I finally bit the bullet

[Kartorbust]

That's quite discrepancy at either end.

 

[Denny]

I was always taught dyno hp numbers are basicly bull. Just a number that can be manipulated. The real value of a dyno is as a tuning aid, like a oil pressure gauge, temp gauge, afr gauge, vacuum gauge. It is just another tool to measure engine performance and make sure the changes you make are heading in the correct direction. Final or additional tuning was done on the race track. with the cars we never worried about dyno numbers just time slip numbers. You can take one engine and run it on 10 different dynos and end up with 10 different sets of numbers.


Denny

 

[KartFab]

I was always taught dyno hp numbers are basicly bull. Just a number that can be manipulated. The real value of a dyno is as a tuning aid, like a oil pressure gauge, temp gauge, afr gauge, vacuum gauge. It is just another tool to measure engine performance and make sure the changes you make are heading in the correct direction. Final or additional tuning was done on the race track. with the cars we never worried about dyno numbers just time slip numbers. You can take one engine and run it on 10 different dynos and end up with 10 different sets of numbers.


Denny

I agree 100%.

Here is a direct quote from the dyno guy that is employed by one of (if not THE major manufacturers of small 4 stroke industrial engines)

"A dyno at a dealer or recreational level is a tool and as such repeatability is going to be all one should be after. When someone talks a tenth or two gain most of the time there is more error in the test standard and definitely more error in the correction factor software than allows for this level of accuracy."

But I also believe that there is a standard method that can be followed to fix this (but it aint gunna happen).

The problem is the vast majority of people buying parts and engines out there, including those in the karting scene, will buy based on hp numbers alone. The bigger the numbers, the bigger the sales. The companies where dynamometer operators have lawyers breathing down their necks will produce the "true" hp numbers. That doesn't exist in the karting world.

There are no consistent materials or methods or actual verification of calibration with inertia dynamometers in the karting scene. If there were and people were held accountable to those materials and methods, everybody's would read the same numbers.

There is a standard for engine horsepower rating that must be followed by engine manufacturers. The standard is SAE J1349 and is used for net and gross continuous horsepower ratings. Honda uses it, and briggs uses it. I know this because they state that they use it. Honda reports their NET hp numbers, and briggs reports their GROSS (no muffler and air box) hp numbers with all of their industrial 4 stroke engines.

http://www.honda-engines-eu.com/web/eec-public-site/consistent-method-of-rating-engine-power

Not sure about lifan and duromax, but i have a sneaky suspicion that they are not since their numbers skew high, and they dont have class action lawsuits to worry about. Briggs, honda, kawasaki, tecumseh all have settled class action lawsuits as late as October 2018 in court over this inflation of hp numbers.

So what happens when these companies actually follow the numbers? You can see results like this for STOCK engines:

Given: displacement and a 1 hp per 32.5cc produced, you can divide displacement by 32.5 to find a ball park figure for any small, bone stock, all original, industrial 4 stroke engine will make at 3600 rpm.

http://www3.telus.net/findNchoose/engine_cc_to_hp_calculator.html

212/32.5= 6.523 calculated (predator states 6.5)
196/32.5=6.030 (honda gx200 states 5.5 but the power curve shows maybe 5.7-5.8ish hp)
389/32.5=11.96 hp (honda states 11.7)
420/32.5=12.92 (predator states 13 hp, and duromax states a whopping 16hp... yeah somethings not adding up here with duromax... a quick calculation here for duromax's MAX torque rating, you find that they actually fudged the numbers in the hp=rpm*tq/5252 and you see that duromax is basically saying that their max torque happens at 3600 rpm, which we all know is not how any small engine works. Max torque happens at 2200-2500ish rpm, so there is at least a feasible explanation as to why duromax is knee deep in BS, they cant or choose not to use a hp formula correctly by substituting 2500 rpm with 3600 rpm for a rather juicy (and completely false) hp number.)

So why does this matter? I have seen more than 1 or 2 people comparing the duromax 420cc to the predator 420cc and when you see about all else the same, one makes 13 hp and one makes 16 hp, well, are you going to blame the guy that thinks 3 more hp is the deciding factor here and buys the duromax? Spoiler alert, im fairly confident that the duromax wont make significantly more or less hp than the predator, and you have been duped into thinking it has. THIS is why I am frustrated.

