Britbike forum Forums British motorcycles in general British motorcycles in general Measuring alternator output

Quote: One element missing from all this but we have partly spoken about is rotor magnetism. With all the old rotors around it would be nice to have a simple(ish) test for that. No doubt someone will suggest something <s>.

Yes, install the rotor on the bike and place a one ohm, 100 watt resistor resistor across the outputof the stator and compare the reading against the figures in the Triumph Workshop manual... they are there for a reason.

You could spend some big pounds on a Gauss meter, but we can't get past the resistor. The "if it will holds its own weight when placed on the side of the tool box" doesn't work!

Geez! Like a bad penny, that resistor just keeps rearing its ugly head. So you spent 15 quid at the local automotive store on a Volt ohmmeter, couldn't you spend a couple of bucks more and buy a resistor. No, email me and I will send you one!
John

Derry,

Well, at this point it is hard to miss John's frustration. We must sound like a broken record, but for a reason. If you start with a perfectly good, healthy, working, alternator and perform your continuity/resistance, no-load test, I can absolutely assure you that the results will show that it is a perfectly good, healthy, working alternator. Other than that, all bets are off. There are conditions that would prevent an alternator from working correctly that your test will NOT detect. Period... end of story. There are even conditions that can cause malbehavior in an alternator that even the 1.0 Ohm load test may not detect, but as we have said, in those cases, the cost of the test materials far exceeds that cost of a simple replacement. The load test is cheap and simple to perform, well-documented (every old brit bike service manual), and even old Lucas themselves recommend it. No single test is definitive, but this one comes close. Truthfully, I am awaiting the arrival of the Star Trek Universal Test Scanner. It will put my company out of business and then I can retire.

Don

Well John,as you know,I'm completely with you on this issue,however,some people are never satisfied and have to go above and beyond. Why I know of someone who even cut a perfectly good Triumph crankshaft apart and readjusted the way the rods went up and down.Yep,some people are never satisfied!! Dick

The 1 ohm makes things easy. If you have alternator outputting across 1 ohm, and the volt meter across the resistor shows 11.1 volts then it's a fact that 11.1 amps are flowing thru the resistor and the output wattage is 11.1^2 X 1 = 123watts

Current^2 X resistance = watts

Voltage X current = watts

Makes the maths easy using a 1 ohm load.

A 100 watt resistor make be difficult to get, maybe better to use 10 X 10 ohm 10 watt resistors in parallel.

You would need a bridge type milliohm meter to check the resistance. Even a hand held fluke or similar good VOM will not give you even close to enough resolution to find if 20% of your windings are shorted. I think a stator is about 2/3 DC ohms. The leads of the hand held are almost as much resistance as the stator itself.
yes you will find if ALL stator leads are open but not if just one section is....without a bridge...
The 1 ohm load test is the cheapest and best stator test.
A 120vac test would work for the insulation check if a VOM does not find a short to the core, though this should not be a common failure without some evidence of burning.
I would suspect that winding to winding is the much more common failure and hardest to detect.
Probably an inductance test would give some clue.
I have data but I'd have to review it prior to posting it.

Not really, the hardest part would be to detect WHICH bobbin had the shorted turn.

Shorted turn(s) in a bobbin would ruin the "Q" of that particular coil, most shorted turns testers use a fixed frequency oscillator and the coil forms part of the resonant
circuit, if a coil has a shorted turn it stops the coil oscillating. Time to break out the soldering iron Derry.

turns tester

Trying to use a multimeter to find shorted turns is a waste of time, a shorted turn would only register milliohms difference.

Has anyone noticed that most multimeter leads have a resistance of 0.2-0.3 ohms?.

Dave Comeau wrote... well, everything he said. In the world of test & measurement it is critical to understand the limitations of the physics of the phenomenon your are measuring and the limitations (specifications) of the instrument that you are using to take the measurement. There is another thread running on the forums about ignition coils with various purported low measured values with fractional amounts of an ohm being tossed about. Now, as Dave has pointed out, these kinds of measurements require specialized instruments to be accurate. Any high resolution resistance measurement of small resistance values requires, at a minimum, a four wire test instrument with very, very good credentials. Everything comes to play here... lead resistance, cold junction errors, temperatures.

1.0 Ohm 100 Watt resistors are easy to come by, although, not at Radio Shack on a Sunday afternoon. Every major electronics part house on the web can ship you one quickly for a very reasonable price (see my earlier posts). The suggestion to bundle in parallel is a workable solution to the Radio Shack issue. Just keep in mind that your end result can be off in resistance by up to 20% with no effort, and thus your measurement will suffer the same lack of precision. Buy a precision resistor and be done with it.

