AustinT

I finally got the dyno done . It took awhile to get all of the bugs worked out and I went through a couple of different designs before I got it to work. I don't have a picture right now but have some numbers that I cant make any sense of. The Loadcell is place in a chain that holds the car back. Only the chain is holding the car so all of the force is accounted for. I backed the car onto the dyno and brought the throttle up and applied the dyno brakes. 769 pounds of force at 2370 rpms. It does not seem Like a lot to me. I thought I would see closer to 2000 lbs of force. I know that my calibration is correct for the load cell. The rpms were taken at the engine. Trans is a t56 in 4th gear 1-1. rear ratio is 4.10 the back tires on the car are 245-45-17 All of the Data was logged in a laptop so I could go back and compare numbers. Do these numbers make any sense?

JP84Z430HP

The main thing you're going to be looking for would be trends. I think it has been stated enough around here that the inaccuracies from dyno to dyno, day to day, make them pretty useless, unless you use the same one, and you can't rely on the numbers to be actual HP. For your purpose, just compare one run to another and see what the change is in load on the load cell at a certain RPM compared to your previous run.

I wouldn't worry about an actual number for horsepower though.

Sounds like it could be a useful tuning tool. Is there any way to feed the loadcell info back into the ecm? You could then have it spit that out in the datastream, and the runs would be EASY to compare!

JP84Z430HP

Oh, and I'd loave to see some pics of this thing!

AustinT

Thanks for the replies guys. The loadcell is converted to a 0-5 volt signal. I use the innovate wideband to record the rpms of the car and the force. I calibrated the loacell into pounds so it actually gives me a pounds and rpm readout on the laptop. I agree trends are the most usefull to compare but I would like to use both numbers . I can download the data into excell and write a formula for hp. Sure maybe My horsepower could be off from another dyno but It would give me a good number to compare with. However I think the data is really close to the actual output of the car. the dyno has no pivots or bearings that would affect the torque. All of the pulling force is accounted for in the loadcell.

I have some pics on a disk . They are with the first design. At first I used a trailor axle for the dyno. The right and left tires are independed of each other. My car has a zexel torson posi. Most of the time both dyno tires would run even but sometimes one tire would do all of the turning. Must not be a true posi? The dyno design that is working uses a semi driver axle. The ring pinion and spider gears are removed and the axles are tied together with pipe. Both Dyno tires must turn the same rpms so if one dyno brake is a little more tight than the other it does not matter. Ill get some pics up as soon as I can

AustinT

I just noticed your username. Do you have a zz430? The car I am using on the dyno is a zz430

RBob

iTrader: (1)

I did some quick math and came up with 218 ft/lbs of torque. At 2370 RPM that does sound correct. I guess-ti-mated that the tires are 28" in diameter. Here is what I did:

769 pounds of force divided by the gear ratio: 769 / 4.11 = 187

Then the tire radius as a torque arm: 14" radius divided by 12" (a foot) = 1.1666

Multiply 187 by 1.1666 for 218 ft/lbs of torque.

RBob.

BJM

The calculation above is correct. It sounds oddly low though, its only 99 hp. Is the chain level with the plane of the car, or at least the direction the car would move if allowed to roll, is the car directly on top of the rollers as opposed to being ahead or behind of the top? Does something stretch under load to allow the car to move off the top of the rollers?

A couple of pictures might clue us in.

AustinT

The chain only has a slight inclination. I tried to get the chain as level as possible so none of the force goes down. Nothing can stretch and allow the car to move forward. When I tried it, it was not directly on top of the dyno tires, I dont know if this would make a difference or not. Sitting in the car with the loadcell attached I can push on the ground with my foot and change the loadcell reading. It is very sensitive

kdrolt

I assumed 27" tire diameter, so I got 211 ftlbs and 95 hp. Those would be the numbers at the tire-road interface with zero loss anywhere in the drivetrain.

But the drivetrain has loss. If we assume 15% total loss (manual trans, differential), then the corrected numbers at the flywheel would be 248 fw ftlbs and 111 fwhp. A ZZ430 should be making well over 300 fw ftlbs at 2400 rpm, so there is still something missing.

I agree on the other questions (is the tire directly atop the roller during WOT, is the chain level).

