Engine Dyno Practise

[topcatproduction]

Hi all, first post here and joined as I have read another good thread on water brake engine dyno's and looks like a really helpful place!

I'm looking to build a really simple dyno and had a thought which I cant find anywhere! Is there any reason why you couldn't just mount the engine on a good frame, rig up a clutch of some kind, have a set length bar (say 5') from centre of clutch plate to a point on some scales and when you hold the engine at set rpm's, push the clutch in to engage and note the highest reading the scales go to. I'm assuming as the revs dip the load will slightly reduce so you could get an accurate reading? I know this would eventually burn the clutch out but only be a couple of seconds a time...

This way all you would need to do is multiply the reading by 5 (5ft bar) and you have ftlbs at set rpms to plot onto a graph.

Simple in theory but would it work?! I attached a very basic drawing.

Also all this talk of using truck eddy brakes- how do you take any sort of reading from them? I'm getting a bit confused!

 

[MikeHalloran]

That's pretty much how all dynamometers work.
The trick is that you have to get your data before the clutch smokes.
The problem is that in order to get repeatable, accurate numbers, you have to wait for the oil and coolant temperatures to stabilize at the applied load level. No clutch will last that long.



Mike Halloran
Pembroke Pines, FL, USA

 

[topcatproduction]

So would it work using a torque converter type pump, or water or oil pump constantly slipping thus pushing on the scales or is that exactly how they work anyway?!

 

[patprimmer]

That is exactly how a water brake works. The load is regulated by controlling the water feed to the pump.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[MikeHalloran]

That's exactly how a classical dynamometer works:
Some kind of mechanism with a controllable slip torque (the classical mechanism is called a Prony brake), coupled to a lever arm that bears on a scale. The brake is free to rotate, restrained only by the lever arm.
Reaction torque is computed from the scale reading and the lever arm length.
RPM is measured separately.
HP is computed from RPM and torque.



Mike Halloran
Pembroke Pines, FL, USA

 

[Alec186]

Topcat - I have tried making a simple dyno similar to that you described - your idea should work quite well.
Rather than a clutch I used a front wheel drive (Morris 1100) drive shaft, suspension upright and disc brake. I think a disc brake is better able to take the punishment and is also easier to load more progressively than a clutch.
One problem I didn't forsee (and you will have to allow for this) is the enormous amount of smoke, sparks, dust etc. that come off the disc brake when it was working hard under load.
You will also probably need two operators to juggle the throttle, brake load and read off the RPM and the weight scale.

 

[topcatproduction]

Cool thanks for the replies. If I used some kind of water brake or eddy brake which was constantly dragging, then should I be able to just use a hand throttle (non-sprung) and increase the revs in increments, taking readings at each, or would I also need to constantly increase or decrease the drag induced by the brake?

If it was dragging a fair amount then would the engine ever show it's full potential or would it not have enough chance to get there?

Basically I'd want to put Rover V8's on, some land rover diesels and maybe even little A series engines, so I'd also need an easy way of adapting to different flywheels. The A-series might not be possible actually as the gearbox has to be attached as they share the same oil (but ignore that as it will go way off topic!!)

 

[patprimmer]

You need to test at full throttle and adjust the brake to hold the revs you want.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[topcatproduction]

Ah- so run at full throttle, quickly bring in the brake to hold at required revs at WOT, then could use the same brake resistance for all the testing through the scale? The only thing I'm not sure about is how you would know if it was at its full potential or whether you could increase resistance a fraction more to show more torque before the revs dropped further...

I can see it being a balancing act where you could run at WOT but having so much resistance it only runs at half speed but shows plenty or torque, where you could also run at the same speed with half throttle and a less resistive braking effect whilst showing much less torque.

Would the best way to test it to take to required revs, then dump the power onto the scales taking the highest reading, release, rev to next increment, repeat process etc? Not how the big expensive dyno's do it but hey, were not talking thousands of £'s investment here...

 

[patprimmer]

NO

you test at full throttle. You adjudt the load to control the speed. You change the load to allow the speed to change for the the different rpm increments, but still at full throttle.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[topcatproduction]

Really? Didn't know that. So for example an engine will scream to say 7k but peak power will be much less, say 6k, do you push to WOT (7k) and then increase the load until it shows the highest reading (around the 6k mark), then keep increasing the load will show less and less torque as the revs cant keep up? And continue right down the rev range or is it only good for measuring peak?

 

[topcatproduction]

Also these should be pretty good yes?

http://cgi.ebay.co.uk/Jship-150Kg-P...cales-/130324608708?pt=UK_Packaging_Materials

 

[Warpspeed]

You will never be able to get a steady enough reading on your scales to be useful or readable. You will be juggling that throttle and brake lever, and the rpm and torque will both be wobbling up, down and sideways, all over the place.

