What would the brains trust recommend ?

craigcurtin

Senior Member
Guys,

I am in the process of deploying a diesel generator - based on the venerable Lister engines.

The engine will operate in the 850rpm range.

I would like to be able to monitor the speed of rotation of the flywheel with a picaxe.

I do not believe mounting anything (mechanical) onto the motor will survive for long as these motors develop a fair bit of vibration.

I was going to use a Kiwi patch board to develop the board on and am happy to use an X2 (or maybe even the new M2 !!) when it is available.

I will be using a regulated wall wart power supply.

I am in Australia.

regards

Craig
 

kd5crs

Senior Member
based on the venerable Lister engines
I'd say don't let anyone named Rimmer secure the drive plate.

More seriously, this seems like the perfect time to use a timing light. Can you draw a bright line on the flywheel for the light to reflect? Vary the strobe until the light matches the rotation speed. Use a light sensor to read the flashes.

That's my thought, I've never tried anything like it, and in fact have only seen a timing light being used once when I was 19 over a decade ago.

Brian
Member of the brian trust
 

MartinM57

Moderator
First questions:
- how accurately?
- updated how frequently?

Suppose the answers are 10rpm and 1 second...not unreasonable.

So 'something' (to be defined) exists to indicate each revolution. In 1 second you count whole 14 revolutions. How fast is it going round? Answer - something between (or equal to) 840 rpm and less than 900 rpm. So your accuracy is 60 rpm. Oops.

So you need to sense more than each revolution. Ideas for sensors:
- reflective (is this what the engine is like - http://www.me.columbia.edu/me3410/spring06/group01/UserManual/UserManual.pdf ?)
- variable reluctance (http://en.wikipedia.org/wiki/Variable_reluctance_sensor)
- hall effect (stuck on magnet?)
- out-the-box thinking - push rod/rocker/valve movement?

M57
- I'd never trust a Brian
 

Jaguarjoe

Senior Member
I'd say don't let anyone named Rimmer secure the drive plate.

More seriously, this seems like the perfect time to use a timing light. Can you draw a bright line on the flywheel for the light to reflect? Vary the strobe until the light matches the rotation speed. Use a light sensor to read the flashes.

That's my thought, I've never tried anything like it, and in fact have only seen a timing light being used once when I was 19 over a decade ago.

Brian
Member of the brian trust
What happens when the speed changes?
 

hippy

Technical Support
Staff member
My first though is use PULSIN, use a trigger per revolution and feed that through a flip-flop so on for one revolution, off for the next. At 850rpm that's a revolution every 70ms or so which is annoying as it's greater than the 65535 max of PULSIN. You could simply underclock ( or use the pulse to gate an oscillator in and use COUNT but now getting complicated ).

Assuming you underclock, 850rpm should read 35000 so quite high resolution.

If it's a question of sensor ... no idea.

Added : Maybe that gated clock isn't such a bad idea. Adjust the clock oscillator so you read 8500 pulses in however long a revolution is at 850rpm and it's instant rpm to 0.1%

Extra Added : ... except it's an inverse.
 
Last edited:

Andrew Cowan

Senior Member
Does a hall effect sensor and magnet count as mechanical? If you can attach a magnet to the flywheel, you'll get a solid reading no matter how much dirt, grime or dust is around. Reflected light or light gates are OK until the equipment gets dirty.

A
 

papaof2

Senior Member
A series of magnets around the circumference would provide more pulses and not affect the balance of the flywheel (a single magnet *might* affect the balance). With magnets 30 degrees apart, you would have 12 pulses/rev times 850 rpm = 10200 pulses/minute or 170 pulses/second. You should be able to use pulsin to get the speed with updates of less than one second and also be able to see small variations in the speed.

If one side of the flywheel stays relatively clean, then one or more pieces of reflective tape (perhaps just the aluminum tape used on heating ductwork) can be used with a light source and a detector. I have a very old analog RPM meter that uses a focussed light source and a detector to read speeds up to 12000 rpm. (Looks like this: http://cgi.ebay.com/PIONEER-PHOTO-TACHOMETER-MODEL-36-CASE-W-MANUAL-/350299637588?cmd=ViewItem&pt=LH_DefaultDomain_0&hash=item518f7c4754)
3M Metal Mending Tape is available in small rolls and works well - a single piece on a small gasoline engine crankshaft gives a stable reading at 3600 rpm (yard power equipment).

