. Now I can thank that old turd for educating me enough to fix this one right!

Sent from my Moto G (5) Plus using Tapatalk

[snow663]

I installed a new radiator cap and Immediately noticed the car running better. I can't get the o2 sensor out though and now I'm having bad o2 symptoms. Accelerates fairly normal but it hunts at idle and sometime jerks taking off. Code 1221 I have an o2 sensor wrench but it's slipping on the O2 sensor. Also the sensor is on top of a y pipe under the driveshaft in between the Y I'd have to drop the entire exhaust to put a socket on it.

Sent from my Moto G (5) Plus using Tapatalk

[E36328Coupe]

I installed a new radiator cap and Immediately noticed the car running better.

Placebo

I can't get the o2 sensor out though and now I'm having bad o2 symptoms. Accelerates fairly normal but it hunts at idle and sometime jerks taking off. Code 1221 I have an o2 sensor wrench but it's slipping on the O2 sensor. Also the sensor is on top of a y pipe under the driveshaft in between the Y I'd have to drop the entire exhaust to put a socket on it.

Sent from my Moto G (5) Plus using Tapatalk

lambda04.jpg

[snow663]

Sorry but no. It's seriously different with the cooling system under pressure. This is something people need to know. Once I put the cap on the engine took a few minutes to adapt but it started running smoothly. It's stopped spitting at idle and throttle responds twice as fast. I think it was reading hot and making the ECU go lean. Once I did that I plugged the O2 back in and now I have a code and light for it. And now the car behaves how I'd expect it to with a bad O2 sensor. Same thing with my old car. I chased a problem for a year on it. Had all bimmerforums best giving me feedback and they all got stumped. Till one day a plastic water hose barb broke and I lost all my coolant. I replaced it with metal and refilled the system and it ran fine ever since. I think the cooling system is far more sensitive than you think. Any air, leaks, whether it's leaking coolant noticeably or not. I have grounds to believe it is a problem. Once I get my inpa up and running I'll do some lab tests if I have to to prove it.

Sent from my Moto G (5) Plus using Tapatalk

[ed323i]

Placebo

I believe that you see the difference, and believe it to be real, but to me it's very difficult to understand why that change would be there. There is a chance it's placebo.. Perhaps it would be interesting to retry the test. Try the bad cap, then the good cap, and vice versa, a couple of times and see if you can reproduce the changes every single time.

When my radiator cap went bad on my E30 318i (1984, M10 originally) in which I installed an E36 M43 engine (nice combination, low weight, high torque, very economic), it didn't hold any pressure, and a few moments after turning off the engine, the coolant would boil, because of the remaining heat in the head, and the water standing still there (water pump not moving), at no pressure (so little over 100 degrees C boiling point).. It took a while for me to diagnose the problem, and then to order and receive the new radiator cap, and, during that time, besides the inconvenience of losing coolant I didn't notice any difference in idle or HP or torque or smoothness whatsoever.

So, probably there is another problem going on, and perhaps, due to some circumstances unknown to me, the problem gets activated when there is low pressure in the coolant?
In itself, the DME and engine don't give a damn whether or not the coolant is pressurized (yeah, well, the engine would prefer it to be pressurized so it knows its head won't crack that easily)..

[spyderg0d]

Forgive me, I didn't read all the posts above. From my experience, a mpg gauge weeping indicates that the 02 sensor is not detecting properly (won't throw a code until it's values maximise for a period of 10 seconds or if the heater element is not working properly within The sensor). Or there is a exhaust leak large enough to make the properly reading 02 sensor find the missing exhaust gas that its assuming is there and drastically lean or riches the mixture, causing horrible performance. If the 02 sensor was grounded or shorted there would be a code stored undoubtedly. It wouldn't be a bad maf, as is when a maf goes, the dme won't sweep the afr to compensate, as it won't read alot, then alittle, then alot, ect. Bad mafs just read consistently bad.

[snow663]

Forgive me, I didn't read all the posts above. From my experience, a mpg gauge weeping indicates that the 02 sensor is not detecting properly (won't throw a code until it's values maximise for a period of 10 seconds or if the heater element is not working properly within The sensor). Or there is a exhaust leak large enough to make the properly reading 02 sensor find the missing exhaust gas that its assuming is there and drastically lean or riches the mixture, causing horrible performance. If the 02 sensor was grounded or shorted there would be a code stored undoubtedly. It wouldn't be a bad maf, as is when a maf goes, the dme won't sweep the afr to compensate, as it won't read alot, then alittle, then alot, ect. Bad mafs just read consistently bad.

It used to sweep under load, hesitate a lil at launch, no codes or CEL unplugging the O2 sensor got rid of the sweeping gauge and acceleration that comes and goes throughout the Rev range but it still didn't feel right. I replaced the rad cap cause they had it on-hand and I could visually see it and hear it leaking with compressed air being fed to the rad. Now that the cooling system holds pressure it gained some power, the throttle response is "tighter" and sounds more like a fuel injected engine. plugged the O2 back in and now I have a code 1221 and it displays the symptoms of a bad O2 more clearly. I have a new sensor but I'm having a hell of a time getting the old one out. My O2 sensor wrench is expanding at the slot for the wire allowing it to slip on the hex head of the sensor body. My actual wrenches are too large to fit in the tunnel of the car feasably. And no chance at a real socket without dropping the exhaust. My TinyADS interface is in the mail but when it gets here I'm gonna run the car for two identical runs. One with the new cap and one with the old and I'm gonna see what that leaky cap was doing to the engine management system. No more speculation. Just real sience.

PS I gained back about 7 mpg avg with the CEL on from the bad O2. And the new rad cap. The only negative thing I have to say about it now is the idle hunts up and down forever. Typical bad O2 sensor symptom. I'm at 26.4 mpg avg over 25 miles. Where two days ago it was 19.5 avg.

Sent from my Moto G (5) Plus using Tapatalk

[jc43089]

Hunting idle is also commonly caused by vacuum leaks or a sticky idle control valve. A vacuum leak will have similar symptoms to a bad o2 sensor because it is altering the mixture the ECU is trying to control. If your o2 wrench is bending try a higher quality one? Or lower the exhaust and use a socket. Another option is to cut the wire off and use a box wrench.

[ed323i]

I also just learned that running with both O2 sensors disconnected, at least with the MS41 DME (M52), the timing gets retarded quite a bit, resulting in quite a performance hit (very, very noticeable).. It's probably the same with the M50, if you disconnect the single O2 sensor. So, with sub-optimal ignition timing, fuel economy will also suffer.