Problem with threading

[Chaz]

If you make just one pass, rather than cutting the thread all the way to depth, do you get a clean cut with the correct lead?

It looks to me like the starts are not synchronized. That would indicate a problem reading the index pulse (Z channel) from the spindle encoder.

The index pulse should come around just once per revolution, at the same place every time. The data sheet for your encoder indicates that it is gated with A and B high, so the index pulse is only one count wide. On the PID Configuration screen, when the index pulse comes around, you will see an asterisk ('*') flash next to the label 'N' in the left-most column, for the 5th axis (spindle encoder).

Check this first while turning the spindle by hand. Does it come around once per turn, in the same place every time?

If you have an oscilloscope, use it to watch the Z and /Z wires of your spindle encoder, first while turning the spindle by hand, then while running it with power from the spindle drive.

I am guessing that you will see nice clean index pulses once per turn when you turn the spindle by hand, and that you will see a lot of noise (which the controller may read as index pulses in the wrong places) when the spindle drive is running.

My assumption is this too, inconsistent index pulse.

[Chaz]

Thanks. I will check with the oscilloscope and manually. Maybe you're right. It’s just very expensive to buy an encoder from Centroid in my country. With shipping, its price is about $ 650 ...
Therefore, you have to choose from what you can really buy here.

Which country? You dont need one from Centroid. I buy decent encoders from Ebay in the UK for £50 - 60.

[avangard]

I'm from Ukraine.
I also think the problem is in the encoder. Counts from 0 to 16384 and gives an asterisk to 16384. When reversed, it can give an asterisk at a value of +2. If the counter had a value of +2 when returning, then with a positive rotation, an asterisk may appear at a value of 16386, but when returning, give the asterisk already at 0.
There may be a problem in the connection of the spindle and encoder. Now I'm redoing the encoder to a belt drive 1:1
Thank you so much for the help you are giving me. This helps a lot.

[tblough]

It looked like your encoder was connected just fine with the coupling directly to the spindle. That should work fine although the downside of that method is you loose the ability to handle long stock through the spindle. I suspect your problem is with the index pulse itself and not your mounting method. Either the encoder is having problems generating the index pulse consistently, or you have noise making the control think it's receiving an index pulse when it should not be.

[Dave_C]

I'll weigh in on this as I had the same looking threads when I first set up my lathe.

Same issues but my Asterisk seemed to appear at the same index point, still I could not make a repeat pass for a thread. (All other settings were correct)

The solution was to buy an encoder from centroid and also buy the cables from centroid!

Problem solved and I have not had any issue since. So after a lot of research we discovered that my encoder was fine but the cables that came attached to it were not twisted shielded pairs and I was getting noise into the encoder signal causing random index pulses. (Also had quadrature errors on the spindle and had to disable them but was able to re-enable them after the new encoder install)

As we all know, or should know, VFD's can induce a lot of high frequency noise and we have to use the proper shielded drive cable, single point grounds and all the rest of the "best practices" for wiring a CNC control.

Dave C.

[frijoli]

Slow it down for testing and finding the errors. Once you find it you should be able to thread that fast if everything mechanical is tight, and you maintain a good starting distance.

[avangard]

Hi
Today moved the encoder and encoder cable away from the VFD and the motor. Native encoder cable twisted pair. Checked. Nothing helped. At any speed, the thread is not correct.
I noticed the following. If you turn on the spindle and set 600 rpm on the screen, the revolutions float stably between 590rpm and 610rpm.
The encoder shows approximately 10 revolutions of the spindle as 590rpm and the next 10 revolutions shows as 610rpm.
The same thing happens at any speed. If 1000rpm then it will be 990rpm - 1010rpm and so on.

[Dave_C]

Thanks for posting the pictures, I now see why you can't thread properly and why the index pulse does not appear in the same spot each time!

You HAVE to use a toothed timing belt to drive the encoder and not a round belt or a vee belt as they have a small amount of slip. No way to get around that!

The encoder must stay in perfect sync with the spindle in order to make multiple threading passes. The only way is to lock it to the spindle using a timing belt or in some cases an encoder that mounts to the spindle directly.

No Vee belts or round belt, no matter how you mount them, they will allow for a small amount of slip.

I enlarged the pics and the belt does not look like a toothed timing belt. My apologies if I am not seeing that correctly.

Dave C.

[avangard]

Today I set the encoder this way. Yesterday it was tightly connected to the spindle and the result was the same as today.

[avangard]

There is one more question. I want to put a servo motor on the spindle drive. It is possible to purchase a TECO 1.5KW 3000rpm 4.8Nm 220V servomotor.
Judging by the description, this servo motor may come up. How to solve the reverse issue I have an idea. I'm not sure if I can use the native encoder of the servo with Centroid. I will be very grateful for the help. Datasheet on the servomotor in the attachment