And as another poster has said, the noise, smoke, sparks, and smells will be really exciting. It may even fly apart and hurt you. But as a serious repeatable measurement tool for any useful engine development testing, it will be all but useless.

Get yourself either a water brake, or an eddy current retarder, mount it all in a very sturdy frame, and save yourself a lot of grief.
Then build a closed loop speed control system for it.

You set the target speed to maybe (?) 4,800 rpm.
Below that rpm, the dyno offers zero resistance, and you can start and warm up the engine under no load.

When you open the throttle, the rpm quickly rises to 4,800 rpm, and as it tries to rev any higher, the dyno load comes in harder and harder, so it holds that rpm rock steady at any throttle opening above idle.

You can then read off the exact steady torque on your scales at 4,800 rpm, and that is it......

Rinse and repeat at a different rpm setting.

 

[topcatproduction]

Thats exactly what I was thinking- big strong frame with scales mounted well and eddy brake, but I may not have actually stated that since my original post...

Still not sure how you would know how much load to put on the brake (eddy/water etc) as it must be adjustable, and too much wouldn't let the engine rev enough and not enough it will just rev past it without showing much load on the scales, or do you just juggle it with WOT so it revs at a certain speed but always WOT, then you always have 2 accurate reference points- 0 and WOT.

 

[Warpspeed]

That is exactly why you need a closed loop speed control system.

It will automatically adjust the dyno load resistance to whatever speed you wish to test your engine at.
All commercial dynos do it that way.

 

[SomptingGuy]

Simple dynos (e.g. a manually adjusted water brake) will provide torque as an increasing function of speed. There will be a speed at which the brake torque matches the engine torque.

As you open or close the dyno, its torque vs speed relationship will change - the curve will steepen or shallow. This will move the equilibrium point and the engine will naturally speed up or slow down.

In practice it's quite easy to run a power curve by gently winding a dyno in or out to control the engine speed. Of course if you want a part-load point, that's more complicated and where closed loop control comes in.

- Steve

 

[Alec186]

Topcat - It is interesting that you mention A-Series engines. This is what I was testing. I used a Mini subframe with Morris 1100 drive shafts and disc brakes. The subframe pivoted on pieces of steel pipe. I used bathroom scales under the gearlever extension to measure the load.
Oddly enough it is not too difficult to balance the throttle and brake load.
My main problem was deciding just how long the lever arm was, where to measure it from etc. With a Mini engine/gearbox you also have to factor in the diff ratio and gear ratio (if not in top gear). Mini engine/gearboxes are (in)famous for losing almost half the engine power on the way to the wheels - so I was never sure what I was measuring.
Another problem was to decide whether the torque/power readings I got were the equivalent of power at the flywheel, at the wheels or whatever. But as a comparison with a standard test engine the method worked very reproduceably and consistently.

 

[topcatproduction]

Well even if I can only get a WOT reading at say 1000rpm increments then it will be good enough to be handy, so I will start hunting for an eddy brake! I'm a bit put off by water brakes as I can't see cavitation not causing problems.

 

[Warpspeed]

An eddy current retarder will be far simpler and much easier to control, because it can be controlled electrically.
The only down side to using one as an engine dyno, is the rpm limit on most is typically around 4,000 rpm.

But you can always fit the gearbox to your engine and run it in a suitable gear.

While a water brake will certainly increase it's resistance with rpm, if the torque also rises steeply with rpm, accurately controlling the speed can still be tricky. It works far better with a flat or a falling torque curve.

 

[patprimmer]

For many years before electronic controllers where even thought of, let alone invented people ran water brake dynos with a hand held string to pull on the throttle and a simple water flow valve to control the load.

You pull the string and turn up the load. If the revs are dying you back off the load before you stall it. If the rpm is increasing to fast you ease off the throttle to give you time to wind on more load then apply more throttle.

Once you have a steady rpm at full load, you slowly change the load and read the rpm. You can do this at however many rpm increments you like, but keep in mind the resolution of your instruments so you can keep it realistic.

I really can't simplify it any further.

Fancy controllers are nice, but dynos existed and operated long before these controllers existed.

I can remember dynos where you slid a weight out along the arm attached to the brake so as to measure the reaction. Load cells either electronic or hydraulic are a lot easier.

The device does need to be durable.

The load does need to be able to be maintained in a stable manner. A water brake or eddy field or even an adjustable electric generator set with somewhere to dissipate the charge can all do that.

Monitoring of temperatures and humidity is important to get accurate reproducible results.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules