Frequency comparator - engine sync

[Wiggo]

After some ideas here, if anyone can help.

I have a twin engine/twin screw boat and I want a simple indicator to show when the engine revs are synchronised. The analogue tachos are not desparately accurate, and I know I have a higher fuel burn on one engine so I want to make sure they're running at the same revs.

I could easily use the tacho output from the alternators, or I could simply mark up the shafts and use an optical sensor to get direct shaft rpm, or attach a small magnet to each and get pulses that way.

In terms of display, a simple 3 LED system would be fine - Port too high - Equal - Starboard too high, or it could get fancy and use multi colour LEDs.

I've some stuff around using a pair of flip flops with their outputs NANDed and feeding an op-amp, but my electronics is too rusty to figure out the display part.

Any ideas gratefully received...


Edit: FWIW, the frequency range is either 600-3800rpm at the engines, or half that at the shaft output.

 

[BrownOut]

What type of engines do you have? (Deisel or Gasoline) If you have gas engines, you can just pick up the signal from your ignition system. That would be the easiest way, IMO. It they are deisel, then I don't quite know.

 

[kpatz]

For diesel you could pick up the injector signals, but chances are there's an rpm sensor that drives the tachs anyway, why not use that?

A phase comparator like that in a PLL could be used. If the filtered output of the comparator is unchanging, the RPMs are equal. Or read the RPMs with a microcontroller and implement the "phase comparator" in software.

Can't you tell from the sound/vibration when the RPMs are synchronized?

 

[Boncuk]

Can't you tell from the sound/vibration when the RPMs are synchronized?

I know from flying that two turboprop engines spinning at 18,000 rpm it's hard to know if the engines are synchronized (resulting in a small rpm difference of the props (dia 5.5m) not causing feelable vibrations). However the air frame "feels" those vibrations resulting in material fatigue.

Boncuk

 

[Wiggo]

The engines are diesel. There's no real issue with picking up a signal - although they're mechanically governed, I can pick up the tacho output signal from the alternators if I have to. I'd rather go direct from the crankshaft or propshaft (output from a 2:1 reduction gearbox), as the alternator drivebelts can sometimes slip causing low readings, particularly when the electric clutches for the superchargers engage.

At a cruising speed of 3400 rpm, 100 rpm difference in revs is impossible to detect - you need ear defenders in the engine room, and any vibration doesn't seem to get through to the hull structure, so I need an electronic solution.

A comparator that give a +ve output for port too high, 0v for even and -ve output for starboard too high would do - I can work out the LED indication bit from there. My challenge is how to reliably detect the frequency differences: at tickover, the shafts are turning at 300 rpm or only 5Hz, so the frequency range I'm interested in comparing is around 5-65Hz. I've never used PLLs - would they work reliably at such low frequencies? If need be, I guess I could attach multiple magnets to the shafts and multiply the frequency by a factor of 4, if that would help.

 

[stevez]

I'd think a digital tach on each engine with display is the place to start - you have to read the display and decide if the difference is ok or not. Sounds like you do not intend to control your engines differently - just manually adjust the throttle for same RPM.

If you do evolve to some simpler way of signalling "synchronization" you need to decide on how close you want to be or can be as you might never achieve or hold perfect synch and you'll make yourself crazy with the lights switching from one side to the other. I could be wrong but hard wiring a digital circuit so that it can ignore small differences would be difficult.

Maybe digital tachs with analog comparison that you can fine tune. The alternative would likely include a PIC or small computer.

 

[Wiggo]

Could this be used as a basis? rather than use the LM311, use an op-amp to give an analog output, so 0-2V = port too slow, 3V = sync, 4-5V = stbd too slow? I could then set a reasonable dead zone for the sync range that didn't involve the LEDs flickering all the time.

 

[Diver300]

A PLL or Phase Locked Loop needs some sort of control of one or other motor.

If you want good synchronisation, you shouldn't run from the alternators, as the belt slip would mess up the speeds. I guess that the tachometers are run from the alternators.

I suggest a proximity sensor pickup on each engine. You then need to divide both signals by the same number, maybe 100. You then compare the phase of the divided signals. That can be done by connecting the two signals to the two inputs of an exclusive or gate. If you smooth the output of the exclusive or gate, you get an analogue voltage, which can be used to control the speed of either engine.

That works like this:

The divided signals from the engines speeds will be square waves with a period of around 3 seconds. When two signals like that are connected to the exclusive or gate, the output waveform will have a period of half in the input period, but the duty cycle will depend on the phase.

If the phases are exactly the same, the output will be 0 all the time. If the phases are exactly opposite the output will be 1 all the time. The amount of time that the output is 1 can be anything, depending on the phase of the difference between the two inputs.