Then there is the engine (ZZ430 crate motor). Are you running EFI on that ZZ430 engine, and has the ECM/PCM been retuned with a wideband to make sure you are getting all of the potential power in ther engine? Many people assume that if they buy a crate motor (like the ZZ430) that it will make the advertised power when it gets installed into their car -- forgetting that the fuel delivery has to be adequate (large enough fuel pump), and that the powertrain control of the injections & ignition has to be retuned from 3rdgen factory spec. So there is the chance that the engine isn't making close to the correct numbers because the tuning isn't there.

JoBy

245-45-17 => 12.84" radius = 1.07 feet.

769 lbs * 1.07 ft = 822.83 lbs ft at the wheels.

822.83 / 4.10 = 200.7 lbs ft at the crank.

200.7 lbs ft * 2370 rpm / 5252 = 90.5 hp

Corrected with 15% loss. ( 1/0.85 = 1.1765 )

200.7 * 1.1765 = 236 lbs ft

90.5 * 1.1765 = 106.5 hp

AustinT

Here are the pics of the dyno.

http://www.josephthompson.net/austin/

RBob

iTrader: (1)

Awesome.

Have you been able to check the load cell? Such as by hanging a known weight on it.

RBob.

Grumpy

Coolest thing I've seen in years!.
Doc's making ALL the lil guys take notes, and study the details.

Neat..

SBNova

I am sure you beleive that axle looking anchor is strong, but how about strengthening it? It may removed some flex with may make your results more consistent, and it would be safer too. Neat idea though, congrats!

wanarace

Just going through a few things in my head. The basic idea is to use the semi axel as brake, then see how much force can be made on load cell correct? Simple enough idea. What I think the tricky thing would be, is the exact proper ammount of brake. How are you regulating braking force? Obviously there is a fine line between too little brake and too much and it is constantly changing as the power changes. I believe the low numbers could be due to not enough braking force.

Maybe this is how you are already performing the tests, but here's how I would go about it. Pick an RPM, say 3500, bring the car up to that with no load. Then start to add braking force, while adding throttle to keep the motor at 3500. There should be a point where max force is reached on the load cell, then it starts to drop off. The only problem is I really don't see any way you could maintain the exact maximum braking force over the entire rpm range.

Another way to make the dyno, would be to take the weight of the rollers and measure their acceleration. That would in turn tell you the force needed to acclerate the rollers. I believe this is how dynojet dynos work. Really wouldn't be that hard. Get everything weighed and use a toothed wheel and VR sensor to measure acceleration.

Sorry if this has already been disscussed in another post.

Steve

wanarace

Ok I found some of the old posts and see this has been disscued in great detail. The thing with load dyno is you need to know the exact force created by the load in order to calculate real world horsepower. Think of it this way, if the load on the rollers was really light the car could easily spin them with out much force. The load on the chain would also be small. As the load of the rollers increases, the load on the chain also increase, even though the car always puts out the same power.

It will still work if you don't know the load, but you'll never be able to calculate real world horsepower. You'll get a number like you have now. It will work great to measure any gains or losses made from tuning.

In order to do comparisons the load must remain the same or change exactly the same way with every run. Unfortunaltey I don't think the brakes will provide you with enough consistency. They change too much with heat and wear.

But that is still one sweet setup
Steve

steve8586iroc

I thought this thing was going to be portable.

Steve

AustinT

I think I will check the load cell again just to make sure. when I calibrated the cell I used a cell on a cement silo to get my weight. I used a welder as a test weight. The welder weighs 633 pounds. The loadcell is rated at 2500lbs. 633 is about 1/4 of the total capacity of the cell it should be close. If I wanted to get it closer I could use a weight that is closer to the total capacity of the load cell (2500 lbs)


Brakes are applied using a regulator to feed the semi brakes. I do have a little trouble getting a nice smooth load, it seems to come on all at once. I am going to try another regulator, maybe a semi foot pedal to get a more gradual load. I am able to get plenty of braking force, I can stall the engine if I want. The torque number increases as the engine load increases.

I might be looking at this wrong but the load on the chain is the load on the car. I will also monitor the TPS sensor when I switch to fuel injection so I know that the throttle is constant when I do my comparisons



It should not matter if the brakes fade a little other than I dont want to glaze them. If it takes 40 psi to reach a load one one run and 50 psi to reach the same load on the next run it doesnt matter since Im not comparing the psi required to apply the brakes. If I know how hard the car will pull at a given rpm I should know the HP, Like you say my numbers dont seem to be right . The motor is rated at 430 hp, but this high altitude chokes it a little. I would like to see hp numbers in the 300-350 range.

