Author Topic: Dynamometer and teststation  (Read 6567 times)

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Dynamometer and teststation
« on: October 28, 2019, 09:49:17 PM »
It is time to think and make plans/drawings for building a dynamometer and combined teststation. The engine that should tested is the one cylinder 55 cc combustion engine. Named SPE-1-55.

A time ago I already take some quick 3d drawings of a idea to build a dynamometer around a water pump of a production car. It happened to get this water pump from a neighbor who collected local scrap iron in his backyard.
On this water pump there is a HTD-8M pulley fitted (20mm). I am thinking of connecting the engine to the water pump by the same type of timing pulley. I can play with pulleys with different amount of teeth if needed.
The pump should be fixed to the teststation with a bearing setup. So only rotating movement is possible.


The idea is to be able to test the enge both at full load and part load. It would also be nice if an engine could be tested with a higher power (and in combination with a higher speed).
Currently I do not know the performance and maximum speeds or the 55cc engine. I had already calculated a global estimate and the power would come in the direction of 1 hp with a speed of 3000 rpm and a torque around 2 Nm.
For designing the boundary conditions are 3 times the guessed performance of the engine. So 3 hp /  6Nm @ 3000 rpm ; 9 hp / 6 Nm @ 9000 rpm.


I can see that there should be enough stress and bouncing forces when the pulley transporting power to the dynamometer. A rigid mounting of the pump is required. Also for mounting the engine to the teststation.
For first try I would start with a simple setup of controlling the load and speed of the engine. There should be a ball-valve on the outside of the pump. Adding a pressure relieve valve for max pressure. And the carburetor is operated with a gas cable.
The torque should be measured by a 5 kg load cell I guess. Speed by the digital output of the ignition hall sensor or a sensor on the pump.



If this work and things seems okay, I would upgrade the system. I am thinking of using a Arduino Mega with Labview or Simulink interface.

- Pressure and temperature sensor water to be sure there should be no big bang.
- Servo controlled ball-valve water output.
- Servo controlled throttle carburetor
- Waterflow sensor

- Air Pressure and Humidity and temperature
- Fuel tank measurement try to indicate usage (load cell)
- Intake maf sensor
- Exhaust temperature / pressure sensors
- Fan cooling

« Last Edit: October 28, 2019, 09:53:01 PM by AlexS »

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #1 on: October 29, 2019, 01:41:17 PM »
I have extensive experience in small engine dyno testing.  We ran engines from 26 cc to 35 cc with up to 12 hp and with rpm up to 25,000.  An article on some of this is at namba.com/content/library/propwash/2019/april/16/

Lohring Miller

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #2 on: October 29, 2019, 09:45:10 PM »
Great info on the article. I can remember seeing a topic come up about the construction of your dyno.

The intertial dynos look robust. How long did a run last on average? For the 55cc engine it should have required a relatively large inertia to have the engine measured for a few seconds. The current flywheel and crankshaft have a mass inertia around the 3 * 10 ^ -3 kg.m ^ 2 / 1.5 kg flywheel, and the engine revs quite smoothly.

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #3 on: October 30, 2019, 12:14:55 AM »
The advantage of an inertial dyno is a short run that duplicates the acceleration of a racing engine.  Unlike marine or aircraft engines that run at a steady, high load; a racing engine spends most of its time accelerating and decelerating.  Our dyno was designed to match the acceleration time into a measured time trap.  Your engine operates at a relatively low power and modest rpm.  You could just play with it's existing flywheel and data log the rpm vs time.  An online guide with a spreadsheet for flywheel size calculations is at http://www.dtec.net.au/Inertia%20Dyno%20Design%20Guide.htm  They would love to sell you the hardware for a complete setup.  We used Performance Trends system which is lower cost.  See http://performancetrends.com/dtm-dyno.htm.  If you understand the physics, any method that can log rpm vs time and calculate the torque from the change in rpm will work.