Don

Duke, it COULD be done quite cheaply with a wheatstone bridge using 2 precision resistors and the old "metre" wire. I remeber at college we were given basic instruction in physical electricity with a metre bridge. We were then given a piece of wire and told to calculate it's resistance in milliohms then use a micrometer to gauage it's diameter and thereby it's resistivity in M^2, afterwards to make a "best guess" as to its composition. The item to measure was a paperclip

The maths was hard but the equipment was basically a 19th century metre wire, 2 fixed "standard" resistors, a galvanometer and a 1.5 volt pencell. When the bridge balanced is was p/q = r/s
and given that q and s were known, r was determined on the metre rule then p was calculated to a minute number BUT very accurately.

Andrew,

Your post crossed mine. I was responding to Dave’s post.

Yes, it can be done several different ways by anyone that is not math challenged and remembers their high school physics. As a practical matter, the point I was making was:

1) that, when folks bandy about very low resistance values measured with their trusty hand-held, and are comparing those to someone else’s values taken with their trusty hand-held on the other side of the planet (see the ignition coil postings) and expect them to match within 0.1 ohms (or even within 1 ohm), it is a fruitless exercise, and

2) rather than buy an expensive instrument to make milliohm measurements, or fabricate a bridge using precision resistors, you can buy a single 1.0 ohm 100 Watt precision resistor for around $12.00 US, conduct a load test, without the need to perform ANY math, and have the “go-no-go” test that Derry has been pursuing, over and done with.

Wow, was that a run on sentence or what?

An alternative is to calibrate a length of nichrome wire using a 6 Volt battery, an ammeter and a voltmeter using the procedure as outlined in Triumph’s workshop manuals from the day. Cheap and pretty accurate. Only ask yourself where, on a Sunday afternoon, are you going to scare up 12 feet or so of nichrome wire? Not in my garage.

Don

Just in case anybody is interested I made the following stator resistance measurements (based on three meters, two digital and one analogue for comparison.)

RM21 47205 two-wire 0.4 Ohms

RM24 47252 three-wire 0.9 Ohm between any two wires

RE Bullet four-wire Violet-Violet 0.9 Ohm
....................Orange-Yellow 0.6 Ohm

When I get my multi-speed rig up and running I'll make some more stator output measurements (loaded and unloaded) for reference.

A couple of 55/60 watt headlight bults with all the filaments conected in circuit would be alright, but you would need 2 meters, one for the voltage across the buld and one in series measuring the current drawn. ~Given that meter leads seem to have about 0.2 ohm resistance and the shunt inside the meter (say another 0.2 ohms) thats a total series resistance of 0.4 ohms, more accutrately a supply impedance of 0.4 ohms to take into account.

Not a lot of multimeters have an AC current range, I always check that they do have before I buy. Last meter I bought (new anyway) was from Maplins, it has AC current range (to 10 amps but handles 20 amps OK) and an RS232 output to log into the serial port on my laptop. Comes with the software for Windows from 3:11 upwards.

Maplin meter here

Thanks for that Andrew. Useful having the PC log when making measurements. I've been using the UT58C from the same range to make my measurements and this will read up to 20A.

Do the bulbs change resistance very much as they warm-up do you know?

Bulbs are PTC to stop them blowing. Good point Derry.

I've seem to have missed reading a couple of posts here previously for some reason.

Don like you I can see issues with measuring low-value resistances. As you know you need to be aware that test-set leads have a resistance and this must be subtracted from your test reading. But once you've done it a time or two it becomes second nature. Ah for the good-old days of analogue when you could zero your meter.

On another front I have quoted Suzuki's method of alternator testing so I thought I would do a quick check of other manufactures and models and see how they cope. Permanent magnet alternators only.

Of five manufacturers investigated (about 20 different models) three do their testing by measuring stator resistance and no-load voltage, the other two by measuring resistance only. Kawasaki were the most compehensive I thought and even spoke of loss of magnetism and alternator output caused by dropping your rotor on the floor! Email me if you want the details.

Ah, Derry, and how do you know the resistance of your test leads unless they are spec’d by the manufacturer? And if you have checked the manufacturer’s specs, look for the accuracy and resolution figures for very low resistances. Oh, and don’t forget added resistance from lead connections, cold junction effects, etc. You will be disappointed, unless you have spent a fair amount of money on your meter. I recommend, as a value purchase, an Agilent (formerly Hewlett Packard) 34401A. It runs sub-$1,000.00, but it is a benchtop unit with excellent specs.

Kawasaki has been very good over the years in writing their manuals. Even back in the late ‘60s and early ‘70s their manuals covered charging system troubleshooting fairly comprehensively, including a form of load testing.

See my earlier posting regarding “shock” damage to rotor magnetization.