John
 

LizzieB

Senior Member
Is your Lister flywheel one with spokes or maybe round holes? If so an optical interrupter could be a good sensor solution.
 

westaust55

Moderator
850rpm? Most Listers run 600-650 rpm. Lister (& "Listeroid") type motors -well thought of down under since the late 1800s- suit bike computer tachometer persuasion. See => http://diesel-bike.com/Lister_Gen/Lister1.html
850 rpm does not sound correct. typo ? :confused:

for 50Hz operation
generator vs Speed
2 pole = 3000 rpm
4 pole = 1500 rpm
8 pole = 750 rpm

even at 60 Hz with 8 pole generator the engine would run at 900 rpm not 850 rpm .

Have had Lister diesel engine driven gensets running at 1500 rpm before today.
 

westaust55

Moderator
Sensing the speed of a genset

@Craig,

you could always consider sensing speed from the generator output.
Have a read of this microchip App note:
http://ww1.microchip.com/downloads/en/AppNotes/00521c.pdf

That way there is no modification to the generator to detect the speed.
Presumes the onboard AVR and exciter will bring genset output volts up as soon as the engine starts to rotate.


(now just wait for the posts on concerns over connection to "High Voltage")
 

Marcwolf

Senior Member
Ok WestAust55 I'll bite

"Oh my god.. High Voltages"

But it sounds a very interesting project. When I sometimes perform at country faires I love to look at the old engine and steam groups with the old faithfuls chugging there.

Ok - there has been some suggestions of dividers etc but it one considers this

900rpm = 450 turns every 30secs = 225 turns every 15secs - so by using a smaller timeslice you can work with bytes or words for the pulse counting.

After you have the figure then you multiply by whatever factor you want to get the correct RPM for display or processing

Take Care
Dave
 

craigcurtin

Senior Member
thanks for all the info - that got everyone going !!

Yes it is an Indian Listeroid and it is definitely 850rpm - it is an 8/1 (8HP, single cylinder)

It uses 20 inch flywheels and a different size pulley on the Genhead to get the correct Hz/Volts for Australia.

From what i have seen with these once they have been in operation for a while then they will get pretty dirty - so not so sure about longevity of stick on reflective pieces etc.

It is important to monitor engine RPMs to detect runaway conditions - no good monitoring the alternator output if the belt stretches and slips or breaks completely. Longer term i would also like to be able to implement an electronic governor to get the output more stable but that is not part of the first round.

Within 10 RPM is fine in terms of monitoring and reporting

regards

Craig
 

Marcwolf

Senior Member
Hi.

Well a magnet and hall effect will do ok as they are no effected by grease etc. If you think about it 14.16 counts per second is not a great amount (850/60) so I am sure that a Picaxe can handle that easily
Please Put photo's up when your finished as I'd love to see the engine

Take Care
Dave
 

Dippy

Moderator
Can you tap into the alternator side of things to get a signal that can be twiddled to get a suitable PICAXE-friendly pulse?

Otherwise it's surely simply a matter of finding a solid place to mount a sensor - whether magnetic or optical.
You have the engine in front of you, I don't sadly.;)

At those sorts of speeds the RPM detection should be easy peasy.
 

manuka

Senior Member
I've just recalled the old piezo sensor trick used to detect petrol pre ignition "knock". They're available commercially as the Tiny-Tach, but for such a slow donk of a diesel engine it'd be a doodle to rustle one up with a PICAXEd piezo I'd say. Give it a try- even flicking a piezo element with your finger will produce a spike capable of flashing a LED.
 

InvaderZim

Senior Member
If you mount a magnet close to the hub (as opposed to the rim) I would think the disturbance would be minimal. One magnet and a hall sensor should work great.

At 70ms between pulses, perhaps using an interrupt would be the most accurate. You could service an interrupt, debounce the input, etc, every count.

In my mind capturing the pulses isn't the hard part. The hard part will be accurately measuring time, or calibrating the Picaxe to measure time precisely enough for this application. But I'm sure that topic has been covered here before!
 

westaust55

Moderator
Yes it is an Indian Listeroid and it is definitely 850rpm - it is an 8/1 (8HP, single cylinder)

It uses 20 inch flywheels and a different size pulley on the Genhead to get the correct Hz/Volts for Australia.
Ah okay - new information. Alternator is belt drive. I only use direct driven/couples sets.

