Hot Rod Forum banner

Home Built Engine Dyno

37K views 55 replies 10 participants last post by  dvnfarrell 
#1 ·
I'm starting this thread to continue the discussion from the "Engine Test Stand" thread we were hijacking.

I put water in the brake today. It's was leaking everywhere. Venting it helped most of it. I think a little Teflon and some hose clamps will fix the majority of the rest but it looks like the shaft seals are going to leak some. Completely full, it seems to be stalling my na SBC at around 2000rpm. Megatune kept locking up. I put 6 inlets and outlets for both sides. It has a double sided rotor and two stators. My intent for all the fittings is to be able to experiment with computer controlled load with multiple valves and to possibly recirculate the water through a cooling system. I think the way I have the hoses routed from one fitting to the other right now may be reducing the load by allowing to act more like a pump than a brake.


Watayahknow,
To answer your questions: I tried working on other people's stuff but wanted to work on my own and it's easier and more profitable to just get up and go to my day job. My intent was to make money to by tools to build and eventually sell my own parts. I just decided to go ahead and buy the tools. I've spent the past couple of years trying to learn how to maintain and operate the 5 axis machining center I stuffed in my garage. The Quad Turbo runs but I have some oil return issues with the lower turbos to sort out. My plan is to finish it when I have parts to show off.


Kevin
 
See less See more
#3 ·
I have a 26' Teleflex on the way. $16 new on ebay. I did some more testing yesterday and couldn't get the water back out of it. It turns out I was completely wrong about how and where pressure would be generated in the brake. Today I added a couple ports on either side of the rotor perpendicular to the shaft on the bottom. It comes out now and was able to get the motor to rev past 1700 rpm. I'm trying to combine the data from the OPTO and Megasquirt because I don't have the RPM part working on the OPTO, but I do have torque readings. This pull was part throttle from 2000-4000 rpm. It has a carb with Megasquirt running timing right now. I think the dips are where it needs more timing.

Kevin
 

Attachments

#7 ·
Gow,
I started from scratch. I drew the brake in cad and had it laser cut. The rest I did on the fly. This brake is my first attempt. I already have some ideas for the next gen. I will most likely mill the next rotor from solid steel and cast the stators from aluminum. My intent is to make them easily stackable as my power requirements go up. The next one will probably be smaller but it seems the volume (about 2 gal) of this one may let me run it with less water flow. We'll see how much steam it makes when I try doing some steady state testing under boost.

Right now I'm working on electronics. I'll reiterate some of our previous PM to bring others up to speed. I'm planning on building a microprocessor based card to do all my I/O with a single GUI, but while I'm learning, I have a kluge of stuff. I'm using a Megasquirt-II to datalog from the engine. I have an Opto 22 ether I/O to collect torque, RPM, and MAP. I'm planning on using a DATAQ di-148u to collect 8 egt's. The Opto and Dataq can easily be combined in Excel.

I've spent the past few days trying to sort out collecting RPM. I have a LM2907N ( http://www.ti.com/lit/ds/symlink/lm2907-n.pdf ) using the minimum component tach circuit. I've tried it with an inductive clamp and a VR sensor pointed at different rotating things. Both were erratic. Now I have a trigger on the back of the brake and am getting a steady signal I just need to calibrate it. I'm about to just stick a dc motor on the back and read voltage off of it. If anyone has any simple ideas to convert RPM to 0-10vdc let me know.

Kevin
 
#8 ·
You might need a de-bounce circuit.

I think I told you but I used the PAK VII chip to measure pulses. It worked well for RPM. I was actually able to read PWM from radio controlled airplane receivers using it.

With the PAK-VII (and other time chips can do this as well) I was also using this to measure timing measuring the difference in time from the crank trigger to the spark relative to RPM.

The problem I had is it worked briefly on the actual engine because I don't have the spark induction isolated enough. It was locking up my USB interface. What I needed was an opto-isolater....at least that was the next thing I was going to try.
So from that I would look at de-bounce and ops-isolators.
 
#11 ·
I guess my question with that for you would be can your software communicate with something such as this instead of reading an analog input? This would communicate with serial communication which would basically ask the chip what was the last high or low duration on pin 0-7 (8 channels in total).

If not you could use say an Arduino to do the communication with the chip and sort out the details then return with an analog output value it could read.

Infact if you google "aurduino rpm" you'll come up with MANY ways to read rpm which may be easier then the PAK VII. With the arduino you could pick up a several inputs your program could not process and output them digitally or analog.
 
#13 ·
I notice you have no way to adjust the load ,did i miss something in the pics? :)
Also, you need gland seals on your shaft to cope with the pressure .
. Froude dynos use a valve that opens or closes to adjust the water which is fed in one side and it gets pushed out the exit port .The exit is close to the inlet . Froude invented the dyno to test Ship engines up to 110,000 Hp in 1877.

Here is one in Finland which is being used to test a 27 liter Rolls Meteor engine. This is the land based version of the merlin V12 and were commonly found in Centurion tanks.
Meteor on Antique Froude dyno.
 
#16 ·
mercmad63,
The shafts have seals that press into the stators from the outside. I have solenoid valves controlling the load. I will probably add a larger motor controlled valve later if PWM doesn't work with solenoids.

Gow,
I hadn't seen that one before. Land and Sea has one also, but the AEM is cheap. I had thought about measuring torque on the shaft like that with some strain gauges. Figuring out the wireless is what has kept from it for now.

