Why Dynojet numbers are always wrong

mbohn said:

Well, they don't! But I suppose others have pointed out where and when the DJ dyno is useful.

Click to expand...

I think the point of the article is that a DynoJet isn't very useful at all. It looks like DynoJet did a classic engineering tradeoff. At the time it was created the existing dynos were expensive because they had to have something to absorb all the torque and measure it accurately (such as a large electric motor).

DynoJet figured out that a heavy roller would absorb the torque also. By knowing the inertia of the roller they could just do the math and calculate torque, and measuring the hp is easy since it's just torque x rpm. All that sounds great in theory, and it made for an inexpensive dyno.

DJ uses a roller that is the same weight for all bikes from an Aprilia RS 125 to a Gold Wing. By physics it produces wrong figures unless the total weight and rotating weight of the bike matches whatever they used for their prototype in picking the roller weight.

So it seems DynoJet did this knowing the power figures would be wrong in every case, but the benefit was a dyno that dealers could actually afford. On the whole that's not a bad tradeoff. My bathroom scale isn't as accurate as a lab scale, but I'm OK with that because I'm not willing to pay $5000 for a lab scale to weigh my fat butt.

Beyond this, the article brings up two other key points. First, DJ can only accurately measure torque and hp under acceleration. Since the dyno can't measure steady rpm power then tuning is limited to only full throttle applications.

In addition DynoJet admits to simply gaming the numbers and adding 15% or so to every calculation it does.

So the way I see it you could call a DynoJet "useful" if what you want to do is a before-and-after comparison of a modification and you want to test under full throttle only. For everything else it's worthless. We can't compare bike-to-bike because all bikes have different weight and rotating inertia. We can't do comparisons on different DynoJets on different days, or in different locations, or with different operators.

It doesn't accurately measure torque and hp. It's not even close. Rather than calling the results horsepower, we should call it "DynoJet units."

I'm fine with all that but the mags should be clear that the numbers are BS rather than just print them like they mean something.

Marc

I did a complete map about 1.5 years ago on my BMW with an aftermarket ecu on a DJ. I told the operator where to go in terms of rpm and tps and he adjusted the dyno to hold that point. We then looked at a/f and I adjusted the fueling at that rpm/tps point to get the a/f I wanted. We then moved on to the next point in the map. I was told that we could not extract torque for these points, that that only could come from a "pull" (using the inertial method). Maybe that has changed since then. I don't really see why if you have to supply the current to hold the rpms fixed that you can't measure it and calculate torque.

but not actually measure if the resulting mixture provides the best HP at that matrix point?

Click to expand...

Yes, that is correct from my experience with DJ.

So how do they actually create a map at less than 100% TPS - they go through each TPS % at a time in inertia mode, and plot the approximate torque/HP through the individual acceleration runs that that TPS %?

Click to expand...

As I said by running rpm/tps points in the map and adjusting a/f to the "desired" value. This value is supposedly the one that gives best torque. In my case the operator recommended a value, 13.2 I recall, and we went for that. Except for wot, there is no confirmation unless you want to do pulls at part throttle. So maybe what we are saying is that trusting that the operator knows that a/f to use for max torque is the weak link in this method, or you have to start iterating between pulls and the a/f adjustment.

How does one then ignore any acceleration factors in the ECU dwell calculations - use only a 4th or 5th gear run, and hope that the acceleration rate does not invoke an ECU acceleration enrichment algorythm?

Click to expand...

If you are talking about accel/decel enrichment, this method doesn't do that at all. I suppose there are two choices: (i) identify poor accel/decel points on the road and adjust from there, or (ii) see how the bike responds in the fixed rpm mode on the dyno when the tps is opened up. In my case I used (i) because with the system I was using it was easy to adjust a bad accel/decel point using an on-board Palm or when I got home with the main software running on a PC.

JEC said:

Related to the original post, Superflow dynos are one of the best around. They measure SAE horsepower, the accepted and most accurate standard, which reads considerably lower than a DJ. For instance, an R1 will put out around 110-120 hp in SAE numbers. Some 1000s can crack 130. Nowhere near the 140-160hp numbers touted on most dynos for litrebikes.

But the nice thing about the Superflow is it will give you DJ numbers as well - so you get a curve showing SAE hp and DJ hp, so you don't have to do any conversion for the less accurate but more common DJ figure. So my old SV650 made 64.5 SAE hp and 75 DJ hp on the same dyno, on the same run.

Yes, Superflow dynos are expensive. They incorporate sniffer units, ram-air systems, an engine brake roller, and meteorological stations to correct for atmospheric conditions. Hence why they aren't common or accepted.

Click to expand...

Wake up, you sleeping thread!

On the Superflow / "DJ" HP number "conversion" issue -

It's VERY approximate.

First: The Basis for inconsistent DJ hp

Cycle Canada tested several different dj dynos and got a 5% or 7% "range" of readings on 3 different bikes on 3 different dj dynos.

Dynojet hp "conversion":

Superflow ran some bikes on their dyno and 1 dj dyno and made an equation that factored up the readings to read about what that dj said on whatever few bikes they ran.
In Superflow's estimation, their DJ channel is +/- 10% of a DJ number.
So -
85 True hp on a SF = ~100 djhp + ~15%, that's 115 and a +/- 5%-7% range depending on the dj dyno you are comparing to. So a DJ might read 110 to 120 djhp, with an average being 115.

If all you want is a "bigger number", maybe their Dj HP channel is good, but, not up to any standards.
...........................

As far as dynos, the SF is a great dyno - but they are by no means the easiest "tuning" dyno, unless you use them in the simplest fashion - unloaded Sweep Test with an O2 sensor. And then, you are only "tuning to an AFR.

