M3 Snnn - Programmed vs Actual Speed

Nightowl

I've finally got my milling motor working properly with On/Speed/Off, but the actual motor speed is different from the programmed speed.

For example, using various S values with the M3 command, the actual speed (measured with a laser tachometer) are those shown alongside (I appreciate some of the S values are below the minimum speed of the motor and DWC would probably return an error if I tried to run the motor out of it's capabilities, but I want to see what happened!):

S1 = 4074rpm
S10 = 4074 rpm
S50 = 4280rpm
S100 = 4569rpm
S250 = 4735rpm
S500 = 4980rpm
S750 = 5145rpm
S1000 = 5306rpm
S2500 = 6456rpm
S5000 = 8430rpm
S7500 = 10320rpm
S10000 = 12250rpm
S12500 = 14080rpm
S15000 = 15980rpm
S17500 = 17940rpm
S20000 = 19820rpm
S22500 = 21800rpm
S25000 = 22780rpm

The minimum speed of the motor (a Sorotec PV-ER) is 4000rpm at 0V and 25000rpm at 10V supplied from the PWM board, but is it critical to set the config line correctly, i.e:

M950 R0 C"vfd" L**4000**:25000 Q1000

, rather than the L0:25000 values I have at the moment, and is it likely this is causing the issue?

Thank you.

jay_s_uk

@Nightowl yes, it won't be scaled correctly otherwise.
Set the scale right and then using the screw on the pwm to analog converter to adjust the speed so it's closer to the requested speed

Nightowl

@jay_s_uk

Thank you. I'll do that tomorrow, as it's too dark to go out now

Nightowl

@jay_s_uk
That's made a significant difference, although I will need to tweak the converter somewhat.

These are the latest results:

M3 S...
4000 = 4050rpm
6000 = 6266rpm
8000 = 8150rpm
10000 = 9955rpm
12000 = 11665rpm
14000 = 13473rpm
16000 = 15304rpm
18000 = 17065rpm
20000 = 18867rpm
22000 = 20726rpm
24000 = 22425rpm
25000 = 22608rpm

My normal range of speeds would be between 10000 and 18000 for my CNC, so they're not too out of whack, I'd like to get them as accurate as I can...

jay_s_uk

@Nightowl do you find the RPM is wondering about a bit? because it does on mine. Or is the value staying stable?

Nightowl

@jay_s_uk
It does seem quite stable. The only exception was with the last list of speeds, where one of the speeds 'hunted' a bit, like a vehicle engine with a touch of fuel starvation, if you know what I mean?

jay_s_uk

@Nightowl yea, that's what mine sounds like. Maybe I need to revisit my setup

Nightowl

@jay_s_uk
Whilst checking the speeds, I also checked the noise. At...

Ambient = 45dBa
S4000 = 65dBa
S10000 = 75dBa
S18000 = 83dBa
S25000 = 84dBa

Nightowl

So, I've done a bit of tuning with mixed results.

M3 Input - Untweaked	- Tweaked
S4000	– 4050		– 4060
S6000	– 6266	 	– 6550
S8000	– 8150	 	– 8485
S10000	– 9955	 	– 10400
S12000	– 11665		– 12270
S14000	– 13473		– 14030
S16000 	– 15304		– 15870
S18000 	– 17065		– 17680
S20000 	– 18867		– 19390
S22000 	– 20726		– 21190
S24000	– 22425		– 22880
S25000	– 22608		– 23050

There doesn't seem to be a regular variation, so I was wondering if anyone else has managed to work this out and achieve accurate speeds.

Thanks

o_lampe

@Nightowl
Did you measure RPM under load or without?
I think, the RPM control loop works best under load? (except with 3-phase PM brushless motor)
All other motors, (brushed DC or asynchron) are less accurate

o_lampe

@Nightowl said in M3 Snnn - Programmed vs Actual Speed:

I'll do that tomorrow, as it's too dark to go out now

How dare you to name yourself nightowl ?

Nightowl

@o_lampe said in M3 Snnn - Programmed vs Actual Speed:

@Nightowl said in M3 Snnn - Programmed vs Actual Speed:

I'll do that tomorrow, as it's too dark to go out now

How dare you to name yourself nightowl ?

Nightowl

@o_lampe said in M3 Snnn - Programmed vs Actual Speed:

@Nightowl
Did you measure RPM under load or without?
I think, the RPM control loop works best under load? (except with 3-phase PM brushless motor)
All other motors, (brushed DC or asynchron) are less accurate

No, I didn't test it under load to be honest. It may have been optimistic on my part, but the blurb indicated the speed is maintained under load, so I'm wondering if I should test the voltage output from the PWM converter (as supplied to the milling motor) to see if the issue is there.

I have more confidence in the German-built milling motor, than I do in the converter and/or the digital tacho I'm using to test the speed.

o_lampe

@Nightowl said in M3 Snnn - Programmed vs Actual Speed:

the digital tacho I'm using to test the speed.

I suspected that too, but those would usually skip pulses. But in you table it reads more pulses at low RPM. Probably false triggering...
I had good results with hall sensors. A simple bicycle speedometer (with programmable wheel circumfence) worked well on my wind turbine.

Nightowl

@o_lampe
Perhaps it's time to start another hobby....

The idea of using an Arduino with a hall sensor sounds like fun, and I've seen an Arduino kit on Amazon which seems to fit the bill (except for the hall sensor, which is cheap enough), but I wonder if this would be more accurate than a laser measuring the reflection from reflective tape stuck to the spindle's collet nut - which is the method I currently use?

Of course, I'd need to learn how to program an Arduino!