Wags,
I had just posted some pics of my adjustable stand on the other thread. I started getting some private messages and answered some questions over there so I started this thread. I have pics of the unassembled brake in my gallery if you want to try to make one.

Kevin
 
#18 ·
Got home tonight and started working on EGT's. Normal EGT probes are shielded, but they aren't cheap and I think bare ones will respond faster. For dyno use they should be plenty durable. I have a bare K-type I use to measure the temp of molten aluminum, which is super corrosive, and it has survived. I have 50' of K-type wire I bought for like $30 online. I decided to try TIG welding the junction. It looks just like the store bought ones. Seven more to go.

Kevin
 

Attachments

#20 ·
That water brake looks nice. I have put together my own rolling road dyno using an old eddie current retarder. I have built my own controller/data aquasition system. I would like to build an engine dyno in the furture as well, so Im keen to see your progress with this.

From my research it seems most control systems use motorised valves (servo valves) to control the load. I'm interested to hear how your PWM valve setup goes. Can you tell me what valves you are using with the PWM setup?

Cheers

Daniel.
 
#21 ·
Me too...........

I have an eddy current roller dyno and have started on the home built closed loop speed control system for it, but it is going to be a very long term project.

Controlling a water brake is pretty much the same, except you need either a motorized valve, or speed controlled pump to vary dyno load, instead of just varying the current through the retarder which is a lot simpler.

But the cosed loop speed control system is going to be pretty much the same, whatever type of dyno you are building.

My system uses an optical chopper disc that generates a frequency, that is used to both control the dyno loading, and read the roller speed for calculating power.

The engine speed setting knob varies the frequency of an oscillator.
The frequencies coming out of the optical chopper, and the oscillator are compared in a frequency/phase comparator chip, and that ramps the dyno lad up and down to hold engine speed EXACTLY where I want it to be.

So I set my oscillator to the equivalent of maybe 4,200 rpm and the engine idles under zero load.
I open the throttle and the revs rise, until the engine reaches 4,200 rpm, then the dyno load builds up very fast and holds the engine at exactly 4,200 rpm.

If it drops to 4,199 rpm the load backs off very slightly, if it rises to 4,201 rpm the load builds up very slightly.
It hold engine rpm steady as a rock at any throttle opening above idle at whatever rpm I want to test at.
 
#22 ·
Controlling the speed of the pump is an interesting idea, I haven't seen that before.

My system is based on a microcontroller which reads pulses from a hall effect sensor. It calculates the rpm and uses a PID loop in the code to control the speed. It also does ramp runs etc. The controller also handles the other sensors like the load cell and air temp as well as other spare analogue channels. The controller then feeds the info to the pc for displaying and logging. I have been working on my system for 3 years and have 2 mates in Melbourne who have my system on their dynos.

Cheers

Daniel.
 
#23 ·
I am using a microcontroller only for monitoring outputs and digital readout.

The control system is all in hardware, and really very simple with few components.
It uses a proportional integral loop (PI) and drives the retarder coils through a PWM system (500Hz), which responds much faster, and is more linear in operation than the usual mains phase control.
The speed and linearity make PI loop tuning far easier, and with a much tighter and faster acting load control.

I have seen a centrifugal pump run off a VFD to pump water from a cooling tower sump into the water brake. The water brake itself discharges hot water through a metering orifice to the spray at the top of the cooling tower.

That seems to respond faster and give much better load control than some of the very slow to respond commercial motorised valves. You will probably need a pump anyway, so might as well control the pump speed directly rather than throttling it.

I am not at the stage of sweep testing, and may never bother with sweep testing.
But for mapping EFI systems, what I have works wonderfully well. It is all in pieces at the moment, must get back into it.....

If you are going to be sweeping your water brake, control of water flow volume into the brake will need to be both fast acting and fairly linear.
If it is slow and hugely non linear, you are going to have a lot more difficulty tuning your PID loop for both speed of response and stability.
 
#24 ·
I have an old Clayton chassis dyno. It's water brake has a water jacket with a temperature actuated water heater safety type valve on it. The water used to load it recirculates in the pump and is cooled by the jacket. It's only good to around 400 whp so I plan on upgrading the brake but it works pretty well within it's ratings. I'd like to play with an eddy current brake, but I haven't found one in my price range and they look expensive to make. Right now I'm just using $15 sprinkler valves on the engine dyno. I think they're way too slow to control load other than fill and drain at one rate. I will most likely end up using a stepper or servo controlled valve. Here's the load cell cal I've been working on this afternoon. I made an arm to easily locate my rusty NIST traceable barbell weights at two and three foot.
 

Attachments

#25 ·
Hi,

Great thread.

I'm very interested in your progress and planning a similar project myself.

I'd love to see more details of how you are coupling the dyno to the test engine. I noticed you are using an equal number of vanes in stator and rotor - I read somewhere that normally there would be a different number?

How is your project progressing?

Cheers.
 
#26 ·
I'm using a PTO flange from a jet boat supplier. They are drilled to bolt to a standard spicer flange. This one accepts 1350 a u-joint. I have an H-bar, double cardan cv, and a 1.5" 10 spline yoke to slide over the shaft on the brake. I don't know about the uneven numbered vanes, but it'd be easy enough to experement with. Right now I'm working on noise control and still learning about microprocessors so I can cheaply collect more data. I have the hardware to log 15 thermocouples, and the load cell, and 70 digital i/o points through one $10 chip. I just need to learn how to program it.

Kevin
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top