Lohring Miller

ChuckKey

  • Guest
Re: Dynamometer and teststation
« Reply #4 on: October 30, 2019, 12:17:24 AM »
A thing that spins up a flywheel is not what I would call a dynamometer. A dynamometer provides a varable load to allow as a minimum, speed and torque to be
measured properly under steady-state conditions. This allows torque and power curves to be plotted under different conditions.

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #5 on: October 30, 2019, 01:50:51 PM »
You obviously don't understand how an inertial dyno works.  You get a complete power curve over an rpm range very quickly.  The result we got compared very favorably to tests of the same engine on a water brake dyno.  We could run a lot more tests in the same time period with a lot less wear on expensive racing engines.  If you are testing a tuned pipe two stroke, it is a lot harder to hold a constant rpm with a brake dyno.  There will be areas in the power curve when the engine comes on the pipe with a huge torque surge.  An inertial dyno just increases its rpm a little faster.  Nearly all serious two stroke motorcycle and kart engine tuners use inertial dynos for the above reasons and for their simplicity.

Lohring Miller

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #6 on: October 30, 2019, 08:03:32 PM »
Ah right, I can see the difference for the use of the two type of dynometers. Given that the engine is a 4 stroke and low spec cams and compression, the torque should be really flat from low to max rpm (amount depend on load of course).
It would be nice to test engine under different load types. And maybe for longer times under load. So preference is to use a water brake dyno or disk brake. If it doesn't work out, a flywheel-based dyno can be made for determining the power curve.

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #7 on: October 30, 2019, 08:08:48 PM »
I do like the software that Performance Trends sells. The amount of extra sensor logging and interface.

ChuckKey

  • Guest
Re: Dynamometer and teststation
« Reply #8 on: October 31, 2019, 12:16:23 AM »
I stand corrected on the inertial device. Do you make allowance for bearing friction or air resistance? For a steady-state arrangement, the load can also be a fan. Easy to make nicely inefficient with flat radial paddles. Not recommended at 25,000 rpm, though.

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #9 on: October 31, 2019, 03:29:39 PM »
We never made any allowance for bearing drag or windage.  We did approximately correct for the inertia of the engines rotating parts, but it was small.  We believed that other factors were even smaller, but can be found by a coast down test.  That's the reverse of a standard test where you measure the rpm vs time from a high rpm point down.  You can also measure the engines internal friction the same way.  We never did this.  We were mostly interested in comparing the difference changes made on various engines.  We only corrected for atmospheric conditions so tests done years apart should be compared. 

My favorite series of tests were fuel and gasoline tests on a single engine.  See below.

Lohring


Offline Admiral_dk

  • Full Member
  • *****
  • Posts: 3752
  • Søften - Denmark
Re: Dynamometer and teststation
« Reply #10 on: October 31, 2019, 08:11:45 PM »
You have shown this before Lohring and I have seen/read this from you before, .... but this time I couldn't help thinking about the exhaust temperature of the different fuels and your very informative graph on Jim's thread about the extra power you got from adding water to lower the temperature => extending the pipes useful range.
So my question (and I'm aware that you might not be able to answer, if you didn't measure it) - do you think that temperature difference is so small between those fuels, that the same pipe would be the optimal one ?

Best wishes

Per

Offline Allen Smithee

  • Full Member
  • ****
  • Posts: 1130
  • Mordor, Middle Earth
Re: Dynamometer and teststation
« Reply #11 on: November 01, 2019, 07:59:56 AM »
When bench testing engine configurations for FAI-F3d pylon engines we found that water fed into the pipe could effectively lengthen it. We could fit a prop which was normally too "heavy" to get on the pipe on the ground and use the "water injection" to get it going. Once in the air the load reduced and the engine would turn it without the water feed. I tried it a few times, with a 4oz (100cc) water tank fed by crankcase pressure (the fuel system ran on pipe pressure) from startup until half way down the back straight on the first lap. It allowed around 1/2" more pitch and was worth at least 6 secs (10% better) in race times. When the water was switched off in the air there was around a 1200rpm rise. The water was injected near the front of the pipe (25mm from the piston face) to give it maximum cooling effect. In that ~30secs of running the system used about 60% of the water tank.

