Hi,

That's not normal. After calibration it should hover around 0 Nm, and noise is typically in the 0.1-0.2 Nm range or less. This does depend on the size of the load cell and the gear ratio (it is calculated engine torque noise you see), but what you see is way too much. What is the size of load cell you have? 

Your system also does not show 0 Nm at rest. You should check first that there is no mechanical binding or similar. You should use rod ends on your load cell (sorry if this is obvious, but some people don't know). Your reading should come back to 0 after calibration, or something is mechanically wrong. It could also be the load cell, but I have never seen them fail like that. 

We had the same type of noise and calibration fail when we first installed YourDyno on our eddy brake chassis dyno. It turned out to be the torque arm was not completely free to move, it touched the side of the safety thing that catches the arm in case the load cell breaks.

For the noise, if it is not mechanical, then try grounding the box and make sure the cables are not routed together with high powered cables or close to noisy power supplies or fans or similar. If you don't get anywhere with that, you can add a capacitor (try 10uF ceramic) between signal+ and signal- on the load cell. But don't start there, I bet something in the setup causes it, and it is better to remove the cause.

See a noise test I did with a previous version of YourDyno here:  https://www.youtube.com/watch?v=mEHJjJzAn9s&t=0s. The newest HW gives similar results. 

BTW are you still using a very long USB cable? Check with the standard cable.

Here it is after another calibration.
https://youtu.be/x293oE45eEw

The load cell wiring does run alone the retarder fan wiring but thats not turned on until after a run and you can see that make lots of noise when its switched on so it stays off during runs.

Using the USB cable it came with it.
Ill have to check the ends on the load cell. Lol had it apart yesterday and cant remember if it had rose joints.
I guess I can test the load cell just on the bench with a short bit of wiring to see if there is still noise there since it is 30 years old.
As well as try grounding the box. Im guessing grounded to the PC case is the best place.

Never had noise problems on my hub dyno, indeed, last sunday i dissamble one for Paint everything, assemeble everything again and checked if need again to calibrate and not, it was needed. 

https://www.youtube.com/watch?v=-1yHntLzeYI

I use a 1000kg loadcell. Have calibrated it with 400kg load. I get fluctations at zero load with retarder not turning, on aprox 5Nm. Should i worry? 

i use a very expenssive load cell. 

What is the engine torque it will measure when load on load cell is 1000kg?

It's 3300nm

sorry 3300nm retarder torque. aprox 1650 engine Nm. 

wow i wanna see that torque arm. 400kg is a lot and 1000kg overkill! what do you test on your dyno? mine is 500kg load cells and 24kg for calibrate.

Its a tractor retarder dyno. A 300hp tractor usually haves 1500Nm of engine torque. So my dyno is actually a bit to small for the newer ones. Gearratio for power take off is usually about 1:2, so retarder torque is double of engine torque. 

I calibrated it with a torque arm on aprox 1,5meter and 80kg weight which gives aprox 1200Nm. Length of torque arm is 0,33m

So your noise is about 0.3% of full scale. I typically get less than that (< 0.1%), but in practice 0.3% is not really going to make a problem since there are other contributors. 

Pure noise can be filtered. Make sure that the IIR filter is on and adjust the Noise filtering such that the curve is smooth but you still see the details.

So been though everything twice and got a new load cell on it. Managed to 1/2 the noise but there is still quite a bit there.

Was 10uF 16v ceramic capacitor the correct one since every where iv tried dont have any source for one so im stuck buying a reel of 10 of them.

10uF is super common so you should be able to buy 1. Voltage rating is not important.

But do a noise recording with the noise filters on and check in time domain. You will not see much.

Posted by: bmgjet

Iv got a bit of noise on the load cell reading and wonder how I can go about cleaning that up.
Its all new wiring ran under ground with shielding but im getting 2-3nm spikes in readings during steady state some motor and tune on old system had 0.2nm spikes.

The old filter box from the old system on the dyno seemed to keep readings within 0.2nm spikes over old wiring so im wondering If I can wire that in to the yourdyno control box over the old wiring to see if it rules out the problem. Filter sits in dyno bed it hooks directly into the load cell then just has 2 wires output. Earth and 0-5V or 0-10V depending on switch.
Is there any way this can be connected up.

Here is a extreme example of whats happening by turning the arm length up with no load on the loadcell.
https://youtu.be/WnJR-JAazU0

And iv attached a image of the filter box.

Checking voltages with multi-meter.
USB voltage iv got cycling 5.01 - 4.98v (all usb ports on computer follow same pattern).

Wideband/Map/Knock are powered off PC PSU with molex plug. Doesnt make a difference having them disconnected.

IMG_20180829_105915.jpg

Have you solved this problem?