You'll probably have a heart attack when you see the cost of a suitable accelerometer and its amplifier. If you do there is an alternative.

Balancing machines are either hard mount or soft mount. Soft mount lets the crank and support bearings float around, and you measure the accelerations.
 
Hard mount balancers rigidly (and I mean rigidly) fix the bearings, and then measure the loads in the bearing supports, using strain gauges, which are cheap. I've never used one of these, but I have seen a couple.

As far as software goes you can do the whole job in Excel. Getting the numbers into Excel will be the hardest part. There are VBA extensions that will read data in from the ports. Don't get too hung up on the calculations, you can do them by hand in about 30 minutes.

You might want to price out a two plane balancing analyser, this will give you all the instrumentation and calculations in one tidy package.

I strongly suggest you go and have a look at a succesful balancing rig so you can see how much work is required - not a huge amount, but not trivial either.

Cheers

Greg Locock

Greg
Thanks for the input. I have looked at both lvdt and piezo transducers. I already have an encoder to get rotation. Am now looking at accelerometers or strain gage. Looking to hook to a data acquisition card  to get data then dump data to excel spread sheet or just do it in visual basic. But still quite fuzzy on how to's. I just completed building a engine dyno using three dac cards using two strain gages along with 12 temperature probes, tachometer output, fuel usage meter, wide band O2 sensors. I keep thinking I'm thru, but I keep adding too this project. Want to get to the balancer before I spend all my money on the dyno.
thanks
reed grant

LVDTs won't have the frequency response or the resolution. Do you know how to do a two plane balance (ie the actual test procedure?)

If you get some readings down the track I'll be happy to crunch them for correlation.

Cheers

Greg Locock

Greg
No, I have never performed a test procedure. It appears that I will be using load cells from Omega. Trying to decide what weight cells to purchase. I surmise that a crank (big block chevy) with all the rotating weights will approach a max of 200 lbs. Do you have any idea how much out of balance weights will add to this figure? Will be using advantech dac card reading 0-10 volts channel. haven't worked out on the encoder. I have a accu-coder, but have been unable to find out what the output is. Also trying to decipher how to tell how much out of balance I will be and how much weight to add as well as where. I realize that the more out of balance the assembly is, the higher the amplitude. If I am off base, please advise.
I appreciate your kind help in knowledge and time.
thanks
reed

Typically we'd balance at 3000 - 5000 rpm, by choice. Slower is OK, but you get smaller signals.

If you are trying to balance race motors you might want to balance at a speed representative of what you'd see in a race, but maybe you should leave that til later, it gets complicated.

So, suppose you have 3000 rpm. Typical test balance weights are about 10 g (0.3 oz), so the forces involved are m*r*w^2

r~0.1m, w=3000/60*2*pi=300 rad/s, so that is 100N, or 20 lbf

You add balance to each end of the crank, when testing, and then drill out the opposite web when you have a solution. In prodcution it can be hard to find material to drill out,
hopefully your cranks are a bit straighter than ours!

www.sdrl.uc.edu/ucme571/balancing_lab.pdf

describes two plane balancing fairly well.

Cheers

Greg Locock

Greg
No doubt you are much more knowledgeable on dynamic balancing than I am. Shortly I will list what I am contempating on building. Please advise and make any recommedations.
Using two Omega load cells 0-250 lbs capacity
Using Advantech 818high gain dac board with conditioner hookup card to read mv signals at 100 khz
Accu-Coder encoder 360 cycles per rev
Durant products model 57820-400 counter to the encoder
Questions--- will the load cells be sensitive enough to pick up the oscillations if I mount them in rigid bearings.
I plan to wire all to PC to read load cell readings at every degree location. Will this suffice. I will dump this to Excel spreadsheet to see what shows??
I welcome all recommendations.
thanks
reed

Fairly well balanced stuff exerts less than 10% of it's own weight as "centrifugal" force on rigidly supported  journals.  A Cevy 350 V-8 crank with radically different components can be out of balance several hundred gram-inches as measured at the end counterweights.  Spinning a Packard straight 8 crank at a few hundred rpm while supported in its own journals gets it a-whippin' a visible amount in either a soft or hard bearing balancer, and tricks the balancer into reporting that it has a big heavy spot in the middle.  Since in real life it lives in a block and is supported and guided by several main bearings, whipping when running it on the end journals is not particularly realistic or useful test.  Most crank balancers I have known (Stewart Warner, Hoffman) support the crank (or any rotor) at 2 locations, and can be run at less than less than 1000 rpm, and I believe with good reason, to avoid being misled by dynamic distortion that >>may<< be the product of an unrealistic test situation.