Going back to the relationship between cc/32.5 might not be a linear relationship (e.g accurate at all cc ratings since there are so many other variables) but from my initial review of honda, predator, and briggs engines... the results are very clear. They all match pretty well. So now lets get on to modified engines:

Here is a number I trust for a 204cc 8.5:1 compression engine with a 22mm roundslide carb on 87 octane gas. Unsure of header or air filter, but it is according to SAE correction: just under 8 hp at 5000 rpm... see attached screen grab of page 6 here:
https://www.briggsandstratton.com/content/dam/briggsandstratton/na/en_us/Files/pdps/ms5701.pdf

So lets just stop right here and think for a minute. briggs animal makes just under 8 hp with a nice carb on pump gas and 29 DBTDC with PVL ignition. Also makes just about 10 ft lb torque MAX. okay... so how do we get inertia dynos running a predator 212 spitting out 14 ft lbs and 9.2-10 hp with a carb that flows less, inferior timing, and about the same displacement/cam profile? THIS is what i am getting at here with the HUGE discrepancy. You cant just take a stock predator 212, change out the jets and whamo it makes insane 54% jump in hp magically. An old engine builder once said, remember the days when the 8.1-8.8 hp briggs flathead stockers claimed 11-12hp? Now we have the 11 hp OHV stockers claiming 15-16hp (sarcasm).

new LO206 204cc engines with 29 DBTDC, .255 cam, 8.5:1 compression and 87 octane gas, and you get 8.8 hp
http://www.briggsracing.com/racing-engines/206

bump the compression to 9.5:1, 91 octane fuel, .308 cam, and you get 11.5hp (world formula)
http://www.briggsracing.com/racing-engines/world-formula

So why is it that the honda gx200 doesn't say 6.5hp anymore? Because it doesnt make 6.5 hp, and it never has (even though it may have said that on a different dyno that cannot actually conform to SAE J1349). It actually might make between 5.5-5.8 net horsepower in the real world per SAE J1349 methods for net horsepower. Take off the muffler and airbox and you do see a jump in hp.

UPDATE:
I was able to determine the optimal engine inertia compensation of a box stock 196cc engine. Things are looking good right now as far as repeatability. That was one of the major concerns i had, and still have for this water brake. But I believe I will be able to verify torque to raw torque with the weights (calibration for accuracy) as well as pre/post viariance in zero due to strain gauge torque drift. Best practices for this is to zero before a pull, but I want to take it a step further and zero before pull, then verify torque reading immediately after pull to see if any drift (even .01ft/lbs) occured, and potentially discard that reading until i can have a pre run zero and a post run still read zero (this is to account for the significant con of a strain gauge in temperature change, this is already compensated for by zeroing out before the run, i just want to make sure it still reads zero immediately after the run). This is in addition to keeping the instruments on for 30 minutes prior to making a pull https://www.dynomitedynamometer.com/dyno-tech-talk/repeatability.htm as well as maintaining head temperature the same and sweeping in the same up/down direction and at about the same rate.

Still will have to think about hp measurements under low rpm (lugging) and high rpm (inconsistent unloading of water as max rpm is approached). I believe this is a limitation of the water brake as far as repeatability goes, or operator error (could just be my dumb self not being able to be smooth with the water loading valve).

So my questions before buying an inertia dyno are as follows:

• Can you replicate the hp/torque curve of the same engine on a water brake?
• Can you replicate the hp/torque curve of your "average engine builder's dyno run" on a water brake without fudging the numbers?
• Are we comparing apples to oranges already? Probably.
• Will I be able to determine repeatability of inertia vs brake dyno (YES! I will even be able to do it at each rpm step in 100 rpm increments as well as overall repeatability)

Final statement till next update. probably going to be apples to oranges when comparing "brake hp" to "inertia hp".But I'm getting the numbers just to show this point, and im going to show both of the numbers, even though in an idealistic world they should be the same. The real world is not kind and real world I don't think they will match. I dont even know if im smart enough to know why, especially when you have professionals that do this for a living saying that "there is more error in the test standard and definitely more error in the correction factor software than allows for this level of accuracy" But accuracy is the holy grail here and we are going to get as close as possible while using holy hand grenades to break through myths, and calculating the airspeed velocity of unladen sparrows. Will anybody make it out without a mere flesh would? Only time will tell.