Don

Bulb resistance changes a *LOT*, a cold bulb could almost be considered a short, but as the filament goes up 1000+ degrees it increases, as I remember it's almost a straight line slope on a graph if you plot resistance / temperature. At 12 volts in a 55 watt bulb would exibit a resistance of 2.6 ohms but at room temperature I don't know, only that it would be very small. You could ohm-meter a bulb.

I pulled this snippet from Wiki

<how do you know the resistance of your test leads>
Is that a rhetorical question Don?

On the face of it it seems that Lucas alternators don't fail very often by the feedback here from owners that have actually 'tested' theirs.

Indeed I had lunch yesterday with the retired owner of a local motorcycle repair shop and I quizzed him about the most common failure mode in his experience. Straightaway he said worn main bearings followed by loose mainshaft nut. So mechanical rather that electrical failure.

Anyway moving on I now have my 2HP motor so am just looking to build a test rig. Then I will be able to run tests from 1-5k ish rpm on and off load. It will be interesting to see how my results match Dave's.

I'm puzzling at the moment how Lucas and others list their output 'at 12V' or 'at 14V' as it implies a variable DC load. Did they just accept the ripple? Or is it just done for advertising.

Hmm, have you finished the motor rig up yet? I have a new (to me) aquisition, I was given by a motorcycling mate (free, gratis) a Fluke 863 graphical multimeter / osciloscope that didn't work, the reason? a broken off battery conector, so out with the soldering iron and a slip of berylium copper sheet.......... result now is a real fine piece of kit, although I am going to offer him the meter back as it must have cost over £1k when new.

Nope, I have just won an inverter to drive the motor on ebay so am awaiting delivery of that. I have made some scruff wretches of the lay-out and am currently puzzling over a motor driveshaft to rotor adapter. So another 2-4 weeks I would have thought.

I still haven't come to a conclusion about the variable load. I have found a couple of sources of low-value variable loads including one with a 300W rating that would cope with the 'nominal 200W' stator.

Derry.

I would have thought a belt driven shaft alongside the motor would be the easiest, maybe a variable ratio with stepped belt the same as an olde worlde lathe setup? A magnetic pulser pickup for measuring shaft RPM.


As regards loads, a few headlight bulbs with jumpers so you can vary the load, a 30 amp (100mV/A) shunt for current readings?

I have some 6" lengths of 1" diameter EN8 that would make an ideal "stubby" rotor shaft if mounted in plummer blocks, but I would need dimensions, threads, drawings etc.

True Andrew.

I originally thought of belt drive but a member here suggested an AC motor and varying the frequency. Now that I liked so I now have three-phase 2hp motor and an variable frequency inverter to drive it.

With respect to loads I am happy with fixed loads but my problem is how to vary the load so as to match Lucas and others when they plot their alternator output verses RPM. Lucas used 12v others use 14v.

Thank you for the offer of building me an adapter but that is now in hand. Are you sure you don't want to build your own system?

Hi Derry.

At the moment me,the wife and 2 kids live in an small house with only a little tiny patio. No real workshop space, just a little benchspace in the yard. We are moving house in July, to a *much* bigger place with 2 double car garages, so I look forward to a fixed shop.

I don't really need a test setup for alternators, I thought to offer to help you with parts and ideas.
I noted you had the VFO drive, I thought that a stepped pulley system would enable you to operate stuff outside the RPM range of your motor with a 1:2 thru 2:1 ratio, even going to a little chain drive if you were going into repairing E3L/Miller dynamos as well.

Theres a whole world full of exotic electric stuff out there on old cars, bikes and trucks and few people today know how any of it works and even less idea how to diagnose/repair it. Stuff like distributors, regulator boxes etc.

Your house sounds a bit like mine Andrew except I have a small garage attached and one son living with me.

I have connected my inverter to my 3-phase motor and I am impressed I must say. Quiet, smooth, controllable springs to mind but with a Lucas alternator attached we will see <s>.

By then end of the Easter break the whole thing should be up and running and I should have some more alternator output plots at various RPM available.

Derry, can you post some fotos of this motor of yours? I have an old alloy mounting ring for a Lucas stator, believe it's off a Cub, would be of any use to you? I think the next thing to buy myself (when we move house) is a milling machine, a grand old lump that needs a bit of a refurb but is as "cheap as chips" and comes with loads of stuff.

What do you do for a living? I manage a petrol station 28 hours a week, crap job, rubbish wage but free fuel and lots of free time, wife is/was apprentice blacksmith, expert bee keeper and gasket maker.

Hi Andrew thanks for the offer but I am almost there. I just need to get the whole thing bolted down to a piece of board tomorrow and I can start testing.











Interesting that you offered a piece of crankcase as a mount as I bid for the same on eBay this weekend (from a Velo). I only bid peanuts and was outbid by a Brazilnut.