But the technical committee decided to call the water "fuel" and so the setup violated the rule defining the fuel spec (80/20 methanol/castor oil). I tried injecting fuel rather than water but the results were very unpredictable. Sometimes it worked, but other times I think the fuel was actually burning rather than just evaporating which made the pipe hotter rather than cooler. I could probably have fixed this by moving the injection point further down the pipe, but I was advised that this would be deemed "not two-stroke power" and would violate the engine spec rules...

I got their point, but it was fun to try it!

AS
Quidquid latine dictum sit altum sonatur

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #12 on: November 01, 2019, 03:14:14 PM »
One of the biggest flaws in our testing was EGT measurement.  We didn't do it seriously.  That would have told us a lot more about the proper fuel/air ratios.  The ear of an experienced engine man, Mike Bontoft, worked fairly well for our purposes.  An oxygen meter would also have been valuable.  I have a design that fits over the stinger and uses a standard sensor.

We used water injection several different times.  It's really easy on a boat.  I'm not sure why water injection isn't used instead of variable length pipes on full size outboards.  I also don't understand how water that doesn't go through an engine can be considered fuel.   Below is an onboard data log that clearly shows the plateau with the water on and the rpm rise after the water is turned off.

Lohring Miller
« Last Edit: November 01, 2019, 03:18:17 PM by lohring »

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #13 on: November 05, 2019, 09:32:33 PM »
I have been thinking a few things. A fan or propeller like dyno is certainly a good alternative. It requires a continuous load at a certain speed. Only I can not imagine that you can vary the load (so that the engine sinks in terms of speed).

I think I give the water brake a go. It will not be the simplest, but if everything works respectively, I can go in more directions. Can possibly also test small moped engines.



Thinking out loud.

I have tried to make an estimate of what the current pump will yield in terms of water and thus the required back pressure of the valve to decrease the power. Based on the output of the pump, something could be said about the required valve size.

As you can see in the first image, I have determined the area and therefore the rotation volume of the impeller.
This should be around 27 cc. Only I can see that it is not 100% giving the amount of water to the outlet. So the outlet flow would be less as a volumetric efficiency. I have no idea what that would be and would depend on speed. So I say this would be 50%.

There should be a gear drive between the engine and the pump. Smallest pulleys available are 22. Pump is 20. So speed would be slightly higher than rpm from this point.
Only when running engine speed around 3000 rpm, this would result to 44 L/min. With the 50% factor in mind.

So why not a extra gear between the engine and the pump to lower the flow output. A set of 22 and 44 teeth pulleys result in a i1 of 2:1. And from there 22 to 20 as i2 of 0.9:1
Pump speed would be 412 when the engine 750 rpm. And 1650 when running 3000 rpm. Output around 22 L/min.



What is a reasonable size of the valve? With a flow of 22 L/min and assumed maximum water speed of 2.5 m/s this result to a pipe/open valve diameter of 0.54 inch.
Speed go higher when valves closed an cause pressure.
I was thinking to use a 1 inch ball-valve. This connect directly to the pump mainbody.


A quick look on the needed pressure for result in absorption of power.

Assumed a linear build-up of power consumption from 0.07 to 0.74 kW at an increasing pressure of 2 to 20 bar at 3000 rpm (engine).

So might output of of pump be higher that of course the needed pressure would be lower to eat power. And if output is to low it is possible to run the pump without the extra gearing.



What do you think guys?

ChuckKey

  • Guest
Re: Dynamometer and teststation
« Reply #14 on: November 06, 2019, 12:24:56 AM »
For a centrifugal pump, at any particular speed, a curve of pressure vs. flow rate can be plotted. There is a point on that curve where the pump is at its most efficient. A series of curves can be plotted for different speeds. If you join up the points of most efficiency, you find that the flow rate increases linearly with speed and the pressure increases as the square of the speed. (The kinetic energy of the liquid leaving the impeller is largely converted to pressure in the volute casing as the flow slows down, in accordance with Bernoulli's equation). This means that the power absorbed increases as the cube of the rotational speed.

