MiniDyno motor test bench..scaleable 0-50kW

[larsb]

I'm building a small dyno that can test most motors here, first revision max power is 30kW. It will be scaleable so it can probably go up to the double power with changed sensors and bigger flywheels. Or smaller with just other sensors.

I have based the frame on alu profiles in 90x90mm. It is an expensive choice but it allows for simple reconfiguration and scaling of the dyno just by changing/adding profiles.

The first revision is a flywheel dyno. It is the simplest form and easily gives repeatable results. Downside is that steady state testing cannot be performed, it will be best for full load optimisation - (hotrodding ). I think it can be used to optimise phase angle seeting for hall sensors, finding saturation current of motors and optimal gearing etc.
It will be possible without excessive altering to add a load cell and brake if someone wants to expand it into a brake dyno.

I'll put up all needed stuff and info here so it'll be really easy to do a version of it. It's a slow build as always for me so it'll be ready in something like a year

Edit: or more like 5 years
WIP estimates
Estimated total cost for me: about 2000 USD
Total cost for the diy guy with time&tools: 1000 USD
Weight: 120kg
Dimensions: 58x58x40 cm
Flywheel diameter: 54 cm
Flywheel material: S355J steel, 15mm thick
Flywheel weight: 26kg / each

View attachment 1

 

[larsb]

Reserved

 

[larsb]

Electronics
Dyno control unit: www.yourdyno.com (or Ardyno https://ardyno.weebly.com/ if you want to build cheap diy with arduino)


Sensors
Rough specification current sensor:
Minimum accuracy 1%
DC sensor min 300A / AC sensor min 600A
Output +/-5V or lower

L03S400D15 from Tamura
https://www.mouser.se/ProductDetail/Tamura/L03S400D15?qs=sGAEpiMZZMsPDRSCoHb1X%252bEtsCkzqCDrdETjJXYYyX0%3d

or

HASS 300 from LEM
https://www.digikey.ca/product-detail/en/lem-usa-inc/HASS-300-S/398-1065-ND/1680531

or

HAIS400 from LEM
View attachment LEM HAIS 400.pdf

or

DHAB S/133 from LEM
75 to 750A capable but quite expensive
Datasheet:


Rough specification voltage sensor:
Minimum accuracy: 1%
input VDC: min 100V
Output +/-5V or lower

Voltage sensor:
Voltage divider 40kohm-->2kohm where measurement is done over the 2k resistor. Small (10uF?) cap for smoothing noise might be needed.

Signal isolation/ buffer board:
mikrobuffer proto from mikroelektrika
https://www.thedebugstore.com/mikrobuffer-proto-board-mikroe-324.html
or Sparkfun opto isolation board
https://www.sparkfun.com/products/9118

These are my best guesses right now, if there are better options for sensors, please propose them!

 

[larsb]

Frame and mechanical parts
First set of parts for the axle arrived

12kg solid steel loveliness:
Bolt-on taperlock hubs for the flywheels and brake disc, 30mm pillow block bearings and drive sprocket.
View attachment 1

 

[larsb]

I got a present in the mail, two LEM HAIS current sensors, ready for use

View attachment 1

 

[larsb]

I've ordered the profiles for the frame, it was 600USD total (including fixations, tax, shipping)


I will get the flywheels water jet cut next week.

Then it's "only" the electronics left

 

[larsb]

Profiles on the way.. But flywheels are delayed. Seems i don't have top priority at the water jet cutter 8)

 

[liveforphysics]

I would use brake drums from semi-trucks. They spin nicely and when they're worn-out you can get them for scrap iron prices.

 

[larsb]

Good tip, thanks!

I am starting small and easy with a flywheel setup but eventually and if i keep the interest then there will be upgrades. I found some small eddy brakes for around 1500 USD new and those would be my first choice if i get some tax returns
I'd rather build cleaner but more expensive since i have limited time.

 

[liveforphysics]

I'm excited to see it running and the learning that comes out of it!

I like that you could make that style of dyno configured towards testing small motors or large motors just by bolting more flywheel plates on or removing them.

 

[larsb]

I received two flywheels and sensor trigger wheel from the water cutter. Fit to the hub is excellent, it just barely drops into place.
Coming together nicely

 

[larsb]

I forgot this project for a while :wink:

Started to build the dyno frame, feels super solid.