 

[65ShelbyClone]

Those lawsuits are at least partly why the Honda engines don't have the power rating displayed on the shrouds anymore. There's also the issue of intermittent power ratings (higher) vs the recommended continuous power rating. Engine manufacturers were using the max intermittent rating for marketing which was misleading.

Horsepower sells cars too. Auto manufacturers use the SAE method to get legally-binding net power, but, and this is a big one, they have multimillion dollar dyno cells with complete environmental control. Performance models are often underrated as well. There are many examples where a production car makes the rated crank horsepower at the wheels, which suggests that the engine itself is making 15-25%+ more than the glossy showroom brochure says.

As far as accuracy goes, it comes down to how many conditions you can monitor and how many variables you can control. Now we're back to that multimillion dollar dyno cell. Since few of us enthusiasts will ever see one, that's where the SAE correction come in, but it still requires accurate measurement of the uncontrolled variables like pressure and humidity.

The type of dyno I want to eventually build is an eddy current brake. They have instantaneous response and excellent control(so they can mimic an inertial load), but the typical aircooled designs don't sink nearly as much continuous power as a similarly-sized water brake.

 

[Kartorbust]

Going by the 32.5 to find hp, the Vanguard 16hp V-twin is 479cc (29.333 CID, so 480cc) is 14.75hp. But yet the only difference between the 14 and 16hp engines is just the carburetor. Go figure.

 

[KartFab]

Going by the 32.5 to find hp, the Vanguard 16hp V-twin is 479cc (29.333 CID, so 480cc) is 14.75hp. But yet the only difference between the 14 and 16hp engines is just the carburetor. Go figure.

Also, keep in mind that some manufacturers implement "gross" hp numbers, and some implement "net" hp numbers. If you see a number that looks "high" it could be because they literally tested the engine without a muffler at all (just bare exhaust manifold) and with no airbox at all.

This is according to the SAE standard J1940 for engines 1000cc and less that arent road legal.

"SAE J1940 provides a single method for determining and reporting the power and torque ratings of small engines. Using this procedure to provide small engine rating information is useful for comparing the power and torque ratings of various engines. This procedure depends on two primary SAE test procedures to measure engine power and torque; SAE J1349 (Net) and SAE J1995 (Gross) Engine Power Test Codes, which correct the measured test data to standardized test conditions. Using these well known engine test procedures helps ensure consistent and comparable test results and engine ratings. The small off-road utility engine industry, unlike automotive, motorcycle, and marine industries, does not have complete control of the end configuration of the engines in many engine applications. It is typical for the equipment manufacturer to install their own intake systems, exhaust systems, and other engine accessories in many applications. These systems need to conform to certain minimum performance requirements to meet emissions regulations. Therefore the engine manufacturer does not have direct control of the engine power and torque in the final application. A gross power rating is useful for these types of engines because variables such as intake and exhaust system configurations are eliminated from the test procedure. Some small engine manufacturers produce a variety of configurations of a single engine model, with many combinations of unique components for multiple customers, which makes it difficult and cost prohibitive to monitor the net power of all potential production configurations. A gross power rating is useful for these types of engines because variables such as intake and exhaust system configurations are eliminated from the test procedure. Some small engine manufacturers produce engine models with standardized exhaust and intake systems. A net power rating is useful for these types of engines because variables such as intake and exhaust system configurations are included in the test procedure. The flexibility to use either Net or Gross test procedures and engine ratings, as specified in this procedure, is necessary in order to provide customers with consistent and comparable engine ratings. Because of the wide variety of standardized and customized engine configurations that are typical of the small engine industry, the flexibility to use either Net or Gross test procedures and ratings continues to be relevant and continues to be acceptable going forward. The flexibility to rate engines at speeds other than the unique speed settings required for specific engine applications is also necessary in order to provide customers with consistent and comparable engine ratings. Because of the wide variety of customized engine applications with unique speed requirements, the flexibility to rate engines at speeds other than the unique speed settings for each application continues to be relevant and acceptable going forward. It would be difficult and cost prohibitive for engine manufacturers to identify and rate each unique engine application speed setting. The flexibility to select appropriate engine rating speeds for small engines is acceptable when using Net or Gross test procedures and ratings. This procedure will continue to use statistical methods as part of the methodology to better define the declared power and torque ratings for small engines. Variations in the power and torque of small engines have been reduced over the last decade through improved engine design and improved manufacturing technologies that were implemented to reduce exhaust and evaporative emissions. Therefore, the requirement for the statistical mean values for power and torque of tested production engines has been changed from a minimum of 85% to 95% of the declared rated power or torque."