I really dont why the numbers are so low The pulling force should be higher, I tried to keep everthing as simple as possible so no force would be lost in pivots and bearings. The chain should measure all of the force that the car is making since it is the only thing holding it back. Also the chain is really close to level so only a small force could be going down into the dyno wheels. On the front of the car I have a binder that keeps the car from falling off the back of the dyno. It trys to pull the car forward. before I even started the car the loadcell showed 300 lbs of pull. I zeroed out the readout so I can start at zero.
Thanks everbody for your input this is great

AustinT

That would be nice but I wanted to make this first one low budget. I have 235 into this one, Not bad if it works

Low C1500

That Dyno is great, nice job!

wanarace

How are you performing the tests? Set the brakes to 40 psi, jump in the car a see what the max load is?

Here's what I am trying to say, I could be completley wrong. If you start with a light load the car will have to work however hard to over come it. As the load increases the car has to work harder, more power. There comes apoint where the load is too great and car will no longer be able to over come it. The instant before the load is too great is the point of maximum power.

Think I had an apifany. What the load cell is measuring isn't the force of the car, but the force needed to over come the resistance of the rollers. So in your case it is taking 110hp to turn the rollers.

Then again I just might be crazy.
Steve

JP84Z430HP

So, even though the car is making more power, that is all that is needed to do the work of overcoming the brakes, to turn the rollers. So, basically, he needs to apply more brake to get it to use more power????

BJM

Typically you only care about WOT power. You really want to be able to apply the brakes progressively. Ideally at around 2000 RPM you would slowly add throttle while keeping the RPM fixed by adding the brakes. Once at WOT you slowly release the brakes to allow the RPMs to rise until you reach the max RPM of interest, then you let off the gas. As long as you don't exceed and acceleration greater than say 500 RPM/s you won't lose torque to interia effects. Actually that would be a cool study is to rev it up slowly and really fast and see how much torque is absorbed simply revving up the drivetrain since that big axle doesn't weigh a lot.

You should have a really good set up since you can datalog at a pretty high rate and each pull will be pretty quick. An inertial dyno like a Dyno Jet cannot alter its load at a given speed like yours can.

I am guessing your load cell calibration is messed up somehow. Perhaps you should pay for a normal dyno pull to get a good baseline.

You might have to remove some brake pad material to make the brakes less grabby.

wanarace

Ok let me try this again.

Say you have rollers that take 100 pounds of force to turn at a constant speed. In this case we are using a car to produce the hundred punds of force. The car is not moving and attached to a pole. If the car is producing the 100lbs to move the rollers but not moving there must be 100 pounds of force holding the car in place, this is measured with the load cell. So that is ok. We know the car is making 100 pounds to turn the rollers. But what happens if the car starts making 800 pounds of force? It still only takes 100 pounds to turn the rollers and that is what the load cell will measure. The other 700lbs of force is used to accelerate the rollers. And the load cell should measure the increase to 800 pounds. But the rollers continue to accelerate. So 800 pounds of resistance is need by the brakes to stop acceleration. It is a fine balancing act to increase brake pressure while controlling acceleration. There is no easy way to have controlled acceleration, with out expensive equipment like a water brake. Why would a load dyno like a Mustang go through all the trouble of using a calibrated water brake to create a know ammount of resistance on the rollers, when all they could have done is had a set resistance and measured the force of the car pulling? Seems like a big waste of money to me. Shops are buying $100000 dynos when the could be spending $5000?

I just have a hard time believing it is as simple as everybody is making it out to be.

Hey and if I am completely wrong atleast I will have learned something from this.

Steve

JP84Z430HP

I understand the precision that goes into a production dyno, but something like what has been made here is fine for back to back comparison stuff. With the lack of consistancy, the portable chasis dynos aren't that much better in my book than what has been done here.

As for the cost of a commrcial dyno, yeah, the water brake isn't cheap, but I have to believe that a bunch of the money goes into software. Think about it, if windows was on only a couple thousand machines, they would have to charge much more for each copy to cover development costs, or cut corners on features... Same goes for the dyno harware.

wanarace

JP84Z430HP damn your fast, I edited a little of what I said above.

