How would I measure RPMs in a diesel engine?

[NDT]

So, diesel engines don't have a distributor, or spark plugs synchronized with the engine RPM neither in mechanical or electric or electronic form. They don't have coil packs either.

But diesel engines do have fuel injection, and they therefore have fuel injectors to inject fuel in a precisely controlled way at each cylinder. The fuel injection is absolutely synchronized with the engine RPM (how else would the engine run?) and in the case of electronic direct injection the injectors may look somewhat similar to a coil pack. I read nothing inaccurate in what deth502 wrote.

Again, MODERN diesel engines have EFI. But as I've said already, mechanical pumps or electric pumps with mechanically controlled injectors don't have the surrounding paraphernalia to extract a useful RPM signal of them. Anyway, until more is known about the engine used, we are all speculating.

[deth502]

So, diesel engines don't have a distributor, or spark plugs synchronized with the engine RPM neither in mechanical or electric or electronic form. They don't have coil packs either.

But diesel engines do have fuel injection, and they therefore have fuel injectors to inject fuel in a precisely controlled way at each cylinder. The fuel injection is absolutely synchronized with the engine RPM (how else would the engine run?) and in the case of electronic direct injection the injectors may look somewhat similar to a coil pack. I read nothing inaccurate in what deth502 wrote.

The problem is, you don't know how many injections are being done each stroke. Some diesel engines can do 5 or more injections per cylinder per compression/combustion stroke. This will very based on ambient conditions, engine load, engine operating mode, and many other things...

true, and the on-board computer derives the tach signals from these. which is what i recommended, trying to tap into the computers tach signal that it is already generating. i never recommended taking a pulse directly from an injector for this very reason.

[hechen]

This engine most likely has a electrical method of injecting fuel (no carburetor, etc).  If this is the case there needs to be some kind of crank position input to the ECU (i.e. crank sensor).  The easiest method is tapping into the crank sensor but I would not recommend it.  I would recommend tapping into the CAN network and getting the RPM off of that.

[Paul Moir]

This engine most likely has a electrical method of injecting fuel (no carburetor, etc).

Diesels don't use carburetors, and most don't use an electrical way of injecting it.  What we really need to know here is what kind of engine he's trying to measure the RPMS of.  I'm picturing an old Westerbeke marine diesel.  But maybe it's a car engine with a CAN bus and all sorts of goodies attached.  Or perhaps it's a model airplane glowplug engine?  I'm guessing it's not a locamotive or a Wärtsilä since those often come with tachometers already.

Second left field idea:  Does it have an alternator?  What about measuring the current wave out from it? 

 

[IanB]

My current project is a good one (details will come later but it's something I am proud to be involved in)

I'm rather interested in why this project is so good and why it makes Gregory so proud to be involved with it...?

[peter.mitchell]

I assume the engine is driving something, not just an engine, so it will probs have a water pump or powersteering pump, which may be driven directly from the engine, to which it may be more practical to measure the rpm from.

[Paul Moir]

I'm rather interested in why this project is so good and why it makes Gregory so proud to be involved with it...?

Fool me....   You can't get fooled again!

Damn contests!

[johnwa]

Second left field idea:  Does it have an alternator?  What about measuring the current wave out from it?

Yes, depending on the engine, this would probably be easiest. I recently had a similar need to monitor the speed of an approx. 4kW diesel powered water pump. I was looking at fitting a sensor to run off the cooling fan or the starter ring gear, but in the end I took a signal from the output of the alternator (which was fitted directly onto the crankshaft). No mechanical modifications needed! The alternator must have been a 6 pole design too, as the speed turned out to be a simple factor of ten times the output frequency.

EDIT: I meant six pole, single phase - a three pole alternator would not make very much sense!

[gregoryfenton]

The engine would be a nautical one, it is for use on a small (43 feet) boat.

I can't believe how much debate this question has started

I hope that narrowing down the engine type will help choose the correct option.

Thanks again all for your comments thus far and any further commentary

[G7PSK]

The engine presumably has an electric start so there would have to be a starter ring gear, put a hall sensor close to that and count the teeth most ring gears will have enough residual magnetism to be seen by a hall sensor. Not my idea I have seen this commercially done.

[fcb]

The engine presumably has an electric start so there would have to be a starter ring gear, put a hall sensor close to that and count the teeth most ring gears will have enough residual magnetism to be seen by a hall sensor. Not my idea I have seen this commercially done.

