Spindle encoder with one intiator but 4 pulses per revolution

This is a follow on from @Sviper's post: forum.linuxcnc.org/38-general-linuxcnc-q...e-emulate-an-encoder which I was hoping I could use directly. But on dismantling the sensor on my mill it turns out to be a photo interruptor with four 'windows' per revolution. I suspect this is a fairly common implementation on the import mills.

I haven't yet tried to implement the above post as I'm waiting for the hardware to arrive (I'll be running two photo interrupters - one for the mill's 'vfd' display and new one for the 7i96 to consume so I can compare the numbers) but I'm if I'm reading it correctly, I'll be reporting 4x the rpm.

Is there a way to apply a 'divide by 4' in the hal config?
 

Thank you Tommy - in the Sviper post, Andy Pugh replies saying to set counter-mode to 1 which according to the docs would mean x4-mode will be ignored.
I think perhaps I may have asked the wrong question - all encoders 'pulse' multiple times per revolution, so the question I should be asking is where do I set the encoder resolution for the spindle?

Edit: Now that I've thought about the question, position-scale seems like a suitable candidate?

Aha! Thank you. I think I'd searched my .hal file for 'encoder' before actually assigning 'Spindle' to the 'Quad Enc-A/B/I' inputs - Doh!
I will be able to test as soon as the parts arrive (and I summon my inner Binky to make a custom bracket)...

You know I can't leave this alone, don't you? So I've started designing a quadrature encoder+index that I can strap to the spindle inside the mill head (where the current single pulse encoder is):



This is currently designed with 128 'windows' which if I understand quadrature correctly and set the two readers x.5 windows apart I should get 512 position 'indications' per revolution - rigid tapping FTW?
I've tested 3D printing up to 250 windows but I'm not sure the opto switches I've ordered (EE-SX1025, 0.5mm window) will read that resolution reliably..
On the Mesa 7i96 card, I have the option of single ended or differential mode - initially for testing I'll use single ended mode as is currently working for four 'windows' non-quadrature, but was wondering about the value of trying to implement differential mode and what that circuit/electronics would look like? Probably not the right place to ask, but everyone on here seems much smarter than me...

To convert single ended to differential, you typically use a
differential driver chip like a 26ls31 or equivalent.

You can use lower resolution opto devices by using a Moire mask
(as most high resolution encoders use)

This mask is just a sector of your disk with the same pitch slots
but 2 sections with the slots in the second section having a 1/4
pitch (1/2 slot width) offset from the first section.

These allow large area light source and sensors.

Brilliant! Thank you. The 26LS31 is something I shall have to investigate - a bit above my pay grade, but I'm sure I can learn...

I thought this is what you mean by Moire:

 

with each opto switch reading a different 'track', but now I'm thinking I've got it wrong because I can't see the difference between what I've drawn and just offsetting the two opto switches by half a slot?

I found this industrytoday.com/encoder-performance/ but they are showing an led+mask+photodiodes - the EE-SX1025 I'll be using already has a 0.5mm 'mask'. I suspect much experimentation is needed - I'm only going to 4,000rpm - the only question is, can I get the resolution good enough for rigid tapping...

I think I've got my head around the moire mask now - if I was using a opto receiver & separate light source it would be the right thing but don't think it would make any difference to the EE-SX1025 's 0.5mm receiving slot.