 

[KartFab]

Update: Had build break repeat over to my house today. We ran their turbo predator 212 on the dyno. After some messing around with "pressurizing" their fuel tank at run #7 we got a good run that made some good hp on the dyno. Still having issues with going full lean at WOT, but at partial throttle we were getting 15:1 A/F, yes this is lean, but the engine ran and made some sort of boost at partial throttle (we think). Made more than what i have seen with stage 1 mods on my dyno. They will release the video in the next week or two. Was cool to see the engine make a lot more hp than a stock predator 212 does, but also was exciting to see how good it perfomed even with the air fuel ratio as messed up as it was. If we would have got it to 12.5-13.2 A/F ratio at full throttle it would have made a lot more power, but I dont think we can get it to that point unless we do EFI.

either way it was really cool to see how boost works under load and under no load. It was kind of dissapointing to see the A/F peg out at 20+:1 if you went full throttle. Probably was pushing fuel back into the carb under full throttle/boost.

 

[65ShelbyClone]

I've seen the more recent videos and they have just about all the hardware there to make a blow-through carb work, just a few teaks and tricks left to get there. "Ohio" George Montgomery was making it work 43 years ago.

www.georgesspeedshop.com/Ohio_George/74_Pinto.html
https://www.motor1.com/news/87626/ford-ohio-george-pinto-turbocharged/

 

[KartFab]

So, I'm not a numbers guy. I had a full ride in the college of engineering, but dropped it because I hated math.

With that said, I was able to compare inertia dyno numbers to my water brake numbers. More to come later, but this is just the initial finding.

Step 1: averaged the inertia dyno runs (3 runs performed)
Step 2: subtracted each inertia hp/tq figure at each rpm step (100 rpm steps from 2200-7600 rpm) from each brake hp/tq figure
Step 3: divided the difference found in step 2 by my brake dynamometer values, and made that a % increase (no decrease occurred at any step)
Step 4: graphed % hp and torque increase by each rpm step
Step 5: Fitted a trendline to the hp data. The trendline did not match up at all, so i ended up going with a 6th degree polynomial, which has an R squared coefficient of 0.9911 (meaning this equation encompasses 99.11% of the data, a very close match.

Some people will say YAWWWN BOOORRRINNNG.... So what?

This means that I have an equation that will predict with 99.11% confidence? what this engine would read on an inertia dyno. It is really weird how this percent difference matches up...... almost like I have somehow derived the relationship between a water brake and an inertia dyno with the engine being the control.....

Notice how the horsepower and torque over-inflation is most pronounced at low rpm and high rpm? This explains so much to me, even though I haven't applied it to other engines.

I also calculated inertia dyno variance, and the average variance on this one is 0.027 hp, which is pretty impressive!

 

[65ShelbyClone]

Dynojets have been the de facto standard inertial chassis dyno for decades. With that in mind, have you ever read about the Dynojet "fudge factor" story?

It's about halfway down the page here:
https://www.hotrod.com/articles/dynojet-chassis-dyno/

And a SuperFlow engineer talks about it here:
https://www.rbracing-rsr.com/downloads/dynojet.pdf

I think it may only apply to DynoJets, but it's still an example of the kind of money business that goes on.