+1 for a hall effect looking at the ring-gear.  Buy one off ebay "crank sensor" or "flywheel sensor" and use that.

If you don't want to get *that* involved, then you could always use a small microphone and listen to the exhaust. I don't think you'll need much in the way of signal processing to create a pulse per detonation.

[NDT]

The engine would be a nautical one, it is for use on a small (43 feet) boat.

I can't believe how much debate this question has started

I hope that narrowing down the engine type will help choose the correct option.

Thanks again all for your comments thus far and any further commentary

Besides it being for a boat, could you tell something more about it? The most important thing, does it have EFI (electronic fuel injection) or the fuel injection system is of the mechanical type?

[NDT]

Second left field idea:  Does it have an alternator?  What about measuring the current wave out from it?

Yes, depending on the engine, this would probably be easiest. I recently had a similar need to monitor the speed of an approx. 4kW diesel powered water pump. I was looking at fitting a sensor to run off the cooling fan or the starter ring gear, but in the end I took a signal from the output of the alternator (which was fitted directly onto the crankshaft). No mechanical modifications needed! The alternator must have been a 6 pole design too, as the speed turned out to be a simple factor of ten times the output frequency.

EDIT: I meant six pole, single phase - a three pole alternator would not make very much sense!

In applications where the engine drives loads at constant speeds (i.e. alternators as emergency generators, hydraulic pumps, etc) or if the accessory being driven, as is your case, has a direct or rigid mechanical connection with the engine, it's perfectly fine to do as you tell. On most engines, including marine engines, the accessories are driven by a pulley and belt system and the belt slippage plays ebb important role on the dependability of the RPM indication. If the engine is used in such a way that its power output is rapidly changing, or if the belt has wear, or if the tension isn't adequate, belt slippage could be significant.

[gregoryfenton]

I'm rather interested in why this project is so good and why it makes Gregory so proud to be involved with it...?

Fool me....   You can't get fooled again!
Damn contests!

The timing is rather iffy with the contest and all I appreciate (not trying to bump my post count), but this is a genuine thread.

I am not at liberty currently to reveal the nature of the project but when it is revealed you will understand my pride, you may even have pangs of "I am so envious, what a great thing to be involved in."

I will see if Dave will comment and inform you of my genuine intentions. @eevblog message incoming!

[EEVblog]

I will see if Dave will comment and inform you of my genuine intentions. @eevblog message incoming!

Yep, confirmed. It's a real project.

[gregoryfenton]

I will see if Dave will comment and inform you of my genuine intentions. @eevblog message incoming!

Yep, confirmed. It's a real project.

Your reply is very much appreciated Dave

[mjrandle]

Not sure what sort of accuracy you requie, but if the diesel engine has an alternator then it probably has a tacho output on the alternator.  Given the crank and alternator pulley effective diameter you would then be able to calculate engine speed.


Cheers,

Mike

[johnwa]

In applications where the engine drives loads at constant speeds (i.e. alternators as emergency generators, hydraulic pumps, etc) or if the accessory being driven, as is your case, has a direct or rigid mechanical connection with the engine, it's perfectly fine to do as you tell. On most engines, including marine engines, the accessories are driven by a pulley and belt system and the belt slippage plays ebb important role on the dependability of the RPM indication. If the engine is used in such a way that its power output is rapidly changing, or if the belt has wear, or if the tension isn't adequate, belt slippage could be significant.

Yes, I agree, if there is a belt involved, this is probably not the best method. Although a properly designed belt drive in good condition should not have much slip, there are too many variables for this to give an absolute indication of speed.

[Lifeboat_Jim]

Hi Guys

I'm Gregs partner in crime in this project. Firstly a massive thank you for the interest shown in helping out with advice & suggestions, very much appreciated!

Many of you have asked for details, so Greg gave me a nudge to register here and provide them directly.

Background and details

As mentioned previously, it's a boat (in a couple of weeks time we'll unveil all the details) project but this is an ancillary aspect to it.

The engine in question is single cylinder diesel, direct injection, mechanical governor (for those diesel heads it's a 1995 Lister Petter AD1C, Build 02... so Ring Gear with the optional electric starter and variable speed 1,000-3,600RPM circa 7.5HP @ 3,600RPM). Currently hooked up to an Ingersoll Rand 'Euro' Air Compressor Pump Head via belts. It has absolutely no modern features, no alternator, no electronics etc. It doesn't have a Heater Plug (this is an option, but not fitted). This is the sort of engine used on Concrete Mixers, temporary Traffic Lights, Compactors, Water Pumps etc.