Short restate; Impossible to do a WOT run without precise control of acceleration. Will work for runs without acceleration but no consistency.

Steve

AustinT

Tried the dyno again tonight. I was a little more brave tonight and ran it at wide open throttle. Yesterday I must have not reached Wot and I also did not pull long enough to get a good reading. I forgot to record the run tonight but here is what happend. With the car off the loadcell reads 242 lbs of force. this force is the force between the binder pulling the car forward and the chain (loadcell) holding it back. Next the car is started, and the brakes are released. In fourth gear I let out the clutch and the car idles at 750 rpm while it turns over the dyno. The loadcell at this point reads 300 lbs. The additional 60 lbs is mainly from the drag of the brakes. even with the brakes off they drag slightly. Some of this could also be resistance in turning over the tires but not much.
In the drivers seat I can control the air going to the dyno brakes with a regulator. I slowly bring up the psi and the throttle at the same time. I let the pull last a little longer today and the rpms went up to 3900 while the loadcell reading said 1050. After the deadweight is subtracted the car pulled 810 lbs at 3900 rpms.


I still dont understand why the numbers are so low. As mentioned earlier in this post I am going to test the loadcell again. The only other problem I can see is getting a gradual load. It seems like the brakes want to come on all at once. It might be the regulator and it might be resistance in the brake assembly. Im going to try another regulator and the footpedal off of a semi to see if the load can be applied more gradual. If this doesnt work Ill replace the aircans with small hydraulic rams. The hyraulics would not compress like the air so the brake push could be feathered

onebinky

iTrader: (1)

How are you lubricating the bearings in the axle? Are they grease packed or oil bath? If they aren't getting proper lube they could be heating up ALOT and skewing the results some.

Very cool idea though, good job

RBob

iTrader: (1)

If the binder/drag load of 300 lbs is kept in the calculation would the numbers be closer to what is expected?

The reason I ask is that maybe under load (WOT, 3900 rpm) the binding system is slacking up. With the pull on the load cell, chain, anchor point, is it possible that the binder is no longer under tension? If it isn't then that portion of the load shouldn't be subtracted out. (it would give about 300 ft/lbs of torque).

RBob.

JoBy

That makes perfect sense .... the 60 lbs drag should be included in the braking force, but the 242 lbs in the forward chain should probably not be.

245-45-17 => 12.84" radius = 1.07 feet.

769+242 lbs * 1.07 ft = 1011 lbs ft at the wheels.

1011 / 4.10 = 246.6 lbs ft at the crank.

246.6 lbs ft * 2370 rpm / 5252 = 111.2 hp

Corrected with 15% loss. ( 1/0.85 = 1.1765 )

246.6 * 1.1765 = 290.1 lbs ft

111.2 * 1.1765 = 130.9 hp

contactpatch

Awesome setup.

contactpatch

Perhaps stiction from the front wheels is reducing the reading.
Put two blocks of (melting)ice under the front wheels.

AustinT

The bearings are lubricated with oil. The differential case was put back in the axle without the gears. Then the axle was filled with oil so the bearings are halfway under oil



According to JoBy's numbers the hp is still low. Im not sure what is going on




The loadcell was calibrated before installed at the dyno. A known weight was placed on the cell 633 pounds to be exact


I could use a torque wrench and weld a nut at the end of the axle

I dont know the size of the tires off the top of my head, proabably around 40 inches in diameter, Ill measure today
\ There isnt a torque arm on the dyno, Originaly I was going to put the whole dyno axle in bearings and weld on a torque arm. This design should be better becuase it removes the lost torue due to bearings and pivots. The torque measured is the force it takes to hold the car from driving away. The torque arm would be the radius of the car tires, 13 inches

Again Thanks to everone for your help

AustinT

Yes the chain is attached to both sides of the diffential on the axle tube. The height is level back to the load cell.

11sORbust

Shouldn't the load decrease with rpm. I mean the brake's psi should decrease as rpm rises..?

BJM

I changed my mind the load cell is fine.