So yes, detection from alternator output is not "safe"
 

Dippy

Moderator
Ah yes.
Well, with every downside there is an opportunity.

Firstly, Craig, you have this thing sitting in front of you , we don't.
Surely you can come up with some robust method of mounting something?
You've been on this Forum long enough to have seen the plethora of hardware/software techniques to measure RPM.:confused: It's not difficult.

And just think, if you measure engine RPM and alternator output you'll be able to tell other things too....
 

hippy

Technical Support
Staff member
Ah okay - new information. Alternator is belt drive. I only use direct driven/couples sets.

So yes, detection from alternator output is not "safe"
In an absolute sense that is correct, but what are the likely failure modes, and what exactly are we controlling or more simply monitoring ?

I don't understand these engines so don't know how they work; are we trying to detect an engine whose 'feedback' has somehow failed and is going into meltdown in an effort to compensate, or is at an open loop system where the engine may race out of control ? Is it the case for the most part everything will be fine and we just want some indication that it may not be ?

It's true that the alternator output will not indicate what the motor is actually and absolutely doing, but alternator giving out of spec indicates a system problem, complete belt breakage would be detected, ideal revving and belt slippage would be detected, over-revving with no belt slippage would be detected, the only case I can see which could slip through the net is over-revving combined with belt slippage which gives the impression everything is hunky-dory when it isn't. Is that likely to occur in a manner which would not be detected ? I can't answer that.

I think we're missing the big picture, why direct monitoring is necessary and why indirect monitoring will not work or why we even need such monitoring beyond it being A Good Thing (TM).

In terms of what I can relate to it's like my indirect monitoring of the electricity grid. If my lights go out, are dim or too bright, I will know there's a grid problem, the only thing I wouldn't observe is plant going critical while the lights remained at the expected level.
 

Dippy

Moderator
Doh, and I was trying to promote some self-thinking in my last sentence....;)

Come on Craig, get on with it... you've got enough ideas by now :)
 

papaof2

Senior Member
In an absolute sense that is correct, but what are the likely failure modes, and what exactly are we controlling or more simply monitoring ?

I don't understand these engines so don't know how they work; are we trying to detect an engine whose 'feedback' has somehow failed and is going into meltdown in an effort to compensate, or is at an open loop system where the engine may race out of control ?
Abbreviated info on diesel runaway - and still probably more than you ever wanted to know.

There are failure modes for diesel engines that can cause a runaway condition which is not controllable in the manner of a gasoline engine (close the throttle or turn off ignition to slow down/stop the engine). Since diesel ignition is from the heat of the compression stroke, there is no ignition switch in the sense of an electrical cutoff - the "ignition" switch actually controls fuel flow (pump or solenoid). If the engine wears enough to allow crankcase oil to be sucked up into the combustion chamber, the engine has (for a short while) unlimited fuel supply and can go into runaway - not a good condition for an engine designed for relatively low speed operation.

John
 

Dippy

Moderator
Do they have a compress release thingy?
If so, that could prevent a TMI.
You could all this done in a couple of days.
 

Johnmb

Member
Brain Trust Recomendation?

Go see the Wizard of OZ before you do anything else and 9 times out of 10 that's all you need to do.
 

craigcurtin

Senior Member
OK plenty of ideas

The system will have an engine controller that will monitor a variety of parameters and will manage the startup/shutdown/safety aspects - fairly straight forward and for about $200 (AUD) it is a no brainer.

My interest is in more longer term analysis of what is happening with the engine - yes i will be monitoring alternator ouput (current) with a hall effect sensor to enable me to bring loads in and out on the motor to keep it operating at the optimal level.

A diesel engine in a runaway mode is not a pretty sight and is really scary if it also has Indian cast iron flywheels !

I will also look to replace the mechanical governor in the longer term with an electronic one to get more accurate control of the motor.

I think i will initially go for a magnet mounted close to the crankshaft and see how close we can get with that.

I will post some pictures as the build progresses. Currently in excavation stage to enable a shed to be built for it to live in

regards

Craig
 
Top