This was bought to drive a Shot Blaster for the boat restoration aspect. Once that phase is done I was minded to take the Compressor Pump off and fit a GenSet (or possibly a heavy duty 24v Alternator) and fit it on the boat as a power source (battery charging etc) thereby squeezing every last drop of value out of it

The ultimate plan is to; measure the RPM, fit a Relay to the Starter circuit, add Servos to operate the Decompressor and Engine Stop. In short have the ARM based microcontroller automate/monitor it, but of course each of those one step at a time.

This small standalone simple engine becomes the proof of concept/test bed for...

Big Marine Engine

Well, this is for a several months time but the main propulsion engine is a 1964 Parsons Porbeagle (a marinised Ford(son) 4D, 4 cylinder Diesel) yielding 57HP @ 2,250 RPM iirc. Again, as you may imagine from the vintage, lacking all the mod-cons). The boat has/had two of these directly attached to each of the two prop-shafts.

We're looking hard into switching from two of these directly coupled to just the one. This will then drive a Hydraulic Pump and the props will be, in turn, driven from individually controlled Hydraulic Motors.

End goals

To remotely monitor all aspects of both the small and large engines (RPM, Start/Stop, Fuel, Oil, Temperature etc) and remote control them.

Also to measure the direction and speed of the two prop-shafts.

RPM

So, back to the original task/question 'How would i measure RPMs in a diesel engine?'.

Thoughts about using TCRT5000L Reflective Optical Sensor with Transistor Output (and a Schmitt Trigger to smooth the output)? I see many fans of Hall Effect (and some very good reasons why being put up too). Given measuring the prop-shafts is also a consideration and standardising on a single method has advantages the IR route is certainly attractive, and IR has its supporters here too. Perhaps we need both, Hall Effect on the Engines and IR on the prop-shafts?

Cheers

Jim.

[jeroen74]

One method is to mount a small (piezoelectric) microphone to one of the injector pipes. The high pressure pulse is easy to detect; easily even by ear if you use a stethoscope or the screwdriver method.

[Lifeboat_Jim]

One method is to mount a small (piezoelectric) microphone to one of the injector pipes. The high pressure pulse is easy to detect; easily even by ear if you use a stethoscope or the screwdriver method.

Interesting. you're the 2nd to mention acoustic measurement. When mounted on the boat they'll exhaust through silencers (interesting factoid... 'hospital(ler) silencers' are thought to derive their name by being quite enough to be used in/near Hospitals... sounds sensible enough to be true!). With two diesels running in the same compartment I wonder how effective/accurate it would be to implement. Perhaps going for vibration rather than sound waves?

They do that in F1 by the way to spy on other peoples cars (as they have detailed telemetry on their own of course).

Anyway, that is a possibility for engine RPM (assuming the difficulties, aren't), but an interesting project nevertheless perhaps as a redundancy method, but not for prop-shaft revolutions so we'd still need a solution for that.

[NDT]

I still think that you should go with the hall effect sensor. And since you plan on monitoring other engine parameters, you also should go with the corresponding automotive sensors according with what you want to measure. The sensors that you need (hall effect for the crankshaft and propeller, temp sensor in the form of an NTC or PTC sensor for air, fuel, or other fluid, etc) are already available and specifically designed for engines, so, why reinvent the wheel? They are rugged, dependable, accurate, and cheap to not too expensive.

[Mr Smiley]

Do Diesel engines "Knock" like over lean / low grade fuel petrol engines 

 

[Lifeboat_Jim]

I still think that you should go with the hall effect sensor. And since you plan on monitoring other engine parameters, you also should go with the corresponding automotive sensors according with what you want to measure. The sensors that you need (hall effect for the crankshaft and propeller, temp sensor in the form of an NTC or PTC sensor for air, fuel, or other fluid, etc) are already available and specifically designed for engines, so, why reinvent the wheel? They are rugged, dependable, accurate, and cheap to not too expensive.

Could you expand on that please? NTC/PTC etc. I'm a complete n00b on electronics. Although Greg is quite handy on general electronics I don't think he'd know what those are either.

[Rigby]

J1939 is how I would do it.  If the engine is even close to modern and it goes in a vehicle then it has a data bus.  CAN, MODBUS, ISOBUS, something.  The tach measurement will probably already be in there.  There are small pump-house diesels that don't have a databus, though, and if this is the case with you, then I would find a good spot to paint a white mark and observe it digital laser tachometer or something.