Greg
You are right. Every one that I have heard of (Stewart Warner, Hines) supports at both ends and does spin at less than 1000 rpm; most around 650rpm. I will try to build a unit that will be adjustable up to 1000rpm max. Do you think that load cells mounted into the fixtures on each end will be sensitive enough to read out of balance situations. I build race engines that sometimes see upwards of 9000+ rpm. Accuracy does play a part into what I want to do. My theory is to use a v-block type apparatus with a milled slot for the load cells to be mounted in straight under each bearing support journal. the slot will be milled thin enough so that the any journal movement will force the v-slot to act on the loaad cell. give advise.
thanks
reed

That sounds fine.

Reworking the equation using 1000 rpm and using 100 gram inches gives a dynamic force of .100*.025*(1000/60*2*pi)^2, 25 N or 5 lbf. I think a 250 lbf gauge would have no problems with that.  

Thanks for the heads up on speeds Dan, I worked on a driveshaft balancer, and we try and balance as fast as we can, the current rig is 3000 rpm and we want 5000. This is because we are worried about the balance as we approach whirl, and the balance is different as the tube starts to bend. Typically our process gets a driveshaft down to about 7 g cm each end.

I've never balance cranks by hand, except as a running engine. We balance cranks, harmonic damper and flywheel separately before assembly, and then balance the final engine. I think we balance in production at 1200 rpm, but I could be wrong, that was 10 years ago.

Cheers

Greg Locock

Oops, make that 20 g cm each end for a driveshaft, I forgot about the conversion for inches. We then balance the whole assembled car and /that/ comes in at 7 g cm, tho we only do a single plane balance. (one of my standard gripes is that we can balance a whole vehicle better than our suppliers can balance the components!).

Cheers

Greg Locock

Hey Fellows
Made two responses but they did not posted.
will run this test and repost.
reed

Greg and Dan
Your inputs have put me much closer to the reality of building my balancer. Have found two load cells that will work with my dac card. Their outputs will be in MV. The encoder may present a problem. (Can't locate if it's output is mv or not). It is a Accu-Coder 716-s(360 cycles per rev) I have it conected to a Durant Products counter 57820-400. Don't know what the outputs are from this counter. Have you seen www.crankbalan.com. Is LVDT. Don't understand how it works, but who can argue success. I can make all the fixtures. advise
thanks
reed

Oops is www.crankbalancer.com

Hi guys,

Race car-- Are you trying to build a hard or soft bearing machine? In either case...couple of tips.. mount your transducers to pick up the unbalance in a horizontal direction. You need sensitivity not capacity. You can always reduce the force generated by the unbalance by reducing balance speed. Slower speeds will also reduce the likelyhood that the part will flex.

Use some mehanical stops to protect you transducers.

If you'd like more info post or e'mail me. Id love to see what you finally com up with.

Greg-- How straight is a typical crankshaft? If I were to spin a crank on one of the inner main bearings, how much should I expect the unbalance to change?

When balancing a drive shaft, Is the shaft under tension?
Where can I sketches or dwgs for the tooling? I'd love to give drive shafts a spin ;).

"How straight is a typical crankshaft? If I were to spin a crank on one of the inner main bearings, how much should I expect the unbalance to change?"

Ooer, don't know. I don't like the sound of that, but you are right, logically it should be OK, up to several thousand rpm anyway. I have a feeling we support two mains at each end, that may be to prevent the drilling machine from bending the crank. Ouch. Now I'm wondering if we balance between centres and just use the bearings as static support (it has been 12 years since I worked on that machine).

"When balancing a drive shaft, Is the shaft under tension?"

No, I don't think so. Oh, it might be a bit, if you have 2 UJs depending on how the flanges on the UJs are clamped to the chucks. Ours aren't, they have a slip yoke at one end so you can't really apply any longitudinal force. Inside the back of the flange on the UJ is a circular recess, this is the reference centre for all the assembly and balancing ops. It mates to a boss on the pinion flange on the diff.

"Where can I sketches or dwgs for the tooling? I'd love to give drive shafts a spin ;)."

