[wanni]

Hi,

got this setup on my bench. Looking forward to get it in my sailing boat to get rid of the build in ICE. Have the following problem:

Throttle is ET-164 with "N" in the middle. Pushing throttle forward will turn the motor in forward direction (Got the Sevcon Display as well, and it says "F" too) but getting back to "N" (display will show "N")the motor keeps turning.

I bought the setup files from thunderstruck for this combination. and at least it is turning, but something must be wrong.

any hints to where i should look at first?

I checked the encoder (UVW) with a scope as mentioned in the manual and put the wires togehter like it is wiritten down there.

Wanni

 

[cricketo]

You want to check the raw output from the throttle as seen by the controller.

 

[wanni]

I will do, but if i understand it correctly the controller is seeing the right values because the sevcon display is showing the correct state of the throttle (F-N-R)

 

[cricketo]

Yeah, I was thinking a bit more about this, does it eventually come to a stop, or it just keeps going ?

 

[wanni]

It still keeps going, even if i go over neutral to reverse. It does not speed up either . Just spinning and only stops when i cut the power

 

[cricketo]

Understood. So again then, you may want to check the configured throttle range in the controller and compare it against the actual raw value reported. I don't know how that display works, but the motor spins because controller thinks throttle is being applied

 

[wanni]

Hi, eventually had time to get some data

throttle is an ET-164 (hall sensor with voltage output)

voltage is 0V in „Neutral“
Voltage is 5V in either maximum (forward and reverse)
The digital inputs 1..8 states are switching between
Forward 0x01
Neutral 0x00
Reverse 0x02

 

[cricketo]

so when you run it forward and then release the throttle, does it come back to 0V while the motor keeps spinning ?

 

[wanni]

Exact, but it will always run at the same speed. It starts turning and holds the speed despite the actuall position of the throttle

 

[cricketo]

Well, maybe define a deadband in there ? Like have the throttle start at 0.5V instead of 0.

 

[wanni]

Deadband is up to 1V - the motor wont‘t start until you reach the 1V.

Something else, may be important:

Connection diagram says
U - M1
V - M2
W - M3

And the encoder wires to
5, 17, 29 in the order they are listed in the specs

If i start the motor it will start turning and will immediately stop with an overcurrent failure.

Afterwards i measured the encoders and motorwires with a scope and switched positions of V and W. Then i get the failure wich i listed above.

 

[cricketo]

Understood. I don't have any other ideas at this time. I will try to get into one of my Sevcons later this week and see if anything else stands out with the throttle settings.

Not sure if motor connections matter for the original problem. I believe incorrect motor wiring would immediately manifest in motor going crazy or controller flashing an error.

 

[wanni]

It‘s getting better
I tried ALL combinations of the motor phase wires and found a working combination.
Motor is starting smooth and controllable in F direction. Works like a charme.
throttle to zero - motor stops

bit when i go over to reversethe motor is sort of rumbling - shaking. Then with more power it is turning the same smooth way as forward.

what is this „dead point“ between forward and reverse?

 

[cricketo]

ME0907 seems to have a UVW encoder. It is possible that yours isn't wired correctly. I was actually just troubleshooting a similar problem on ME4201. If you have an oscilloscope, you can hook it up to encoder output channels to see if the waveforms are correctly offset one from another when you spin the motor by hand. Otherwise you can look at the RPM reading in the DVT while spinning it in both direction to see if the numbers are consistent, and aren't jumping back and forth between positive and negative.

 

[wanni]

I tried it with the scope - but that didn‘t work out
I found an encoder setting in the dvt it says something like
Uvw
Uvw reverse
Sin
…

and as i remember it correctly in the original dcf from thunderstruck ther is the „reverse thing“ checked

 

[cricketo]

well, you still have the channel assignment dictated by the wiring on the controller. It basically defines U, V, and W inputs. So the question is if U and V are getting correctly offset signals, and then also the same between V and W. For this test you don't even need the controller involved - as long as you have the voltage applied to power the encoder, you can scope it while spinning the shaft with a drill or something.

My understanding of the reverse thing is basically that Motenergy docs say
S1 (U1), S2 (U2), S3 (U3) for the encoder outputs, and you'd think S1 is U, S2 is V and S3 is W. But they're actually in reverse - S1 is W, S2 is V and S3 is U, so Sevcon allows you to work around that by picking "UVW reversed" option so you don't need to rewire the controller.

Still, if you were changing the phase wires on the motor, there is a good chance you now need to swap the two encoder outputs.

 

[wanni]

So, what i did:
I tried EVERY combination of U, V, W with EVERY combination of S1, S2, S3

there are only a few running combinations but all with the same „failure“.
throttle forward - the motor starts turning smoothly and speeding up as expected

throttle backward - motor is „hanging“. After wiggling F/R some timesmotor will spin backward and also speeds up as expected
Edit: if i hold the throttle in reverse on minimum and turn the motor initially by hand it will also keep turning

(i found a combination where things behave the same inopposite directions, so there is no failure in the motor)

is there a parameter in the dvt which i could improve to get rid of this behavior?

 

[cricketo]

I tried EVERY combination of U, V, W with EVERY combination of S1, S2, S3

Isn't that 6! ? Yields 720 combinations...

 

[wanni]

It is 3! x 3! Only 36 Kombinations. It doesn’t make sense to locate phase wires on encoder positions

 

[cricketo]

I've got no other suggestions. Perhaps indeed there is some different programming for reverse, like maybe check the torque/cutback maps and see if there are any differences between the directions ?