 

[Crotchrocket]

It’s really a difference of how they are measuring the power and how they are calculating it. The only problem with using one test and one engine is it doesn’t account for other variables besides RPM and hp delta %. What if the guiding variable is not RPM but hp how would the numbers turn out and would the equation be simpler?

If you create an X-Y scatter of the brake and interia numbers and then fit the relationship you might see something more interesting. Like is there a constant hp being added? Is there a more direct hp to hp difference?

I would do the same for torque. Since the dyno is actually measuring torque and then calculating up based off of RPM then the real error is probably somewhere in the torque measurement. If you look at the percent difference graph you can see the error is lower somewhere in the middle maybe around peak torque? This would make sense as the torque becomes higher it is easier to read and there is less effect of other parasitic loads from the friction of the bearings in the machine or drag from gears and belts. You may find a much simpler correction for torque based on torque input.

The real test will be to take a completely different engine with a different power curve and predict the difference then run the test and see how close it actually predicted the difference.

 

[KartFab]

It’s really a difference of how they are measuring the power and how they are calculating it. The only problem with using one test and one engine is it doesn’t account for other variables besides RPM and hp delta %. What if the guiding variable is not RPM but hp how would the numbers turn out and would the equation be simpler?

If you create an X-Y scatter of the brake and interia numbers and then fit the relationship you might see something more interesting. Like is there a constant hp being added? Is there a more direct hp to hp difference?

I would do the same for torque. Since the dyno is actually measuring torque and then calculating up based off of RPM then the real error is probably somewhere in the torque measurement. If you look at the percent difference graph you can see the error is lower somewhere in the middle maybe around peak torque? This would make sense as the torque becomes higher it is easier to read and there is less effect of other parasitic loads from the friction of the bearings in the machine or drag from gears and belts. You may find a much simpler correction for torque based on torque input.

The real test will be to take a completely different engine with a different power curve and predict the difference then run the test and see how close it actually predicted the difference.

Numbers are corrected. Both hp and torque are off by the same percentage points as you can see in the graph. It is not a linear relationship, and the best fit equation that captures 99.1% of the relationship is the 6th degree polynomial. I wish it was a linear or exponential relationship but it isn’t. I’m not confident about the 7500+ rpm range. I don’t know if the equation will capture how far off it could be at higher rpm.

Using the percent relationship the equation “accurately” predicted stage 1, 2, and 3 results on the inertia (e.g. spat out the numbers you would expect on the Inertia Dyno with the water brake data and 6th degree polynomial equation).

Results will come back for other things too, but I just have to wait for the data.

 

[KartFab]

Calibration
Since the data I collected has come under question, since it is "low" (e.g. not falsely inflated), I have had to buy some precision calibration weights. I really hoped the water brake would be 2 foot pounds off, and that all I would need to do is correct the offset and then it would align with what other inertial dynos are spitting out. I sat in the garage today with the dyno in some soft jaws on the vise, and made sure the dyno was level in both planes. I maintained the torque arm perfectly level, adjusting as necessary as the weight moved the arm the heavier it got. I did not change any values or actually calibrate the dyno, just recorded the observed raw torque reading with the amount of dead weight that was placed on it.

Weights
These weights are a 2 lb hanger, 1lbs slot, 2 lb slot 5 lb slot, 10 lb slot and 20 lbs slot weights that have a positive correction on them with the calibration sheets that came with the weights (e.g. the 2 lb weight has a nominal value of 2 lbs plus 48.8 mg. I corrected them in an excel sheet and summed them when combining them to determine the column "Weight"

Hanger
I used 3 feet of fishing line, 14 lb test, that i wrapped around a gram weight on a small scale and measured the difference between the gram weight and the fishing line. Fishing line added 0.122 grams to the total weight, which is in the "Total Hung Weight" column

Observed
This is the raw torque reading that was displayed in dynomax, the software that runs the dynamometer. I only went up to 30 pounds because nothing I run pushes more than 30 ft pounds, only 25 ft pounds so far.

Accurate?
So here it is, more data that shows that the numbers that my water brake are putting out are accurate. Not sure about repeatability, that will come later for a comparison between inertia and brake dyno data variance.