The output is smoothed to become the average voltage. That will go up and down depending on how much earlier one engine's divided signal is than the other. If that is used to control one engine, its speed will follow the other one.

 

[ccurtis]

I came up with the attached circuit after some consideration. Phase comparators are a problem since they are too sensitive for your application. Even if you manage to get the two speeds very close, the phase will vary all over the place, though at a slower rate.

The attached circuit counts the pulses from each shaft and compares the upper three bits from each counter. The lower bit (LSB) is ignored to provide a "dead band" so that a difference in speed of more than 15% holds the LEDs in a constant state, with one LED on for the faster shaft, and the other LED off. As the difference becomes less than 15%, the LED corresponding to the faster shaft starts to flash (at higher rates of comparison, in the 100's of Hz or more, the LED brightness decreases). As the difference in speed between the two shafts diminishes, the LED for the faster shaft stays off for more time than it is on. When the speeds are identical both LEDs are off, regardless of phase difference. The LED for the slower engine is always off.

 

[Boncuk]

Hi ccurtis,

do you know the 4029 is obsolete?

 

[tcmtech]

One possible option is to use two frequency to voltage converter IC's with a +- swing analog meter between the two IC outputs. If both outputs have equal RPM'S the needle would stay centered but whichever one is running faster would make the needle point toward it and proportionally more or less depending upon the RPM difference.
Using the right voltage output scale and meter you could easily have a scalable reference that could be accurate from 0 - what ever RPM difference limit you chose and show which engine has the higher or lower speed in comparison to the other.

Using a set of IC's like the LM2907 frequency to voltage converter type would be simple and easy to build and work with.

Just a thought.

 

[Dean Huster]

If you can't find any new, LM2907s are readily available from old 5.25-inch floppies as the spindle drive motor tach circuit.


If you're going to go to all the trouble to detect the RPMs of both engines and still have to do some manual tweaking on the throttles to make them match, why not go all the way, set the speed with the starboard engine, set the port engine close and then flip a switch to "AUTO" and let the output from the starboard tach adjust the throttle of the port engine to match? Hey, if it's going to be a neat project, why not make it an insanely COOL project to boot?

Dean

 

[Boncuk]

Hey, if it's going to be a neat project, why not make it an insanely COOL project to boot?

Dean

Something like that?

 

[marcbarker]

edit: oops! tcmtech said the exact same thing!

Feed each tacho signal into a frequency to voltage converter, so that revs = DC voltage for each engine.

Connect to the 2 outputs, a 'centre-zero' moving coil meter between them.

The meter displays RPM difference.

Trim the engines to zero the needle.

 

[Wiggo]

If you can't find any new, LM2907s are readily available from old 5.25-inch floppies as the spindle drive motor tach circuit.


If you're going to go to all the trouble to detect the RPMs of both engines and still have to do some manual tweaking on the throttles to make them match, why not go all the way, set the speed with the starboard engine, set the port engine close and then flip a switch to "AUTO" and let the output from the starboard tach adjust the throttle of the port engine to match? Hey, if it's going to be a neat project, why not make it an insanely COOL project to boot?

Dean

Good old fashioned mechanically governed 6 cyl turbo- and supercharged intercooled diesels, I'm afraid, so "AUTO" just isn't an option unless I fit a pair of huge linear actuators to drive the throttle levers on the injection pumps...

2907's and a suitable meter or voltage comparators/LEDs sounds like the way to go. Thanks, guys!

 

[Boncuk]

I'm afraid, so "AUTO" just isn't an option unless I fit a pair of huge linear actuators to drive the throttle levers on the injection pumps..

Wow! It must take a bear of a man to operate the throttles of your engines.

No chance for a fragile female to drive the boat.

 

[Mr RB]

This project is just crying out for a PIC. Just measure the 2 RPM inputs on 2 PIC pins, get the PIC to calculate the difference as a nice human-friendly percentage difference between motor RPMs etc, then display to a simple row of leds;
L6%, L4%, L2%, MATCH, R2%, R4%, R6%

It won't need much more than a bottom range PIC and 7 LEDs.

 

[Wiggo]

Wow! It must take a bear of a man to operate the throttles of your engines.

No chance for a fragile female to drive the boat.

Volvo Penta EDC (Electronic Diesel Control) Actuators are physically large - about 18" x 4" x 3", as well as being very expensive...

Never used a PIC, where would I start?

 

[Boncuk]

Volvo Penta EDC (Electronic Diesel Control) Actuators are physically large - about 18" x 4" x 3", as well as being very expensive...

I've seen physically large Volvo trucks being driven by girls.

If it's electronic control it's something like "fly by wire", isn't it?

 

[BrownOut]

I don't think he's talking about large force being needed. He's talking about the length of the "throw" to actuate the controls. A female would have both the strength and arm length to handle it, but an actuator would have to be pretty long.