First job is to check if the ECU is providing any fault codes. As it sounds like the ECU is shutting down the engine it's more than likely.

Things worth checking
1) Ignition switch - these can get bad contacts.
2) Connection to air flow meter.
3) Make sure all fuses and relays are seated properly.

The ECU diagnostic book from Donutz will help greatly, as would the Veloce Enthusiast's Manual.

PLEASE let us know how it goes and DON'T be persuaded to spend big money on new ECU etc. until you're totally sure this is the fault. (5/00)

If you've swapped the AFM then it's either the wiring between the AFM and the ECU (probably a bad contact in the connector at the AFM end) or a bad joint at the ECU end. Really carefully check the AFM connector. Try pulling the connectors off the ECU and replacing them a few times.

If the problem persists consider -
1) running new wires in parallel to the existing ones and connecting them near AFM and ECU. If this fixes it then you have a loom fault.
2) Cutting off the AFM connector and splicing in a new one. Jay Ng should be able to snip one off a scrapped vehicle for you.
3) If it's an ECU fault then pull it out and trace the lines from the connector around the board. It's probably a dry joint where the main connector meets the board or a dry joint elsewhere.

Someone might lend you an ECU to test things and/or tide you over if necessary. (5/00)

We are [the garage who are] actually looking at this one, and I have to say we're a bit stumped!

So far we have swopped the AFM, and the ECU, and checked all wiring between the two for continuity. The ecu still reports the fault as code 10 (afm). All is fine when very cold, and the car will run for a couple of miles, then it just cuts. It will fire when cranking but then just dies. I suspect the water temperature sensor on the back of the engine, but don't understand why the ecu is saying its the afm at fault!

Also its a real pain to swop the sensor, so that will have to wait till the morning! If anyone has come across this kind of problem before, any help would be gratefully received! (5/00)

similar fault on a Citroen.. a crack in a pipe between AFM and throttle would only cause a serious leak when the plastic got hot enough to deform .. after 15 miles... same going lean and dieing engine..

.. but this one sounds hard .. should be able to just pull the water temp sensor wire .. and put it into limp home mode for test.. I only have a 1.8 manual ..so no pointer for you there... (5/00)

Also its a real pain to swop the sensor, so that will have to wait till the morning!

Can you just stick a meter on the wire at the ECU for this sensor and see if the reading gets out of line when the fault happens? I always like to see things being wrong before I try swapping them, but I don't have lots of spares lying around.

You could also meter the AFM signals - tape three meters to the dash and go for a drive! I suppose an air leak is possible. The ECU knows roughly what the AFM should read at different RPM and a leak will cause bad readings.

BTW Did you check all the AFM wiring including the temperature thermistor? (5/00)

How about forcing the fuel pump into operation with the jumper in the diag box (F/P to Ground)?

That will override the fuel cut off function of the air flow meter.

Don't leave it like this, though. The fuel pump will run whenever the ignition is switched on - and the cut-off is a safety feature!

The diagnostics only tell you if a sensor is open circuit or shorted, not if it's providing the wrong info. But failure of the water temp sensor shouldn't stop the engine anyway - in fact I think you could disconnect it and hardly notice. (5/00)

Thanks all for help and advice. The fault has now been fixed!

It turned out to be a dry joint in the ecu (quite why we had the same with a different ecu I don't know, but at the moment don't have time to investigate!) After having exhausted all tests in the manual, I took to tracing the circuits in the ecu, and hey presto, pin 2P had a dry joint. (5/00)

I am glad you found the problem.

Dry joints can be funny things - many a satelite/probe has failed besue of these. You can get continuty or not just by a few degress changes in temperature and/or stress.

As to why you have 2 ECU's with the same problem, well I suspect that there could be pins on the loom connector not exactly in line, so when the ECU is plugged and un-plugged, it stresses the PCB joint on the ECU mother board. If the joint already has an inherent weakness, then it fails.

As to why 2 ECUs with the same weakness - well if they are about the same age then it could be a "consistent" manufacturing defect. Most likely a setup problem on the flow soldering machine, which in my experience are a pig to get set up well for a particular pruduction run. (5/00)

Any garage owner will tell you - don't listen too much to the customer's explanation of how the fault happened - it'll probably be misleading. Cambelt is always worth checking on a non-runner.

Fault codes in the Grainger's 1.6 manual the same as for the 1.8 ?
Yep a few extra ones on the 1.8, but nothing that would be critical except 04 (SGC) 03 is SGT on a 1.8 (G on 1.6). Both are fatal. (Timing signals from camshaft position sensor)

AFM connection had occured to me; is it a similar connector to the 1.6 (ie. secured by a wire clip)?
Yep. Release wire clip, take it off, check for bent pins, and carefully replace fully home. Check all the fuses, too. (12/00)

I've been having problems with the engine cutting out on my Mk2 1.6, which has now done 28k miles. The problem started about three weeks ago, just driving at about 40mph the engine cut out. It restarted ok, but cut out a few times over the next few days. There was no pattern, cold or hot, the bugger just cut out when it felt like it, not very nice on the IDR through Reading in the morning crawl hour.

I took it to Mazda, who after taking it out on a 30mile test drive found it only cut out once. They changed the coil unit the following week (five day wait for part from Belgium). But upon driving to Oldham for the hols with cat and wife, it quit on the M1 just before Sheffield (doing 70 (ish) in the fast lane at the time).