I am actually kind of bummed that the dyno was spot on. I was hoping it would be off by a ways.

Can an Inertia Dyno Be Calibrated?
After speaking with a guy that sells dynos, he was unable to give me any information as to how one could calibrate an inertia dyno, other than you take an engine and if it reads numbers you expect, put it on the shelf and see if it keeps reading those numbers later.... Not actual calibration, just verification of repeatability... Others take their inertia dyno and can do rundown tests for frictional losses etc, but nothing about actual calibration so far. Others basically put in random numbers for their polar moment until they get about what they want on an engine and say that is good enough and roll with it.... It would be cool if someone could actually calibrate an inertia dyno so you could say, i found the accurate polar moment of my dyno, losses, and know that it reads accurately when it displays 1 ft lb all the way up to 30 ft lb. Pendulum test?

Inertial Compensation
Another item, it does look like there are ways to compensate for engine inertia with an inertia dyno. E.g. the question was posed, if you steady state test an engine, inertia plays no part in the equation, so why do a lot of inertia dynos show higher hp numbers if you reduce engine inertia? Dynomax is able to compensate for that by determining the actual inertia of the engine and water brake to avoid reporting higher or lower numbers depending on which direction you sweep.

I remember seeing a spot for engine inertia in performance trends when i downloaded the trial version. I also remember seeing an engine builder post that he saw over a 1 horsepower gain with graphs and everything simply by swapping out the cast iron flywheel with a billet flywheel.... looks like people arent properly compensating for change in engine inertia. In the real world where physics and math apply, when you change the engine inertia, you dont increase torque output, you just decrease the amount of rotating mass, which allows the engine to apply the torque to the ground rather than a portion of the torque going to accelerating the engine's flywheel... Its too bad people think that actually makes more horsepower.... just frees up the available torque... it doesn't create more torque....

 

[65ShelbyClone]

Can an Inertia Dyno Be Calibrated?
After speaking with a guy that sells dynos, he was unable to give me any information as to how one could calibrate an inertia dyno, other than you take an engine and if it reads numbers you expect, put it on the shelf and see if it keeps reading those numbers later....

I bet it could be done with a power source mounted on a trunnion held with a torque arm. The force and acceleration could then be compared....but at that point you don't need to know the MOI because it's already measuring torque directly. There are also inline torque transducers available these days that would eliminate the trunnion.

 

[KartFab]

I bet it could be done with a power source mounted on a trunnion held with a torque arm. The force and acceleration could then be compared....but at that point you don't need to know the MOI because it's already measuring torque directly. There are also inline torque transducers available these days that would eliminate the trunnion.

hmm mm so my water brake would directly mount to a 1" go kart axle

then i can determine the inertia experimentally via this:
https://www.dynomitedynamometer.com/dyno-tech-talk/determining_inertia.htm

So using this:
http://dtec.net.au/inertia dyno design guide.htm

Now all I need is:
1x 1" go kart axle
2x 1" go kart or pillow block bearings
2x #35 sprockets
2x #35 chain regina chain
1x 1" bore clutch 35 pitch (or direct drive with pto sprocket)
1x 3/4" bore clutch 35 pitch (or direct drive with pto sprocket)
1x used large flywheel from a junk yard 280 lb to 350 lb
1x electric starter hopefully from same said junk yard to start flywheel if not using clutch to accelerate flywheel from dead stop.
1x balancing of said flywheel at machine shop, quoted me $300
brake mechanism, maybe can even get that from junkyard. If not, MCP hydraulic brakes with lines running to each side of the brake, having it spaced apart
1x neodymium magnet secured to flywheel somehow, drill hole and epoxy?
1x inductive pickup for hall effect , already have a cheap one for a bicycle, can plug directly into dynomax
1x software computer etc, already have it and it will work with the RPM A/C channels, and can work out ratios, adjust inertia values etc... I can get RPM A from engine, and RPM C from flywheel

Just an idea i have been kicking around. Would be cool to get a machine shop to make a flywheel that is 1.5" thick or 1.75" thick and weld on the hubs and balance, but no luck so far.