I know where your load is going. It clicked when I realized you have the front tied down as well. You have preloaded a system with two springs in parallel (the load cell value you zero out is the preload). Any change in load on the load cell involves changing load in the front tie down as well. Unless you can calculate the stiffness of the front and rear tie downs accurately you don't know what proportion of load change went into relieving the preload on the front versus increasing the preload at the rear. For simplicity, assume the chain is twice as stiff as the front straps (kfront=0.5krear). Both have ~300 pounds in them before you start. Push the car forward 1 unit, the rear chain load goes up by K*1=K. The front load drops by 0.5k because they changed the same amount in length. Your load cell registers the value K in load change. Another .5k was there too relieving the front load but goes unmeasured. You just lost 33% of the car's total actual thrust. Those flexible straps are very non-linear, you can see them stretch at first and then they get much much stiffer so guestimating is tough. The stiffer the front strap is, the more thrust is missing from your load cell. The angle between the front and rear tie downs adds more complication.

On a normal load dyno (not a dynojet), the car is held still while the reaction torque trying to roll the truck axle housing is tied to a lever arm and the load cell measures the load in the line holding the lever arm still. The axle housing has to be mounted on bearings to allow it support the vertical load but not restrain the housing from rotating.

Instead you want to measure thrust, which is fine but you either need a second load cell in the front or the simplest thing to do is to make sure that the front one hangs slack during the pull. To do that you cannot use a strap, you need a stiff chain so that it stores no energy when pulled tight. As soon as the car shifts forward and the front line goes slack its preload is gone and the whole load is on the load cell.

Use the flexible strap at the back if possible, it can stretch all you want as long as the tries stay on top of the rollers (shifts up to 1" are fine). Put the rigid chain at the front and it will hang right away once the car moves slightly forward.

AustinT

BJM , thanks I will give it a try and report. The front tiedown is a cumalong ratchet. It is pulling the car forward against the rear chains. I can remove it completly and chock the back of the front tires. The only reason I have used it is so I dont fall off the back of the dyno wheels. Also I realized today that the chain is not parallel with the car. The chain runs level but the car is tipped forward. The dyno wheels are around 8 inches higher than the concrete the front wheels rest on.
Thanks again, seems youve had your share of physics?

BJM

I am a Mech Eng and besides I love stuff like this.

I thought about the angle of your car and I don't believe it matters. What matters is the that the plane formed by the contact patch of your tire is parallel to your load cell chain. Your car could be vertical hanging nose down but if the tires were still able to be on top of the rollers, they still pull directly away from the load cell, all would be fine, weird but fine.

You could use chocks with a very slack strap on the front that only goes tight if the car really moves backwards for some reason, don't want to harm your car.

AustinT

Im a couple of years out from getting a degree in ME. I wish I was done, it would really help me out on projects like this one

Does it matter where the tire is positioned on the dyno tire?

Can you explain again how the preload is affecting the numbers please. Why does the cumalong pulling forward skew the results.

I know you explained in your previous post, and im not questioning your knowldege. It just seems like that force would not matter since it is pulling forward and not down


I think I know why the dyno brakes are not easy to control. On the cement trucks we have 3 axles that can be dropped when the truck is loaded. Large airbags push down the axles so the amount of weight they carry can be adusted. When the regulator is adjusted it takes a long time to see the difference at the bag. I am using one of these regulators to control the air to the semi brakes. I belive it is too slow so It is hard to know when to stop turning the regulator. Today I took a foot valve off a old truck to try. The brakes should feel like the braking system on a semi.

AustinT

Today I took the cumalong off the front of the car. The load cell read 180lbs. The 180 lbs on the cell is becuase the car tires are a little forward than being centered with the dyno tires. Tried another run and my results do not change other than the reduction of preload. The car has to make more than 130 hp at the wheels. Im going to do as suggested, check the cell for accuracy.

JoBy

Are you sure that it makes more than that at 2370 rpm?

BJM

The last post has me confused now. 130 hp at 2370 RPM is 288 ft.lb which sounds pretty good. Maybe I am confusing 2 runs.

As far as the come-along goes, you need to know where all the forces are going. Assume the post behind the car does not move and the ground does not move where the come-along is tied. The car is restrained front and back with "springs". Ignore the preload for now. Slide the car forward and the rear "spring" must see an increase in force, and you are measuring it. The "spring" at the front must see a reduction in force, but you are not measuring that one. You also do not know the relative stiffnesses of the 2 but we assume the front one is softer. That implies that the for a 1lb increase in the load cell value that the front one drops by a smaller amount, but its not zero. Part of the car's thrust pulled the back one and part of it relieved the front one. You need to know both.

