Intermittent Power Loss / Hesitation / Lean Condition

[mromero]

Alright E30 experts, I need some help trying to diagnose an intermittent engine hesitation/power loss/lean condition on my '87 325i (02/87 build date).

The issue: It will have extremely poor acceleration / lean condition such that it makes driving near impossible; however, its intermittent and doesn't seem to have a common time/condition when it occurs. I.e. can occur after 15-20 min of driving or immediately on start-up. It seems to only occur during partial throttle. At WOT it seems to clear up and revs fine and at extremely low throttle its able to slowly rev cleanly to redline.

One curve-ball with the car is a PO replaced the AFM with a MAF conversion and piggy-back signal box. See link below. Due to the intermittent nature, I don't think its a mapping issue on the box. I've checked all after-market wires/splices and they appeared to have been soldered/installed well. As well as checked MAF signal voltage into the signal box and verifies that its working/changes voltage w/ increased rpm. https://splitsec.com/product/psc1-00...nal-calibrator

Heres everything I've replaced so far:

• Fuel pump, transfer pump, and fuel filter
• Fuel pressure regulator
• Fuel injectors
• Main relay and fuel pump relay
• New battery and alternator (previous ones were dead/dying)
• Ignition coil
• MAF sensor
• O2 sensor
• Coolant sensor
• All vacuum hoses

Items that I've tested and/or noticed:

• CPS is within spec at 540ohms
• TPS continuity check for both idle and WOT
• Plumbed a fuel pressure gauge at the fuel rail and verified that it does not lose pressure during occurrences
• I inspected the fuel tank while changing the transfer pump and it is quite clean / no sludge or rust
• Checked/cleaned grounds at the oil pan, distribution block, battery, and strut locations
• Stomp test yields AFM code, 1215
• I removed the o2 sensor, to see if it would improve if I had a clogged cat, with no discernible difference

[82eye]

i think you know where the issue is. the maf converter is programmable so you'll need to hook up a laptop and see what it is doing and possibly recalibrate it. i notice the mfgr has stated they are discontinuing that product so i'd get on it sooner rather than later.

your other option is revert to afm.

edit : code 1215 is bad afm signal.

[mromero]

i think you know where the issue is. the maf converter is programmable so you'll need to hook up a laptop and see what it is doing and possibly recalibrate it. i notice the mfgr has stated they are discontinuing that product so i'd get on it sooner rather than later.

your other option is revert to afm.

edit : code 1215 is bad afm signal.

Yeah, I haven't been looking forward to diving into that rabbit hole yet. But I managed to find the software online, so I'll go ahead and trying plugging in this afternoon.

[82eye]

Yeah, I haven't been looking forward to diving into that rabbit hole yet. But I managed to find the software online, so I'll go ahead and trying plugging in this afternoon.

e30 maf conversions are getting hard to support now. miller has gone dark, and sssquid has dropped them for now. they seem to go digitally dirty over time and guys can't fix them. our ecu's were never that comfy with them.

some of the MS app plug n play maf conversions are still ok but you'll find guys eventually having issues with even those. them and maybe speeduino(?) might be the best now. cleaner with a standalone ecu.

edit: rock auto has afms now if you need to reverse it on the cheap. bavrest is sort of the recommended go to

[mromero]

Of course when I went to tinker today, the problem wouldn't show itself ha. However, I was able to get plugged in and the box was working as it was supposed to which I was able to verify by comparing the MAF sensor voltage to the conditioned signal voltage. There was a couple of simple functions in the software to send out discrete voltages and when I sent a 0v/no signal it replicated the issue exactly. So yeah, its likely the box is intermittently having issues which should be easy to verify by keeping the multimeter close and checking the conditioned signal voltage going into the ecu.

e30 maf conversions are getting hard to support now. miller has gone dark, and sssquid has dropped them for now. they seem to go digitally dirty over time and guys can't fix them. our ecu's were never that comfy with them.

some of the MS app plug n play maf conversions are still ok but you'll find guys eventually having issues with even those. them and maybe speeduino(?) might be the best now. cleaner with a standalone ecu.

edit: rock auto has afms now if you need to reverse it on the cheap. bavrest is sort of the recommended go to

I appreciate the insight. The speeduino project seems pretty sweet; I was literally thinking earlier that it would be pretty simple to get an arduino rigged up in place of the current box since its just a voltage offset look-up table. But having it as an entire stand-alone is definitely a few steps above that. I'll definitely be looking more into that.

[mromero]

Quick update:

I was able confirm that it is in fact the signal box that is the issue. The issue occurred again today and there was no output from the box. I did call the manufacturer and I can send it in to them for bench testing / possible repair; however, in the interim, I was able to get an arduino rigged up and get the car running. The only issue is that the board I'm using only has a PWM output which I dont think the ECU likes since its quite choppy; but I will be able to try a bit newer board tomorrow with an actual DAC output which I think will help.

If that does work, that should be good enough for me to cruise around with this summer.

[hasa]

What is the motivation for a MAF conversion?

[mromero]

What is the motivation for a MAF conversion?

A previous owner had it done. And right now, it seems to be slightly less work to keep that running than switch back to an AFM setup; though, this may be prove to be the wrong decision when looking back with hindsight.

[hasa]

I'd personally go for standalone aftermarket engine management in our situation. Unreliable black box style MAF conversion doesn't seem comfortable for customer when something goes wrong with the system.

[mromero]

I'd personally go for standalone aftermarket engine management in our situation. Unreliable black box style MAF conversion doesn't seem comfortable for customer when something goes wrong with the system.

I certainly do not disagree; however, if I'm able to get the arduino to work well, then I'm the one controlling the black box and all internal functions. The hope is that this will be sufficient for the summer / until I decide to proceed with a planned S50/52 swap that will most likely be on a standalone system.

An overall intent of: dont get it perfect, get it running.

[hasa]

How does the MAF conversion actually work? As far I remember from my Jetronic times at 90's, the flap type meter outputs flap angle (flow) and intake air temp to the ECU. But MAF is just air mass without any temperature signal (as it is not needed).

[mromero]

I'm definitely not an expert on the sensors; but from a laymen: the MAFs are just a thin wire that can calculate the air flow based on how much power is required to keep it at a constant temperature. Which are inherently a bit better than the AFM style since it doesn't have a flap impeding flow as well as they are more responsive.

In terms of the conversion on these cars, the new MAF signal is sent to a box and is 'conditioned' to be a signal that the ECU expects. Although, AFM and MAF are typically both 0-5v signals, the curves ( voltage vs airflow) are different and thus if you want an identical behavior there needs to be some signal conditioning. In my case, all the black box does is read the MAF sensor and offset the voltage to what the stock AFM value would be. One 'benefit' is that you can 'tune' this to give more/less airflow signal to the ECU which will obviously compensate for fuel/injector pulse. The box that the PO put in my car has a map, similar to a VE map, of RPM vs MAF Voltage that you can define the offset voltage value that you want.

So at the end of the day, its just an expensive voltage lookup table. Which is why now-a days you can accomplish the same thing with a 30$ microprocessor board and ~50 lines of code.