Again it started ok, but when trying to accelerate it started cutting out again, bigtime. Top marks for Mazda Assistance, who arranged for a piggyback on wrecker truck to Oldham and then onto the Mazda dealer in Stockport the following day.

Unfortunately it wouldn't cut out for them, but they did change the plugs (should have been changed at 27k service I reckon), and also changed the fuel pump relay just in case. Well, fingers crossed it hasn't cut out since, (over 300 mile later). Has anyone had similar problems ? (1/01)

only similar was with a load of water getting into the plugs and leads after an over zealous engine bay cleansing. No steam just plenty of water and some leaked down the plug caps and gave the engine a terrible tractor effective and loss of power till high revs.

If base idle is too low then that will cut out when coming off motorways after long periods of driving.

Another possibility could the the immobiliser being intermittant if it just cuts dead, not an area that is thought about if just a factory unit installed but could have same effect. Seem to remember someone in the summer having a problem with the remote and it randomly arming and disarming. (1/01)

If Mr Mazda has used his error code reader for the engine ECU (please check) .. then they should know what is wrong .. and simple things like fuel pump relays would be part covered by this. ..

the other alternative is for an after market alarm acting as an immobiliser .. the Mk2 has a built in immobiliser on the engine ECU linked to the key .. this could have a questionable connection .. Do you have a big bunch of keys on the ignition key?? (this makes car ignition keys unreliable). (1/01)

only similar was with a load of water getting into the plugs and leads after an over zealous engine bay cleansing. No steam just plenty of water and some leaked down the plug caps and gave the engine a terrible tractor effective and loss of power till high revs.

If the engine runs rough then suspect plugs and/or plug leads.
If it dies with an error code suspect a sensor.
If it dies without an error code suspect fuel or electric problem (usually ignition switch). (1/01)

I asked the following on www.mx-5.com: There are a number of well known cases of early '99 Miata's which had an engine failure due to an improperly machined number 4 main cap of the cylinder block. Does this problem also exist in early '99 MX-5 engines (1.6l or 1.8l)? If it is, then which range of VIN numbers might be susceptible to the problem?

and this is the reply I got: Yes this problem also applies to early 99 NB 1,6L engines. The 1,8L never had this problem.

Only a very limited number had this problem and not all engines in a VIN range were affected. Therefore we do not have a fixed VIN range that we can publish.

Significant for this problem was, that in the majority of cases, the engines were younger than 20.000 km. If the symptoms (knocking when releasing the clutch under load, i.e. driving off) did not appear before 20.000 km, the engine bearing is sound and the problem will probably never appear. As most 99 MX-5 have a higher mileage (km) now, we have no such problem any more.

So according to Mazda, 1.8l MX-5's (i.e.: european cars) should be safe while early '99 1.6l MX-5's might also suffer from the dreaded #4 bearing problem. That's not what I hoped :^(.

Granted, I do have 50000kms on my 1.6l, but I'm not sure whether the problems in the US were really limited to low mileage engines. (7/01)

If in doubt get an engineers report to confirm the state of the engine. Also IF you have taken the car into the garage prior to the warranty expiring and they give it a clean bill of health and THEN the fault occurs and it can be proven back to the issue then you have good recourse against the dealer and potentially over the franschise operator in your region as the dealer often works as an agent for them.

If it can be shown that the fault was a manufacturing fault and was evident prior to the expiry of the warranty I would expect them to honour the engine change no matter how old the car - especially as this is a well known issue.

Before anyone suggests the same could be said for the NA LWC issue then that is due to bad servicing that it fails as I understand it - a mechanic not applying the correct torque when working on the engine and is NOT a manufacuring fault so I would not expect the manufacturer to stump up in that situation. (7/01)

I had heard this also last year from my mazda dealer. They have a TSB describing the problem. They said thay had replaced 2 engines, during the first 6 months of the 1.6l NB being available in the UK.

Summary of situation
Affects: US 1.8l cars US. 1.8l protoge (323) European 1.6l cars Other european Mazda models??
Not affects: European 1.8l cars

Why can this be?? Well one thing I do know that here a 2 types of the 1.8l engine BP4W and BP4Y - the BP4Y is used in the US where there are tighter emissions controls - For sure I know the cams are different, as well as the block itself. The TSB on miata.net lists a different part number for the Cal spec short engine, for example.

If you have a European (or Japanese market) 1.8l NB you will find many components are marked BP4W on the engine castings itself, and many ancilliaries. Also (having taken a 1.8l NB engine apart) the crankshaft, pistons valves etc are also marked accordingly. On this engine there was no measurable wear on either thrust washer after 13K miles, just some slight "polishing" of the rear most washer that takes the force when the clutch is applied

(the crankshart weighs a f**king ton by the way - more than the entire cylinder heard). I would be interested to hear what European 1.6l engines are marked with.I guess it could be B6 with a number a letter afterwards.

From what I can gather, the problem is the selection of incorrect sized thrust washers rather than the main cap being incorrectly machined itself. A random problem caused by an individual person or piece of line equipment I would guess.

There is probably more than one engine line in Mazda's factory. We can perhaps summize the US 1.8l and European 1.6l line is shared, while the 1.8l line for Europe and Japan are separate. If we hear of more than one 1.8l NB engine failing in Europe then I guess this throws my theory out of the window. (7/01)