Now add in the preload, that determines how far you can shove the car forward before the front line goes slack, maybe it has in every run, maybe not. Once it goes slack, all the car's thrust shows up on the load cell assuming the tires are near the tops of the rollers, the chain is level, etc. Until that point however, you are sharing the load in some proportion front to back and you do not know what the proportion is.

AustinT

BJM, tried the dyno again today. The front cumalong was removed so the chain in the rear was all that was attached to the car. The results were the same as before. The load cell read 180 lbs because the car is a little forward than being centerd on the dyno. I also took the load cell off the dyno and checked the numbers. It is accurate. I dont know where else the force can be going. The last thing I can change is centering the car on top of the dyno. Maybe Joby is right , maybe the numbers are right. I know Im not getting the 430 hp the motor is rated at but it should be in the 300-400 hp range. I have another 89 camaro with a stock 305 rated a little less than 200 hp. The camaro with the zz430 feels like twice the car.

Pablo

iTrader: (1)

wouldnt you suppose that 300-400 hp number would occur at a much higher rpm than 2300?

My car probably makes like 75 hp at 2300 rpm
then again its no low end torque monster

JoBy

This should be subtracted from the load cell value.




1050-180 = 870 lbs at 3900 rpm

245-45-17 => 12.84" radius = 1.07 feet.

870 lbs * 1.07 ft = 930.9 lbs ft at the wheels.

930.9 / 4.10 = 227.0 lbs ft at the crank.

227.0 lbs ft * 3900 rpm / 5252 = 168.6 hp

Corrected with 15% loss. ( 1/0.85 = 1.1765 )

227.0 * 1.1765 = 267.1 lbs ft

168.6 * 1.1765 = 198.3 hp

---------------------------------------

The first run ( 769 pounds of force at 2370 rpms ) Did you subtract anything from those lbs numbers?

267.1 lbs ft and 168.6 hp at 3900 rpm

AustinT

Yes I subtracted the preload. I think the rpms would have been a lot higher if I would have stayed into the throttle and let things stabilize. The 2nd run with the 3900 rpms was accurate. I held it at WOT until the numbers were stable.

Is 15% average for the loss in a manual transmission drivetrain?
4th gear is 1-1 right?

Tomorrow I am removing the aircans from the dyno. They cannot be controlled very precise. The Footvalve off a semi was better than the regulator but did not work like I had hoped. When i watched the aircan and slowly applied the brake pressure it would not move. Slowly I would apply more pressure and it would move all at once. I was going to go to hydraulics but it would be easier to go completely manual. What I believe will work is tie both slack adjusters together and weld on a big lever. One guy can sit at the back of the car and slowly push down the lever. Here is a link to a website showing how semibrakes work.
Thanks again everbody for your numbers and help

Low C1500

Not sure if it was mentioned earlier or not, but what about friction loss at the contact point of the tires??

AustinT

good point, tire flex could eat up a little of the load. I have 50 psi in the dyno tires. Ill try bumping them up to 100 psi.

AustinT

Here are some pics of the first dyno built. It was only used a couple of times, and I could see it would not work. A semi trailor axle was used in place of the driver axle I now have. The dyno wheels are not tied together on the tralior axle. Even with the brakes on each dyno adjusted real close they did not have even friction. The car has a zexel Torson HD posi. Most of the time both tires would spin together but when a large load was applied one tire would spin twice as fast, like a non posi rearend.
http://www.josephthompson.net/austin_pictures/

JoBy

I think you should add computer control to the brakes on the dyno axle.

To make a good dyno measurement you have to measure at different RPM.

You have a computer for logging now? right?
Measure engine RPM and the load cell.

You let the computer control the brake force. What you should do is start at about 2000 rpm setpoint. Apply WOT and let the computer adjust the brake force so the engine stay at 2000 rpm. Then on a command (still WOT) the computer adjust the brake force and let the RPM rise slow and under control. Perhaps 5 seconds for every 1000 rpm. You will then reach 6000 rpm in 20 seconds, all under WOT.

From the log file you can get the load for each RPM and plot a tourqe / hp graph.