Still stumbling off idle.

[gtmdriver]

I have just finished converting my 1600cc air cooled VW Beetle engine to EFI using a MS2 V3 with port injection but a single throttle body.

I have driven the car a number of times and VE analyse live has produced a workable tune but one definite problem still remains. While the engine is idling a gentle opening of the throttle causes the engine to stumble and almost die. This same problem also causes the engine to misfire when accelerating from a closed throttle at low engine speeds.

I am using an IAC valve to control the idle and it seems the problem is in the transition from the IACV to the main throttle butterfly.

I have logged the problem and the AFR spikes high when the throttle is slowly opened. To try to prevent this I lowered the tps.dot threshold so that the acceleration enrichment is triggered at very slow throttle openings but this has had little or no effect on the problem.

I would be grateful if someone could have a look at my tune and log files to see if I am missing something obvious.

John

[142 guy]

I only looked at your log, not your tune.

First off, being air cooled what are you using as your proxy for coolant temperature, oil temperature? Coolant temperature does not have a huge effect on fuel; but, yours is lurching around like a sailor on shore leave and that should not be happening.

What is your req fuel? Your injector PW is swinging like crazy in part because your accel PW seems to be really huge. If you look at the start of your log, at record 28 your PW swings from 2.7 to 6.7 ms. Unless your injectors are tiny, that is a huge swing in PW. Shortly after the injectors open the flood gate (record 30) your RPM takes a dive. So the fuel came first and then the RPM took a nose dive. Although it doesn't show in the AFR, unless you have a really high Req fuel, my reaction is that you are applying way to much accel enrichment which is killing the motor. I would be inclined to completely eliminate acceleration enrichments. Most engines will work satisfactorily with out acceleration enrichments although the transition during acceleration. If you can get it to run satisfactorily, and it is hesitating on acceleration, then try adding in small amounts of acceleration enrichment.

[slow_hemi6]

Agree probably going too rich but to compound that, you simply have the totally opposite spark behavior to what you need. When ever you throttle on it rips spark out. First one goes from 15 degrees to 8.7, second small dip goes from 11.2 to 9.1 before it recovers. It continues on like that, Every rpm nose dive has a timing nose dive with it. You need to be adding 5 or 10 degrees in those areas to get the engine to respond.

[gtmdriver]

Thank you both for that.

My initial estimate was that the old VW engine would need a lot more fuel than a modern engine so I upped the AE.

I am rapidly coming round to the idea that I was very wrong and, as you suggested, have backed the AE off to a very low value.

I have not yet driven it to test the new settings so I will report back when I have.

I am using an OEM VW cylinder head temperature sensor as fitted to the very last injected Mexican Beetles. It fits into a pre-tapped boss on the later heads.



I calibrated it very carefully and entered the calibrations into the controller.

Under normal running it seems to stabilise at 110 - 115C


As regards the ignition timing, in the absence of any other information I set up the timing table as best I could to replicate the OEM distributor.



It had 22 degrees of centrifugal advance and 6-8 degrees of vac advance.

It is timed at 7 degrees btdc at idle with no vac signal but as soon as it's running and the vac signal is included this rises to around 15 degrees btdc at idle. When I press the pedal this kills the vacuum and the timing falls back, which is how it would have behaved with the original distributor.

Is this wrong?

[slow_hemi6]

The area I have highlighted here is really only ever accessed in a bog, heading towards a stall, type situation. This is where you need extra timing to try and pull it out of that event. I would suggest 2 degrees above your idle timing in the cell just above idle and work up to about 4 degrees above idle timing at 75kpa. That area appears to be used when starting but, when cranking and starting the values given for cranking advance is used till the cranking rpm is exceeded. By that time MAP should be getting close to idle value anyway, so you might get a moment of 2 degrees over idle timing going into idle but that won't even be noticeable.

example.jpg

[gtmdriver]

I've had a look at the tables.

This is the timing when it idles steadily



and this is the timing when it starts to stumble



So I've increased the advance to 12.8 at 45kPa (+2) and 12.5 at 70kPa (+3.7) then interpolated between these values.

I also interpolated between 70kPa and 100kPa.

Is that what you suggested?

[142 guy]

Thank you both for that.

My initial estimate was that the old VW engine would need a lot more fuel than a modern engine so I upped the AE.

I am rapidly coming round to the idea that I was very wrong and, as you suggested, have backed the AE off to a very low value.

I have not yet driven it to test the new settings so I will report back when I have.

I am using an OEM VW cylinder head temperature sensor as fitted to the very last injected Mexican Beetles. It fits into a pre-tapped boss on the later heads.



I calibrated it very carefully and entered the calibrations into the controller.

Under normal running it seems to stabilise at 110 - 115C

I think most first time users of MS (me included) are over enthusiastic on the application of acceleration enrichment. As a word of advice, get your Ve table set first as AE is a % adder based on the base fuel pulse width. If you set your AE values and then go back and modify your Ve values you effectively modify your AE values.

