DTHC2 settings

[kn612]

Can anyone share their optimum settings for their old mach3 dthc2 settings? Z motor speed, acceleration, THC rate, etc.

I've always had pretty good performance with this old blade runner/dthc2. However the thc wasn't seeming to respond very well on some 11 gauge. Didn't have time to play with the settings so I finished the job with a different table that uses a minithc.

Left is dthc2, right was miniTHC. You can see on the perimeter where the torch rose and wouldn't come down. I'm using a powermax65, 45amp consumables, clean dry air. I run hypertherm's book specs for speeds/voltage.

18051.jpeg

Mechanically z axis is tight, coupler isn't slipping.

Tested it on some 16 gauge and had similar results.

IMG_20231221_143718.jpg

Cut some 1/4" at some good angles and the dthc2 will follow it but I bumped up the speeds/accelerations from what I usually run. Don't want to get a sawing motion on thin stuff.

IMG_20231221_143833.jpg

IMG_20231221_143758.jpg

IMG_20231221_143751.jpg

IMG_20231221_143744.jpg

Curious what guys are running their machines at.

[tcaudle]

Most are using 15 year newer controls and software with the ultra high speed DTHC5 and iis fast enough to cut corrugated at full XY speeds..
The DTHCII is limited to MACH's round trip loop speed and response. While you can play with the Z speeds and the DTHC settings for the Z speed but you cannot change the acceleration on the DTHCII because MACH has no setting for that. Practically you are limited to about 20 to 30 IPM Z speed

The DTHC 5 running under Limix and the Real time OS can do up to 350 IPM Z speeds and has things like Velocity anti Dive (VAD) on corners and detail cuts, has variable accuracy (tracking) settings and a host of other options .

Take advance f our Support options and Join the Support Forum , or fillout a request for Support off the website. Wile you are there have a look at some of the new controllers and the options to upgrade what you have.

[kn612]

Is it possible to switch out the parallel port ubob3 card and the dthc card for mesa hardware? While retaining the use of the front panel and reusing the g251 card?

It all seems fairly simple besides wiring/configuring the charge pump/control to enable the gecko drivers. mesa 7i96s, 7i84, 2 thcads fit fairly well with enough room in the bottom for power supplies.

IMG_20231221_200632.jpg

Screenshot_20231222-185504~3.png

[robertspark]

that would be a very easy a good upgrade to qtplasmac on linuxcnc , given you have all the hardware to hand (same as newer candcnc boxes.... just newer versions of linuxcnc, I think that candcnc is running a much older and no longer supported version of linuxcnc, I think it is LinuCNC 2.6, the current stable release is 2.9.2, there are some improvements)

https://linuxcnc.org/docs/stable/html/p ... asmac.html

https://forum.linuxcnc.org/plasmac

https://linuxcnc.org/docs/stable/html/p ... rimer.html

__________________________________

https://www.candcnc.com/downloads/

As of September 2019 we started shipping systems with an updated Linux OS we call CNC-Xf. It is derived from Xubuntu 18.04, and it is now a 64-bit OS. Practically this means that if your system was shipped on or after 9/18/2019 you probably need the 64-bit update

https://xubuntu.org/release/18-04/

Xubuntu 18.04
Xubuntu 18.04 is an LTS release which was released in April 2018.
Codename Bionic Beaver
Release Date April 26, 2018
End of Life April 29, 2021

[tcaudle]

The upgrade we offer swaps pout the whole front end (UBOB/DTHC Tbale I/O) to allow for expanded I/O and the real time kernel.

Here i the problem with doing it piecemeal:

The version of Linuixcnc is older and modified to use the DTHC Hardware. There is a plugin that handles the RS485 communications and the DTHC interface. Its not going to work with the Palsamac release .

We have a current project that upgrades the linux to 22-04 and the latest Linuxcnc. We have to put in all of the support for the DTHC hardware and other devices along with a updating screens to GTK3+ and redoing the Configuration tool .

The forst versions used the UBOB and the 7i92 MESA card but still needs custom compile and added support for DTHC hardware and control.

If you want to go that direction you would be better off to use the MESA BoB and abandon the UBOB and the drivers. Its an integrated system and made to work as a unit .

If you go the PlasmaAC route you may find the PC you are using is not fast enough to support everythign running in software. Beacuse the DTHC has its own 49MHZ processor/memory and runs the Z axis closed loop outside Linuxcnc its a lot more efficient

[tcaudle]

The release date on the 22-04 update is this month.

[kn612]

I have been playing around with qtplasmac and 2.9. I really like the qtplasmac interface. I have all the hardware, installed in a custom metal enclosure, but would rather use the candcnc blade runner enclosure. It's much neater and I like the front panel with the on/off buttons. Along the the good 48v power supply and gecko drivers.