There are others that include better tuning features, like better drive rollers for better traction and integrated high speed 4 gas EGA's (instead of a simple wide band "AFR") and quicker setup features.
...................................
DJ HP scale numbers

Funny thing is, that I found the Cycle magazine that had the "V Max magazine test bike" sleeper in a comparison article.

It made 119 Real HP on a brake dyno, with "chain drive to the dyno" (no rear tire).
That 119 might be a bit skewed, but a gs1150 and FJ1100 made a touch over 100, an 1100 Sabre and Ninja 900 made mid 90's. And "they" sound about "OK", when you subtract 8 to 10 hp for the tire.

In reality, a stock V Max (the one you and I could buy) made about 95 True (measured at the rear wheel, on an EC997 dyno, that I know is correct)

So -
the V Max they tested, chain drive w/o the aprox. 10 hp loss from tire flex made:
119 - 10 or, 109 True (with the rear tire)

A real, store bought V Max makes maybe 95 (with the rear tire).

That means (using a few minor leaps of faith) that Cycle mag got a VMax that made 109 hp (if measured with a tire, at the rear wheel) and a normal VMax really made about 14 hp less (109 sleeper less 95 real bike = ~13%.
Or maybe it's 20% - 24 hp difference / 119 sleeper hp -
So - the DJ inflation is based on a sleeper bike's reported hp vs. a real dealer bike and they didn't want their dyno to read lower than the Kerker dyno -
Not knowing they were comparing their dealer bought bike to a sleeper.

WOW. There's the difference. With a path to their HP scale madness. In order to make a real VMax read high like the magazine bike, they arbitrarily inflated the "real" rear wheel (with a tire) hip up to where it matched the sleeper.

I do know a bit more -
When DJ first brought out their dyno, it read VERY VERY high - at Daytona, a cbr600, with just the addition of a pipe, it showed more power at the rear wheel than Honda claimed at the crank! Those original Super High dj numbers sort of meandered down with software changes as it was obvious that the numbers were so high that only uniformed took them seriously.
That I remember.

So, the arbitrary DJHP scale, as Dobeck admitted, was actually related to a "sleeper bike's" "maybe enhanced" engine...
So.... we are stuck with inflated HP figures until the SAE or FTC looks into it.

So - now - this post will be buried deep down in the Internet's nether regions.... <wink!>

Cheers!

Marc

sae is just a correction factor, dynojet offers it as one of the 4 to choose from. sae and eec gives the lowest number, std and din the highest. it's about 3hp per every 100hp difference as a general rule ime.

i'd expect any correctly calibrated brake dyno to give the same result (within a very small margin) as it's an absolute - the load required to stop a particular engine accelerating. roller grip is probably the biggest factor is the load cell calibration is correct.

inferring hp from the acceleration of a rotating drum is different, no less useful ime, but different. dynojet have done a great marketing job with two different products - chassis dynos and power commanders. they may not be the best way of doing stuff, but they are common and useful enough and that helps. seems many brake dyno manufacturer have been running this "dynojet is wrong" thing simply as a way of competing and differentiating with dynojet, because dynojet is so accepted that it's simply them feeling like they have to use the put down the competition marketing principle.

which says more about dynojet's marketing than anything else. they're very, very good at selling stuff, and in a free consumer society once you have deep market penetration your version of the truth becomes fact. kind of like politics or religion i guess.

a graph from the dyno i use (dj200?) and one owned by the company who does a manufacturer national superbike team out here (250). bike is an ex aust superbike MV 312R, run on the dyno i usually use to see what it made then run on the other to see if it was a comparative dud or not. the dyno i use has a fairly smooth roller these days, but it's the only dyno i use so that doesn't matter for me. only when someone has paid a lot of money based on someone else's power claim from an interstate dyno. bike was untouched between the two runs, done on different days.

altho the dyno i use and the quite new dyno at ducati city were within 1hp for my 75hp 750m and 3 or 4 for a 145hp 1098.

SAE isn't a HP scale.

SAE and EEC are simply standardized correction factors for "non-standard" ATMOSPHERIC conditions. They are small.
The "standard" for "standard air"is SAE's call or the EU's call or whoever's standard you want to use.

What atmospheric "Correction factors" are supposed to do, is to allow for higher or lower atmospheric values and calculates what the power "would be" if it was 59f/0% humidity and 29.92 baro, if we were using SAE J1995 standard.

That way, you can compare runs made a somewhat different atmospheric conditions.
Otherwise you'd need a temperature, pressure and humidity controlled room and air supply to compare runs.

As far as different hp numbers obtained when using SAE (which one?), EEC, DIN Correction factors?
Each has it's own idea of what "Standard air" is and will correct slightly differently. The math is a bit different, too.

Any time there's a correction factor applied to the actual power a run, it should say "SAE J1995" or "EEC" or whatever factor used. If it's SAE, it needs to say which SAE Correction factor it used, as there's more than one.

At any rate, all these ATMOSPHERIC Correction factors are usually a percentage of 0% to around 3% (up or down from ACTUAL hp) - not 10% to 15%+ as is the difference between True and djhp.

At least with "True HP" - ANY dyno company can provide comparable numbers. With djhp, the "other" dyno companies stumble around and guess about some "secret inflater decoder ring" - and sometimes it's kinda close and sometimes it's not.

Don't expect any Brake dyno to replicate a dynojet number unless the dyno mfgr or user fudges the number up.
Sorry, that's well known in the non-dj world -

There's an article in MCN coming out.

Thanks -

Marc