If you are using the OEM spot for a temperature sensor that the original D jet used, that should be reasonable. The arrangement seems correct and you should not be getting the coolant temperature fluctuations that show up in your log. Your sensor may be calibrated correctly; but, based upon the speed of the temperature swings (which should be impossible with the arrangement you have) I would suggest that your sensor may be faulty or that you have a bad connection somewhere. Since warm up enrichment and ASE are coolant temperature based, these coolant temp swings might cause them to fluctuate which would make life interesting. Its not a disaster; but, something to put on your to-do list to be checked out

[gtmdriver]

I hadn't noticed that. Thanks for the heads up.

It's a modern sensor as used on the very last Mexican injected Beetles.

When I calibrated it off the car the resistance readings were rock steady.

I'll check out the connectors but there doesn't seem to be any break in the trace which would indicate an open circuit. The reading just swings about wildly.

It looks more like a faulty sensor.

[slow_hemi6]

12.5 at 70kPa (+3.7)

Don't ramp down from 12.8 to 12.5 as the bog gets worse. Make it at least flat if not slightly increasing as kpa falls.
You seem stuck on the logic of the dissy curve is best. However AFAIK the VW dissy is fed from ported vacuum and including vac advance figures to idle advance is actually incorrect to start with. You have a whole map at your disposal, why be constrained by the limitations of a distributor. The engine likes 10.8 but in reality it would not have had that at idle from factory.

[gtmdriver]

Thanks again. I've amended that.

I understand what you're saying re the advance curve but, as a complete novice, I needed somewhere to start the tune and, in the absence of any other data (although I did appeal ), I went with what I thought was a safe option with a facsimile of the original distributor.

Once I get the fuelling somewhere near it's going on a rolling road which should sort it once and for all.

[LAV1000]

Your temp sensor is a one wire type, check out the groundpath to ms unit.

[gtmdriver]

Definitely. It's my first job.

[gtmdriver]

I've just had a look at the log again and the battery voltage is fluctuating between 12.2v and 12.5 v.

This fluctuation seems to be affecting the dwell and possibly the coolant temperature reading.

The VW Beetle alternator is obviously quite an old design. Is there any way of stabilising the supply voltage to the Megasquirt unit?

[gtmdriver]

I have been out and checked the wiring and it is sound.

There is 0.2 ohm resistance between the body of the temperature sensor and the Ground point on the MS relay board so I don't think that is my problem.

When the engine is not running the temperature reading is rock steady even if I move the wiring and connectors around but with the engine running the temperature fluctuates as does the Batt voltage.

I also noticed that the batt voltage reading is significantly lower than the actual supply voltage by about 1.8v. I noticed this when I was running the JimStim too. I can get the reading correct by adjusting the voltage calibrations in the Tools section but this is clearly not the way to go as that's not the purpose of this adjustment. I am powering the MS, the injectors and the coil from the same source so the voltage calibration should be fine.

I suspect that the apparent temperature fluctuations are actually the result of the old school voltage regulation found in the Beetle alternator. This may also explain why the fluctuations are more pronounced when the engine is hot as the resistance of the sender falls from around 3k ohms at 20C to less than 200 ohms at 100C so the voltage fluctuations will have a greater effect on the current flowing through the hot sensor when the resistance is low.

So to sort this out I need a stabilised 12v supply to the MS unit.

[LAV1000]

On grounding check:
http://www.msextra.com/doc/pdf/html/MS2 ... e-3.4.html
Chapter 3.2 !

As far as the alternator and regulator, Injector opening time and dwell are (battery) voltage related.
If you think those are messing up, run a log with alternator and rectifier disconnected and see what happens.

All sensor are using the same 5 Volt witch is a controller inside the ms unit.
All the ms units are automotive spec. so they designed to withstand some voltage drop.
I thought min batteryvoltage is around 9 volt or a little higher before the system drops out, 5V controller quits.

I think for the tempsensor it is still a grounding isue.

[gtmdriver]

I can't see another option for the sensor.

It is grounded to the engine head/block as it is bolted into it and the MS unit is also grounded to the engine block.

As I said the temperature reading is rock steady when the engine is not running but fluctuates when it is.

[subwoofer]

Proximity to spark leads? Are you running resistor leads and plugs? Copper plug leads is asking for trouble.

[142 guy]

I have been out and checked the wiring and it is sound.

There is 0.2 ohm resistance between the body of the temperature sensor and the Ground point on the MS relay board so I don't think that is my problem.

When the engine is not running the temperature reading is rock steady even if I move the wiring and connectors around but with the engine running the temperature fluctuates as does the Batt voltage.

I also noticed that the batt voltage reading is significantly lower than the actual supply voltage by about 1.8v. I noticed this when I was running the JimStim too. I can get the reading correct by adjusting the voltage calibrations in the Tools section but this is clearly not the way to go as that's not the purpose of this adjustment. I am powering the MS, the injectors and the coil from the same source so the voltage calibration should be fine.