Pay a visit to your local driveshaft balancer, our backstreets are full of them (slight exaggeration). The production machine for balancing driveshafts is a big hard bearing unit. I think it is made by Schenck, most balancing gear round here is. Our post-assembly rig is soft bearing obviously.

Cheers

Greg Locock

jclough and greg
Great Idea on mounting the tranducers horizontally. Was looking at what I would call a hard bearing mount??? I was planning on using a material like teflon (approximately 1/4 to 3/8 " thick), but am open to suggestions. I have secured the drive motor(1 hp DC motor with variable speed) and the encoder and transducer with a dac card. Am in the process of making drawings for the fixtures. Hope to start making within next couple of weeks. Can we talk about how to mount the load cells horizontally. Would like to get yall's(southern drawl) ideas or sketches on how to machine the fixture to hold the load cell horizontally. Crank will be spun by a flat belt around the middle bearing journal. Crank will be suspended/mounted on both ends in a Vsaddle. The tension will only be enough to spin the crank.
Thanks
reed

Hi Racecar,

Just some quick thoughts...

Maybe make your work supports in four pieces(nor including the fasteners or sensors). Make the base like an upside-down T and the bearing support like a regular T. Connect the horizontals ends of the two T's together using two I shaped pieces (springs)using set screws. Connect one of the the two I's at the each end of the horizontal part. You'll probably want to offset the vertical leg of the t's so there will be a space between them put you load cell. Yo may want to add a couple of pieces in this area to allow for adjustment. Would be beter to measure abd machine accurately because as you know adjustable stuff will go out of adjustment.


The work supports.... Vee's are nice for the electronics. low noise. Definitely my first choice. You can visibly see what condition they're in. But they also have some drawbacks. They wear quickly which can cause some sensitivity problems if they're worn. They also require lubrication and often create quite a bit if friction. This friction can increase the load on the drive belt to the point where it slips. This slippage can can cause your electronics to think the part is better balanced than it actually is.

Roller bearings are noisier to your filters but have lower friction and last longer than vee's.


Gotta go,

jon

jclough
Can you do me a quick sketch on your idea of load cell placement. After much thought, will have to agree that horizontal placement will be better. Want to get the best and highest amplitude reading without noise.
I appreciate the help that is coming from you all. This will eliminate many start up problems that go with a project of this type. My Daddy said do it right the first time. And y'all are helping me to do this.
thanks
reed

Racecar,

No guarantee's on the start up problems...We'll work through  them though.

Can yo deal with AutoCad?

jclough
I can't deal with autocad at home but I can carry to work and do. Yes I am familar with using autocad.
reed

Balance machine dangers -
"End drives" are a real pain in the neck because of UNAVOIDABLE centering variations and tight u-joints.  Beautifully balanced adapters do NOTHING to compensate for that. and Do not be complacent Just because the readings go down in response to balance corrections. Balance machines adhere blindly to the rule that IX force/motion phenomena are caused by unbalance. A suspicious nature, CAreful work with dial indicators, and proving runs re-indexing the components are the only protection against the devious pranks that balance machines enjoy pulling on their human operators.

Be sure and run the belt drive perpendicular to the direction of measurement.  Otherwise the once-per-rev variation in belt tension due to small drive journal eccentricities will talk the balance electonics into convincing you to drill holes that should not really be there.  Ditto for any stops or end thrust device.  Balance machines adhere blindly to the rule that IX force/motion phenomena are caused by unbalance. ( I repeated that on purpose.)  Please also run thru the math once or twice comparing ballpark unbalance resulting physical runout to unbalance before becoming too smug about the attainable and attained residual unbalance results.

Still need help on how to get the load cell mounted perpendicular to drive. I thought I was ready to start, now 2nd guessing mounting of load cell. Any help?
thanks
reed

Anyone know how to post an AutoCad dwg here?

Just some ideas and observations. How about floating the vees or roller blocks on air (most workshops have a compressor.)  Both underneath and end float can be air. The machine I worked with some years ago locked down the bearing blocks until the machine was up to speed (to save the sensors from the hammering of startup and in case the shaft was very out of balance.) Before you start drilling, turn the shaft around and check if you get the same results. How about an adjustable strobe light triggered by the load cells to check the out of balance plane?  This could be very simple electronics and no encoder or computer needed.
Let me know how it ends up as I want to make one too! Thanks, derekrwhite@yahoo.com

jclough
send dwg by email if need to.
thanks
reed.grant@att.net