Automotive

Automotive Diagnosis and Repair

 

There have been several discussions recently about how one pays up for automotive repair work, without necessarily getting the problem solved. In my contribution to that thread, I offered a strategy of diagnostics which was indoctrinated into me by General Motors (who for all their faults do actually know some things about cars and their repairs):

Fast forward 18 months. The problem reoccurred last November, but this time it was more intermittent. Run great, stumble and miss. Run like a champ, chugga-chugga, run great. Arggh. Of course, to add to my irritation, I had a client in town from the Netherlands and we were out for a drive in excellent weather - 50 miles from home. Again we limped back and I parked the car for a couple months until I could get time to tackle the irritation.

This particular generation of the Elan (Lotus being owned by GM at the time) uses a GM engine management system parked atop an Isuzu Impulse Turbo block and heads, with all sorts of Lotus customization in-between. The only interrogation you can do to the computer is via a Vetronics Tech1 diagnostic tool enabled by a security key controlled by Lotus itself. Or a laptop-based system using serial cables and a Windows 95 program. Or a paper clip connected between terminals which then puts the ECU into diagnostic mode so it can flash a lamp with codes to direct the technician. 

Only 300-350 cars came to the USA, and there are no competent Lotus Elan dealers nearby. There are specialists in the older Elans from the 1960-70's, and there is a main dealer who presumably knows the Elise and so on. But the task of repair for this car falls on me. 

Luckily it dates from a time when engine computers were relatively simple, and I was one of the authors of the authoritative service manual for engine diagnostics. Which is not to say that I am a "real" mechanic with all the tools.

BASIC DIAGNOSIS

We begin with the basics. 

1. I confirmed the fault with the owner, myself. Yup, it's missing on 2.

2. The car looks ok under the hood

3. No bulletins or recalls, so I checked the Lotus Central repair archives. Plenty of material on missing and stumbling due to fuel and ignition

4. My scan tool can't talk to this car, so I inserted the paper clip. It gave me consistent code 12 (I'm ready to deliver you helpful hints but haven't got any) but none of its other codes, so I knew those particular components were likely OK.

In order to run, an internal combustion engine needs compression, ignition and fuel - all supplied in the right order and timing. Because the problem was intermittent, at times I had everything. But I still checked the compression by inserting a pressure gauge into the spark plug holes.

COMPRESSION

Lotus compression check with 5 cranks; cold 

#1 - 136 psi

#2 - 132 psi

#3 - 130 psi

#4 - 140 psi

That's relatively consistent and adequate for our purposes - I ruled out a mechanical problem (valves, pistons, etc.). Progress, but no improvement. 

NOTE: At this point we have to discuss this motor's particular engine management system. Essentially, my 1.6L DOHC Isuzu motor is treated as two 2-cylinder engines. If you are conversant in 4-stroke cycle engines, you know the 4 parts of the cycle are intake, compression, combustion, exhaust. The spark has to come at the right time to ignite the compressed air/fuel mixture. 

But it doesn't matter if an extra spark occurs during the exhaust cycle. So my car, and many others that lack a distributor, uses half as many computer circuits to fire half as many ignition coils twice as often. This saves processing cycles, manufacturing costs and complicates our diagnosis. There are two 2-ended coils that fire a pair of cylinders each time. The cylinders are paired up with 1 & 4 and 2 & 3. 

IGNITION The ignition coils are common between cylinders - 1&4 and 2&3

I pulled each plug wire in turn as the engine was running. There were sparks every time coming out for every cylinder, but the engine didn't slow down when plugs 1 or 3 were pulled. Those were the dead cylinders. But not always dead - remember this was an intermittent problem. 

I found no visual evidence of damage, rust, water incursion, etc. But on the web I found lots of stories about coil failures. The cost was relatively low ($80) and the access almost unimpeded (on the web photos) so I bought a pair of coils and set about changing them. (photo from Lotus Elan Central)

Silly me -- I soon discovered that the coils on a UK car don't hide under a brake master cylinder, above a steering column, and next to A/C pipes. My coils were a swine to remove and replace -- follow the red ignition wires to the coils ... it only takes an hour upside down with hands jammed into sharp corners and wire ends.