Universal connections for the profiles turned out to be for the wrong profile due to incorrect data on the site i bought from.These have been replaced now.. I bet there are more stuff missing when i get further into the build
Business as usual

 

[larsb]

Shaft collar mounted to sensor wheel:


Time to learn how taperlocks are mounted

 

[larsb]

I got all dyno parts assembled tonight! It’s a monster All went together without any surprises and there is space for one or two extra flywheels in the frame.

The critical part now is if flywheels need major balancing 8) Too bad i haven’t got any spare motors lying around..

Anyway, next thing will be electronics and measurement system

 

[larsb]

I set up the dyno computer tonight, installed yourdyno program and hooked up the rpm sensor. Works without issues. Now i ”only” :wink: need the current and voltage sensors hooked up and working plus a motor/controller kit to test.

 

[ferret]

Looking good!

Keep up the good work,
Avner.

 

[larsb]

Thanks!

Built a trial circuit for the voltage sensing tonight. First time i’ve ever used a breadboard


47kohm resistor and 2.2kohm in series, sensing done over the smaller resistor to take 100V down to less than 5V. Added a cap to smooth out noise.
Could work

 

[larsb]

Turns out that rieju drac -99 moped brake discs (which i happen to have) fit perfectly on the taperlock hubs i bought so soon i’ll have brakes on the dyno.

I think i'll use the dyno frame as a base to mount all parts except the computer (if there's a lot of vibrations during a run). I cut and glued two 15mm plywood boards to make a 30mm thick table top for mounting the controller and batteries.

I might build a brake dyno sooner than i thought.. but it’ll have to wait until i get the basic setup done. Next step is hooking up batteries, controller and current sensors.

 

[Ohbse]

Awesome work Lars! How will you go about calculating the inertia?

Last weekend I did a bunch of reading about doing exactly this! I came across one quite excellent solution which was an enormous AC induction motor. All of the stator copper was removed and a large hole cut in the top of the motor housing, that way a tire can be run directly on the rotor. In addition, the manufacturer was able to dig out the precise measurements of stator inertia.

Your solution is very modular and making up motor mounts with the extrusion lego will be easy

A brake purely for safeties sake seems wise. The majority of dyno loads these days I've seen are eddy based, essentially just some motor coil electromagnets with a mechanism to vary the distance to the rotating mass to adjust torque.

 

[larsb]

I get inertia from the CAD, guess the simplest way would be to connect a weight to the wheel and measure the time it takes to drop a certain distance.

I’ve looked at eddy brakes but it seems powder brakes are cheaper. I’ve also read that QS motor dyno is built on a powder brake

 

[larsb]

Voltage divider works, transfer function is vout=vin/(70.3/3.191) so 22x lower output than input.
if my noname DMM is to be trusted..doesn’t really matter as i am mostly looking for relative measurements on the dyno.

 

[j bjork]

Interesting
I have not been reading there for a while, but at savarturbo.se they have been building some dynos. One for moped had a rim from a car I think.

 

[larsb]

I saw some of their builds when i researched, 800hp diy dynos are impressive. Not my scale, yet! :wink:

Hooked up all sensors now - I am doubting the connections to the buffer board.. i wonder if it needs separate power supply.
connections basically go

from yourdyno box to:
+5V common connection point
common ground connection point
Yourdyno input1 to buffer output 0 (volt measurement)
Yourdyno input2 to buffer output 1 (battery current measurement)
yourdyno input3 to buffer output 2 (phase current measurement)

from +5V point
to current sensor1&2
to buffer board common voltage Vcc on analog input side
to buffer board common voltage Vcc on output side

from ground to current sensor 1&2

from current sensor 1&2 signal to buffer input 1,2
from voltage divider to buffer input
from voltage divider ground to ground point



I take 5v from the ”yourdyno” box, it can drive 100mA and the sensors only need 20mA each.

 

[larsb]

Yey! It works!!
Voltage is displayed correctly and the currents seem to function as they have a small offset from zero. RPM sensor counts when the wheel is turned.

So happy that no magic smoke was let out! Now i need a motor and controller that's not on my bike :wink:

 

[larsb]

Ordered a QS 3000W mid drive motor for chain as my current belt drive setup with the same motor is worn out. Corona virus has put a stop on QS so it'll be a while before i get it.

It'll be interesting to see what the dyno is good for
It'll be interesting to see what this motor is good for on the dyno