Japanese style spindle orient, how to?

It is just a magnetic hall sensor
As soon as orientation command is enabled , the drive will rotate at coarse speed till it detects magnet signal then it will run at a fine speed to the 0 point and hold the spindle in position
This orientation setup will have Magnet+Pickup+Amplifier configuration that generates two signals .
1)A full sinusoidal signal while a magnet passes the pickup , like a linear hall signal
2)A high signal between 90deg~270deg of sinusoidal signal,  like a TTL Hall signal
These two signals fed to the orientation control PCB.
The one cycle of sinusoidal signal is interpolated to get an accurate angle .
There will be adjustment in PCB or parameter to have adjust the offset if any.

If I remember correctly, the magnetic Amplifier needs +/-15V or +/- 12V supply.
With power connected sensor output signals can be verified on an oscilloscope

But again these sensors are expensive to source in case of failure .
Proximity may not be so accurate to control fine position angle .
A linear hall sensor will  be a better choice .
Motor Resolvers are more reliable and you can retain it with a resolver converter that can be used for both speed and position loop.

Thank you for the explanation. This seems like it would be difficult to recreate as a software component, They are obscenely expensive for what they are.

Seems like using the magnet for an initial sense and an offset of encoder counts is the easier way to go. Thanks again for the explanation .

To add insult to injury, they seem to "wear out" with time and eventually must be replaced. We also have a machine with this set up, and homing the spindle is kind of hit or miss now, and unfortunately it seems to require a complete power cycle of the VFD to clear and reset the error when it errors out, requiring power cycling the whole machine the way this machine (a Lagunmatic) is wired up. This, as well as it's Dynapath control's eccentricities have it on my list of machines I wish I had time to retrofit list.

Hey Andy,

It has all of what you referenced. The machine has a gear box, the orientation card has inputs so it knows what gear it’s in, I can post the manual. I suppose the card could be figured out, but it is designed to work with that and only that Yaskawa drive.

There is a pretty good explanation given on this thread, I don’t know how one would read the analog voltage , if I understand this correctly, input from the sensor amplifier into lcnc.

Todd, I’m just getting used to the Dynapath, agreed it different and you have to push a lot of buttons, it is also strangely simple. Maybe it’s my American mind. I would think that the orientation is handled in the drive also. The Dyna is just sending commands and receiving finish signals. The tool change and orient r on my Tree vmcs work like this, they use a Yaskawa,( seems I can’t get away from these things),positioning drive . The control sends move ,commands, when that tool is at change position on the carousel, the control takes over and move the head, unclamps the drawbar ect.
Anyone interested enuff to see the manual? Happy holidays.

I think I'd be interested in seeing what you have.

I've been told that there is possibly an "adjustment" procedure for the spindle orient sensor that may help, but I don't know it or haven't figured it out, and what I've tried so far hasn't helped. The copies of the Yaskawa manual I have is a bit terse on the subject. And the Lagunmatic manuals we bought from Lagun are mostly useless (the mahcine was bought used and didn't come with original manuals) and Lagun wanted us to pay them more money for more specific procedures on the subject, (so far we haven't caved to their extortion.)

Yes, the Dynapath control has nothing to do with the spindle orientation and simply sends an orient command signal to the drive and waits for the completed signal to be replied. (Not sure exactly how the rigid tapping works, but it is there somehow.)

My biggest frustration with it is drip-feeding/loading code into it. If the control detects something it doesn't like in the g-code formatting it rejects it and won't load it. Usually with no hint as to what is wrong with the code. So I waste half a day scratching my head and tweaking the code to try to get it to load into the machine every time I go to use the thing. (I don't use it enough to have all the bugs worked out of the post I tried to make for it in FreeCAD.)

You probably have an MT3 drive. I have a decent HSM post for a Dyna, works with Solidworks or Fusion. I can share if you want.
Rs232 with Dyna path . Program name goes in parentheses (P1234) no line number on this line. Program load end with “ END “ after M30.
All blocks must have a line number.
No more then 1 g code or m code per block.
Do not put (E) at the beginning of the block after the line number
No spaces at all -G0X1.Y-6. Not G0 X1. Y-6.
No $ end of block on eia code.
Turn echo off on the load settings.
Drip feed is called buffered input and has to be enabled by parameter
Depending on what model you have, some use H and D codes, some don’t. Putting a D on a Dyna 20 will get you an error. No m codes on a line with G80.
IJK are absolute, not incremental if your program is absolute, as in G90. If you program an arc in G91, incremental arc centers apply, like the rest of the world. Absolute arc centers were an American thing. 
Control uses 2-2, 3-3 rs232, not a null modem.

I’d bet the orient parameters on a MK3 are set on the key pad , not with pots. In addition, there is a Dynapath parameter that selects encoder offset counts.
I had a crash course over the last 60 days with the Dyna, it wasn’t fun..

Our machine has a Dyna 40 and I think I had most of that already figured out. It doesn't use H codes for the offset instead it uses a 4 digit T code when loading the tool to set the tool it's offset and automatically applies it. (I don't think I have tried to use a D offset on it yet or set an H offset explicitly, but if I did I failed.)

This has strayed way off topic, I should probably stop hijacking this thread

To try to move this more back on toppic. The alarm our drive is giving us is Fd17. Observing the LED feedback on drive when rotating the spindle at 60rpm, the LED flashes. So the drive is seeing the signal from the sensor, but it looks like the drive is missing maybe 30% of them because it isn't a steady 1 second pulse.

I’ll hijack this one more time. The Dyna 40-50-60 and 2000 can be set by parameter to use H and D codes along with a 2 place tool number. Right now you are using it in Dyna 20 mode. The manuals are downloadable from Dynapath. It’s different, but usable once you get used to it. Most annoying of all is MDI, one code at a time, jumping between menus and options. What a pia...

Now let’s get back to getting this orient thing happening.

Working on re-organizing my GUI "spindle-control" gladeVCP panel and I have a "spindle-orient" button I never bothered to make work. it was in preparation for a toolchanger. Now I'm looking into ways to make this work and came across this thread. Just orienting the spindle to a digital "index" pulse is all I want to do as it sounds that's what the OP wants to do as well...

While no HAL component exists (it's kinda what I'm looking for) you don't actually need position feedback to do this if you can stuff a sensor on the spindle itself. I have been using a proximity switch on my spindle for several years. I assume you at least have speed command control?

What I'm looking to do is....
Push "spindle orient" button, or have the spindle orient command come as it does from toolchange or whatever
command spindle drive to move slow until index is high, move slower until index is low and stop
For fine position it should reverse very very slow until high again and then stop then the component outputs a "spindle-oriented" output
On the next low index pulse the "spindle-oriented" pin should go low to signal that the spindle has been moved

Assuming the spindle has some holding ability that simple component would work well with only a few extra pins and parameters. It's can be as accurate as the device triggering the index which in my case is a grease fitting screwed into a collar around the spindle. It would be relatively easy to scale the speed command for the current gear ratio, I do this now with a belt drive for command and feedback and it works flawlessly for rigid tapping. I can't write the component, I have a friend who used to write all my LinuxCNC craziness but not sure I can get that going....

BTW, I feel your pain on the Yaskawa drives. I assume yours is a VFD? I created a device to replace Yaskawa ServoPacks with internal indexers and translate the serial commands to and from to the DMM DYN4 protocol. I had to (bribe somebody) to then recreate the code to do it for Sigma V's as well. That is for a specific brand of printing press where it's actually necessary, I think orienting a spindle can be done much simpler than even the existing hal component.