Retrofiting an old big milling machine

More
12 Sep 2018 19:47 - 12 Sep 2018 19:47 #117426 by vre
The scale is glass because i saw the glass inside.
I found the VCC & GND from TL084CN 4&11 pins.
TL084CN has populated 1&7&8pin but not 14pin.
Are these 3 outputs the U1 U2 U3 ?
The remaining one wire is the 7.5V ?
That is the usage of 7.5V ?
Attachments:
Last edit: 12 Sep 2018 19:47 by vre.

Please Log in or Create an account to join the conversation.

More
12 Sep 2018 21:12 #117430 by tommylight
The TL084 is an operational amplifier so it needs a positive and a negative power supply voltage and a common ground for both of them, so it needs a + and a - 7.5V measured from it's ground in your case.
That looks to be derived from 15V supply in your case.
It can be done simply by using 2 capacitors to quite complicated voltage regulators.

Please Log in or Create an account to join the conversation.

More
12 Sep 2018 22:14 - 12 Sep 2018 22:17 #117433 by vre
The schematic shows +15V 0V GND +7.5V
If understand correct the +15V is the +VCC the 0V is the -VCC and the 7.5V is the GND(but the TL084CN has no GND only +-VCC) right ?
I have a symmetric linear power supply +-15VDC/1A but i think the is not suitable for this case...
I need a symmetric +-7.5VDC right ?
What max current is good for powering 3 scales ?
Last edit: 12 Sep 2018 22:17 by vre.

Please Log in or Create an account to join the conversation.

More
13 Sep 2018 05:04 #117446 by tommylight
Just a guess but they should not use more than 50mA or 0.05A. They rarely use more.

Please Log in or Create an account to join the conversation.

More
26 Feb 2019 15:19 #127164 by vre
I got new scales that output TTL quadrature waveform so now i have 3 scales 3 electromagnets (each electromagnet engages its own axis X/Y/Z)
and one brush DC motor >2kw that moves all axes(engaging correct gears by 3 electromagnets).
The axis will be moved is the axis which its electromagnet is actuated.
The dc motor driven by a big full bridge that accepts direction&pwm signal.
So how to tune the pid in linuxcnc ?
Is there a guide pid tuning for brush dc motor with full bridge and linear encoder feedback ?
I have a mesa 7i80hd .

Please Log in or Create an account to join the conversation.

More
26 Feb 2019 15:37 #127166 by PCW
Tuning a Hbridge driven motor (voltage mode) is quite similar to tuning a velocity mode servo where velocity feed forward (FF1) is set to compensate for the motors BEMF

The difference is that unlike a velocity mode servo, voltage mode will likely require substantial D term to remain stable with reasonable P values.
The following user(s) said Thank You: vre

Please Log in or Create an account to join the conversation.

More
06 Jan 2021 12:59 - 06 Jan 2021 13:02 #194216 by vre
Happy new year;)
I found time to retrofit the machine...
I have installed a huanyang vfd for spindle and a dc motor driver (pwm/direction)
for the motor that moves x/y/z axis.
All axis x/y/z moved with the same dc motor(100V/26A/also has tacho in the side) and the selection which
axis will move each time is done by 3 electromagnets.
So it is possible only 1 axis movement each time.
Each axis has a linear scale that outputs A/B quadrature signals and endstop switches installed.
Also the are are some other hydraulic electrovalves & oil motor pumps
for lublication for cycling cutting fluid and holding/releasing iso50 toolholders.

I have installed linuxcnc 2.9 with mesa 7i80hd and i have 7i33 7i37 and 7i84 daughter cards also i have some modules like this www.aliexpress.com/item/1005001907108856.html and this www.amazon.com/JBtek-Channel-Module-Ardu.../141-3122699-5842460

What configuration can be done for working one axis a time without 2/3 axis interpolation ?
Last edit: 06 Jan 2021 13:02 by vre.

Please Log in or Create an account to join the conversation.

More
07 Jan 2021 16:48 #194376 by andypugh

What configuration can be done for working one axis a time without 2/3 axis interpolation ?


It isn't easy as things stand, but I think it would be a minor change to LinuxCNC to make it easy,

If an axis is set as a LOCKING_INDEXER then pins appear that cycle state when the machine requests a move of that joint.
linuxcnc.org/docs/2.8/html/config/ini-co...lt_letter_gt_section

As things stand only G0 moves are allowed, and (I think) it can only be applied to rotary axes.
But I think that these are probably artificial constraints.

An alternative, but less satisfactory solution would be a hal component that detects which axes have been commanded to move and at the same time toggles feed-inhibit until the system reports that the clutches are engaged.

It would take a bit of HAL investigation to work out how to detect that a joint is active when feed-inhibit (or maybe adaptive-feed) is set.

Please Log in or Create an account to join the conversation.

More
07 Jan 2021 23:34 #194434 by andypugh
I had a quick play around, and it is surprisingly hard to stop the motion and then be able to work out which axis would like to move.

What looks like it would work would be to enable adaptive feed (M52) and then have a HAL component that adjusts the motion.adaptive-feed button.
It would need to set the feed to a small but non-zero value (I experimented with 1e-10) and then watch the joint velocity pins to see which one is trying to move. At such a feed rate there is no _actual_ movement over a period of several seconds.
So the HAL component would then be able to operate the clutches, check for the clutch-engaged feedback (if it exists) then set the feed-override to 1 and route the PID command for that axis to the motor.
I assume that there is glass-scale feedback or similar?
The clutch-engaged signal should also be used to enable/disable the PID for each axis, so that there is no attempt to move disabled axes. If done that way then G1 X100 Y200 Z40 would actually work, but the moves would happen in sequence.
Do you need help with the HAL component?

Please Log in or Create an account to join the conversation.

More
08 Jan 2021 08:15 - 08 Jan 2021 08:16 #194473 by vre
The handwheels in the front are indepedent from motor
so i can use the dc motor for z axis(permanet engaged electromagnet for z)
and put in the front 2 servomotors for x and y axis but these will have serious backlash.
Has linuxcnc some option for backlash removal in linear movements ?
Linear glass scales that output quadrature incremental exist for x/y/z
Last edit: 08 Jan 2021 08:16 by vre.

Please Log in or Create an account to join the conversation.

Moderators: piasdom
Time to create page: 0.139 seconds
Powered by Kunena Forum