I removed the ubob/dthc/table io and am able to power up the front panel and it sends 48v to the gecko drivers. Fan turns on, green light on front panel. I'm just unsure how to configure the wiring from the 7i96s to the gecko board to enable the driver's. I can jog the table around if I remove the gecko board and wire straight to the gecko driver's but that defeats the whole purpose of what I want to do.

I don't know what original pins from the ubob enables the drivers. I assume it is estop but there are other pins (shift, output 18, output 19, charge pump). I'm not very good at explaining my thought but it seems a fairly easy swap. Much easier than building a controller from scratch.

What are your thoughts? Is your gecko board too tightly integrated that the drivers can't be enabled from any signals from mesa board? Worst case I figured I could remove pin 11 of the gecko 251 from the board but I didn't want to start hacking working hardware.

[tcaudle]

Its like taking parts out of your garage and the local scrap yard and building a street legal vehicle.....Its going to be a lot of trial and error. I guess it depends on what value you put on your time .

The older UBOBIII uses a unique input MUX that doubles the I/O on a single parallel port. There is a custom MACH3 plugin to handle that. It also has an output MUX the increases the outputs but its optional and not needed. You simply don't have enough I/O using the older UBOB. The 26 pin ribbon cable to to the G251-4 assembly carries the standard step and direction for up to 5 axis but also logic power and other signals. The steps and dir are on pins 1 - 9 of the cable and pin 10 is port ground. All of the driver protection circuits are on the G251-4 but its design the feedback to a "Smart" panel that handles faults. Later designs have an RS485 port on the front panel that is tied back to the User Interface The system is designed as a functional controller with pluggable modules but designed to work as a unit. The 7i92 to UBOB upgrades could only use a late model (REV16) UBOB tha had expanded inputs . The current EBOB is a direct Ethernet high speed I/O with 16 inputs, 6 axis of signals and up to 9 outputs out of one card. Has the same 26 pin cable to the G251-4 or to our MAT250 card that lets it plug into the newer Dual Gecko driver modules .

[kn612]

I don't mind tinkering here and there, especially to keep a cnc machine running. Does the MUX handle enabling the drivers? I would think estop would be what enables and disables the drivers?

[tcaudle]

No, The drivers are not enabled or disabled . The motion signals (step and direction) are not part of the mux either . All outputs (even the muxed ones have a safety disable (CP) that keeps them off if the software is in RESET or not running. SO in a sense the dives are disabled BUT its not from having an enable or disable signal, its from having no motion sigas . Now, the power section most of our controllers (which is independent of the software and motion control has driver disable that can be tripped by fault conditions in the power section .

[kn612]

Ok, thanks for the information.

When I removed the ubob and swapped in a 7i96s I wired the signal wires to the the g251-4 but had no motion. So I pulled the 4 gecko drives off the candcnc board and wired signals and power straight to one gecko driver and then had motion.

Pin 11 is disable on the gecko drives when grounded so I assume charge pump/mach software reset button has something to do with pin 11. I wasn't sure what kind of voltage goes through pin 11 so I didn't go any further for fear of ruining good electronics.

I can still use this controller to route with at the moment.

[tcaudle]

Drive disable is based on two things: The driver interface has a current overload detection that pulls the disable pin high on THAT driver to rapdily disable it (but only as long as the overload exists which if the driver turns off , is not long but it sends a pulse to the "Smart" panel and trip[s the Driver Fault/ It also ,motors the DC motor voltage when you start and delays turn on until voltage hits about 3/4 of the DC valuer. Its part of the "Soft Start " so the front end is not faced with starting up the bulk DC circuit AND turning on all of the motors at once/ Motors are"on" even when sitting stilloto voltage. NONE of that has to do with what kind of cutting (plasma VS routing) . The [power section and drives have no concept of the kind of cutting being done. its just Volts and Amps . If are getting faults in one mode versus another its something different in the SOFTWARE settings (Profile)

[kn612]

Sounds good Tom. Have never gotten a drive fault. With the ubob, dthc2, and I/o card removed I can power up the box and green light on the front panel turns on and there is 48 volts going to the 251-4. I went as far as to wire step/dir from a 7i96s to the 251-4 but couldn't get movement. When I removed the gecko driver's themself from the 251-4 then went straight from the 7i96s to a gecko I could get movement. I'm just not sure how to interface the 251-4 to the 7i96s as far as charge pump or estop goes. I didn't want to input voltage to the 251-4 on a wrong pin and ruin it and the geckos. I'll make some time to pin it out better.

I ended up building a controller from scratch for this table using linuxcnc that works well but I would still like to pursue this project. There is a lot of good hardware still in that blade runner box.

[tcaudle]

YOu are trying to stink things together that we not designed to work that way.

When we do an upgrade the whole front end including the DTHC is replaced . The g251-4 does a lot more than send step an dur tot he drivers .

It was never designed to be used in pieces. It all has to work together .