I suspect that the apparent temperature fluctuations are actually the result of the old school voltage regulation found in the Beetle alternator. This may also explain why the fluctuations are more pronounced when the engine is hot as the resistance of the sender falls from around 3k ohms at 20C to less than 200 ohms at 100C so the voltage fluctuations will have a greater effect on the current flowing through the hot sensor when the resistance is low.

So to sort this out I need a stabilised 12v supply to the MS unit.

The MS unit is supplied off of the vehicle 12 v system; but, all the sensors and the board electronics run at 5 v. There is an integrated circuit voltage regulator (its mounted on the very left end of the V3 heat sink) which drops the 12 v to 5 v on the board. The external sensors are supplied off of this regulated 5 volt system. The 12 v car voltage can bounce around an awful lot and the 5 v board level voltage will generally remain close to rock solid. Unfortunately, I don't think you can log the internal 5 v bus voltage to confirm that it is OK. If desperate, there are ways that you can check the operation of the internal 5 v regulator; but, I expect that this is likely not an issue (this check would normally be done during the build process). In general, the MS board level electronics and sensors will be very tolerant of fluctuations in the vehicle supply voltage and I don't think this is what is causing the temperature fluctuations. As a note, injectors and ignition coils are not so tolerant of changes in the 12v supply and may result in erratic operation if their voltage sensitivity curves are not correct.

Short of finding an alternate sensor, you are pretty much stuck with the ground return. It might help to apply something like Ox guard to the threads of the sensor to improve its ground connection (Ox gaurd and other products are conducting greases / pastes to improve electrical conductivity - don't use dielectric grease which does the opposite. Any thing that cycles on and off and induces a current flow in the ground connection will generate a voltage offset that shows up and corrupts the CLT signal. Do you have anything that is cycling on and off when these temperature fluctuations occur (I know its air cooled - but have you substituted or added an electric fan by some chance?)?

edit - just a thought. I don't know how your alternator is mounted and how big it is; but, is it possible that the current through the ground connection on the alternator is what is causing the CLT temperature fluctuation? At 100C, the curve of the sensor is getting closer to horizontal making it more sensitive to small offsets. Its would be an easy check to disconnect the alternator and run the engine for a minute or so to see if the fluctuations disappear - reaching for straws here!

You said that this engine was off of a last generation air cooled bug. Was VW still using D jet on the last generation or had they switched to later LH jet? If it was Djet, your characteristic curve looks close to the curve on the D jet coolant sensor on my Volvo B20E. It is a two terminal solid element sensor. Something like that might provide a solution; but, only if it turns out that ground offsets are creating your temperature fluctuations.

[gtmdriver]

Proximity to spark leads? Are you running resistor leads and plugs? Copper plug leads is asking for trouble.

Magnecor leads and NGK resistor plugs.

[gtmdriver]

The MS unit is supplied off of the vehicle 12 v system; but, all the sensors and the board electronics run at 5 v. There is an integrated circuit voltage regulator (its mounted on the very left end of the V3 heat sink) which drops the 12 v to 5 v on the board. The external sensors are supplied off of this regulated 5 volt system. The 12 v car voltage can bounce around an awful lot and the 5 v board level voltage will generally remain close to rock solid. Unfortunately, I don't think you can log the internal 5 v bus voltage to confirm that it is OK. If desperate, there are ways that you can check the operation of the internal 5 v regulator; but, I expect that this is likely not an issue (this check would normally be done during the build process). In general, the MS board level electronics and sensors will be very tolerant of fluctuations in the vehicle supply voltage and I don't think this is what is causing the temperature fluctuations. As a note, injectors and ignition coils are not so tolerant of changes in the 12v supply and may result in erratic operation if their voltage sensitivity curves are not correct.

Short of finding an alternate sensor, you are pretty much stuck with the ground return. It might help to apply something like Ox guard to the threads of the sensor to improve its ground connection (Ox gaurd and other products are conducting greases / pastes to improve electrical conductivity - don't use dielectric grease which does the opposite. Any thing that cycles on and off and induces a current flow in the ground connection will generate a voltage offset that shows up and corrupts the CLT signal. Do you have anything that is cycling on and off when these temperature fluctuations occur (I know its air cooled - but have you substituted or added an electric fan by some chance?)?

You said that this engine was off of a last generation air cooled bug. Was VW still using D jet on the last generation or had they switched to later LH jet? If it was Djet, your characteristic curve looks close to the curve on the D jet coolant sensor on my Volvo B20E. It is a two terminal solid element sensor. Something like that might provide a solution; but, only if it turns out that ground offsets are creating your temperature fluctuations.

This is my sensor



and this is the log of my batt voltage and my 'coolant temperature'.



As you can see the fluctuations are quite rapid.

Tomorrow I will try earthing the sensor body direct to the MS unit so see if that cures the problem.