Perhaps surprisingly, it is also around the maximum efficiency point that the pump will be absorbing the most power. A pump well designed for a particular duty comes on song at that duty, a bit like an engine coming on the cam, or a 2-stroke coming on the pipe, but much less sharply defined.

These effects are less pronounced with the crude simple impeller and large clearances your pump would seem to have.

The volume of liquid encompassed by the impeller tells you very little about the pump's output.

Quite early in my career I had the job of designing a pump test facility. Its exactly the same as a dynamometer except that most of the instrumentation goes on the power absorber rather than the power source. My rig was a driven by a 250 HP thyristor controlled motor. 

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #15 on: November 06, 2019, 03:05:05 PM »
Jim Allen has built both a water and fan brake for model engine dynos.  See the bottom of https://www.intlwaters.com/media/users/jim-allen.2129/albums?page=3 for the index to the pictures.

Lohring Miller

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #16 on: November 06, 2019, 08:01:59 PM »
Thanks guys, this is stuff to think about. I agree that it is difficult to estimate what the output of the pump would be. Also At some point you should just build it and make adjustments on the way. But I need to dig in deeper to estimate absorbed power with this setup.
« Last Edit: November 06, 2019, 08:19:41 PM by AlexS »

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #17 on: November 07, 2019, 03:21:51 PM »
Another quick way to estimate power is with calibrated propellers.  I can't post the .pdf file.  See below for some excerpts.  Send me your email for the complete paper.

Lohring Miller


Offline Allen Smithee

  • Full Member
  • ****
  • Posts: 1130
  • Mordor, Middle Earth
Re: Dynamometer and teststation
« Reply #18 on: November 07, 2019, 06:02:39 PM »
Many years ago I used to have a set of Kavan calibrated props, but I sold them because I didn't use them enough. These were (IIRC) machined and ground steel props which each had a calibration table of RPM vs Torque.

It's not that hard to make your own using ordinary commercial propellers. The calibration table can be established using an electric motor because it is relatively easy to derive the power absorbed by the load of a BLDC motor using just a voltmeter and a clamp-meter. For wooden, composit or plastic propellers it's necessary to measure over a wide range of RPM because they flex, whereas the Kavan metal ones can be assumed to be rigid over their intended speed range, but that just means plotting a table or graph rather than using the single P number as above.

To measure the power absorbsion just set up the prop on a known electric motor. You need to know (or measure) the following:

Motor Constant Kv (in rpm/v)
No Load Current I0 (in Amps)
Motor winding resistance Rm

You then fit the prop and measure the RPM (s), current (i) (with the clamp-meter) and applied voltage (v) as you raise the voltage over the desired range. Output power can then be calculated from simple equations:

P(out) = P(in) - Losses

 = (v x i) - Losses

 = (v - voltage losses) x (i - current losses)

= (v - (s/Kv)) x (i - i0)

Or if you want the answer in BHP rather than watts it becomes:

= ((v - (s/Kv)) x (i - i0)) / 745.7


So you can take some measurements on a set of props and record the data in a table or graph. Not that this tells you the power DIRECTLY. If an engine turns that prop at the RPM recorded in your table/graph then that is the power it is developing.

If you use it in this way - to provide a set of calibrated props - you will need to apply factors to correct for aur density and viscosity (there are standard correwction formulae for these), but the beauty of this approach is that you can use it as a direct test reference. Put a prop on your engine, measure the RPM. Then take the prop off the engine and immediately fit it to the electric test motor. Increase the voltage until the RPM is the same as for your engine and then the calculated power will apply to both - no corrections of any kind are needed. It also means you don't need to spend time testing/calibrating and checking a whole family of props in case you need to use them!

AS
Quidquid latine dictum sit altum sonatur

Offline dieselpilot

  • Full Member
  • ****
  • Posts: 341
Re: Dynamometer and teststation
« Reply #19 on: November 07, 2019, 10:12:01 PM »
Allen have you done this in practice? How was accuracy? The problem with this approach is calculated vs real world efficiency (or motor constants).