Net result - NO CHANGE. (insert image of technician ranting, raving, stomping around the garage banging head on walls, etc.) 

Let's move on to the Fuel System.

NOTE: The fuel injection system takes the same approach as the ignition. You can use smaller injectors if you pulse them twice during the intake cycle. And if you place them upstream of the cylinders in the intake manifold, you can use multiple pulses to provide a more consistent fuel mixture. Thus my car, and many others, uses half as many computer circuits to fire smaller injectors twice as often. This saves CPU cycles, electronic components, and complicates our diagnosis a bit.

There are two driver circuits in the ECU which pulse a pair of injectors each time. The cylinders are paired 2 & 4 then 1 & 3. THE OPPOSITE PAIRING compared to the ignition system.

FUEL The fuel injection circuits are common between cylinders - 1&3 and 2&4

It's relatively difficult to interrupt the fuel injection, compared to ignition. In theory you can pull off the injector wire connector, but they are well hidden under the manifold and impossible to touch (heat!) when the engine is warm. 

So I tried the indirect approach. I spent $40 on a tool called a NOID light set, which are LED bulbs in a special case with connectors to match any fuel injection harness. 

You connect the light to one injector plug, crank or start the engine, and watch for the bulb to flash. No flash means an open circuit; steady light means a short in the circuit. All the lights worked properly, so I knew the wiring (at times anyway) must be connected to the injectors.  I used the factory diagnostic circuit diagnosis flowchart to confirm my thinking.

I used my ohmmeter to check the continuity of each circuit from the fuel injector to the ECU engine management unit. All seemed OK. Then because I was almost certain there was some intermittent problem, I pulled apart the harness and re-soldered all of the splices that joined the injectors and ground connections in pairs. No change. But I did feel better because the splices were pretty lame.

That narrowed down my options to bad fuel injectors or bad driver circuits in the ECU. I'd already replaced all the injectors less than 1500 miles ago but finding them are much less difficult than finding a new ECU. So out they all came for visual inspection and resistance check. AHA! 

The number 3 cylinder injector was 1.6 or 1.7 ohms compared to spec of 1.8 to 2.2. ohms. The other 3 were 2.2 ohms. 

This pointed to a fuel injection reason for losing 1 and 3, not ignition. Why did we lose both cylinders? Because they are fired together, and if the signal is "lost" or shorted to ground, or blocked by high resistance in one injector, the other one will fail to fire as well. We only have a few milliseconds each time the injector is open, so they must be up to snuff.

A long distance phone call followed by an overnight cross-country flight resulted in a pair of injectors arriving from "Injector Warehouse" as a warranty exchange. Replacing number 3 solved the problem almost instantly. It did take about an hour to swap them. Release fuel pressure, remove fuel regulator, unbolt fuel rail, release locking clips, slide clips to side, pry up rail, pry out injectors, confirm new injector resistance, insert, etc. etc.

And nowadays, it's obligatory to take a selfie of the car

and the diagnostic mechanic

Over the space of a couple weeks I spent about 10-12 hours on this car, replacing only injectors (2) and coils (2). I used my existing $20,000 worth of tools, $500 of Lotus service literature, two Internet sites, and 30+ years of industry expertise. I bought $40 worth of tools and $80 of parts, and visited the local parts store 3 times plus mail order from Amazon and Injector Warehouse. 

I can see how the dealer would have charged me well over $1000 to accomplish the same. But I would expect him to work faster, for my car to work flawlessly when returned, and not to experience any other drivability faults. Alas, the guy fixing the car at the dealership may or may not have the same resources that I have had, or the patience ... or could he have done it in seconds because he knew the common faults, had access to piles of parts and the big diagnostic machine? That's the gamble we take.

Cheers,

Cazalea

PS -- This story has been simplified for the sake of our readers ...

More posts: EliseLotusTAG Heuer Connected

  login to reply