Offline Allen Smithee

  • Full Member
  • ****
  • Posts: 1130
  • Mordor, Middle Earth
Re: Dynamometer and teststation
« Reply #20 on: November 07, 2019, 11:06:57 PM »
I've done it, and I got a number that was in the expected range - I have nothing to calibrate it against so I can't make any claims on accuracy. But I would take issue with comments about "calculated" efficiency - with electric motors all these things come down to measurable fundementals: voltage, current, resistance and rpm. Of these the only one that is "tricky" is the winding resistance because it's a very low value (a few tens on milliohms) and as such should really be measured using a four-wire or bridge method, and for absolute accuracy the resulting number should be factored by the thermal resistivity coefficient for the (measured) winding temperature under load. But the differences are tiny.
You mention motor constant - this is very easy to measure, so I don't see the problem.

The beauty of this approach is that all those normally troublesome-to-measure losses (bearing friction, magnetic losses, copper losses, air drag etc) are directly accrued into the measureable parameters. The no-load current represents the sum of all the power lost to friction and magnetic effects. In this way you can directly calculate both input power and output power.

AS
Quidquid latine dictum sit altum sonatur


Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: Dynamometer and teststation
« Reply #22 on: November 10, 2019, 03:15:44 PM »
All these "propeller" brake methods are approximate, but have been used for a very long time.  Nearly all nitro engine development was done by measuring rpm with a select group of propellers.  The big advantage of inertial dynos is that in the time it takes to make one run on a brake dyno, you can get a full power curve.  You can test dozens of things in an afternoon.

The main problem is people that don't understand physics don't understand how they work and why they accurately measure power.  In the days where engines ran at the same rpm with a fixed load, brake dynos simulated the real world.  Where engines are constantly accelerating and decelerating as in racing, an inertial dyno better simulates the situation.

Lohring Miller

Offline Allen Smithee

  • Full Member
  • ****
  • Posts: 1130
  • Mordor, Middle Earth
Re: Dynamometer and teststation
« Reply #23 on: November 10, 2019, 09:07:32 PM »
I agree with you - the inertial dyno is beautifully simple and fundamentally (intrinsically) accurate. I didn't use them on FAI-F3d engines for two reasons:

1. I was concerned that I would not be able to stop the engine running past the upper operating point (call it 30,000rpm) quickly enough to prevent it over-revving an un-necessarily grenading engines.

2. With the very high Q of this kind of engine/pipe combo I was never convinced that the mixture (and thus ignition timing) would "settle" to a representative condition on the way through, so the curve as measured may not have been representative if the in-race condition

3. For FAI-F3d engines we very much DID want a single, constant power setting - the engines weren't equipped with throttles and wouldn't have responded to them in any meaningful way even if they were.

OK, so that's three reasons - sticking within a bidget was never one of my strengths...

AS
Quidquid latine dictum sit altum sonatur

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #24 on: July 26, 2020, 10:45:43 PM »
I am leaning towards to build an dyno on the eddy current principle. I found some video's on the internet about the principles and working. But not much literature about designing an eddy current dyno. It would be nice to see what the effects are of coil specs and voltage/amperage to act like a brake.

I was thinking if it is possible to spin a aluminium disk. And the both outside would be a metal connected to multiple coils. And maybe control the voltage on the coils duo pwm duty cycle(changing in % load)? And different resistors/pot to set the base line load? I am thinking out loud.

Thanks in advance!

Alex

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Dynamometer and teststation
« Reply #25 on: July 27, 2020, 03:42:54 PM »
I believe that some of the fitness (torture) machines use eddy current brakes to set the load for the user. These are often available second hand quite cheaply and may give you something to start with. These should also be capable of disapating a few hundred watts.
Best regards

Roger

Offline AlexS

  • Full Member
  • ****
  • Posts: 269
Re: Dynamometer and teststation
« Reply #26 on: July 27, 2020, 10:49:41 PM »
Thanks for the idea Roger!

 

SimplePortal 2.3.5 © 2008-2012, SimplePortal