Advanced Search

Note our Email is borked for a while

7i84U is probably next week and the 7I96S about 3 weeks out

The blue short Centronics connector on the back of the case/on the board upper left, seems like it is for MPG and knowing how lazy chinese are, i would venture a guess at having step/sir pins for all drives on it, so grab the TB6600 datasheet and a DVM and measure continuance between those pins on the chip and on the connector.
That should save you a lot of time even if you do not have the connector for it as it is easy to solder wires to those connector pins.

Looks great. Took awhile to get that look?
Ya keep us updated.

Hi

I'm new to CNC machines. I want to mill 3D aluminum on a smallish budget. That is the reason why I bough this used machine, my first one. Since I don't want to learn and buy end-of-life software and I like the power and flexibility that comes with open source projects like LinuxCNC the decision to convert it was easy.

It is a Chinese 6040ZH router, 60x40cm bed, 4 axis, rod rails on all axis, 1.5kw spindle, 57BYGH56-4011YD stepper motors (1.26Nm holding torque, 2.8A), 24V system, running mach3 with a USB connection.
Here is a link:  www.lygroupchina.com/gantry-cnc-router/h...c-machine-1500w.html

 

I' ve decided not to invest much into the machine because it has some mechanical design flaws and try to make the best out of it.

1. It has rod rails and not linear guides
2. The leaver arm on Y-axis to the rod rails is bigger than it needs to be. The rods could have been placed higher  

 

Since the controller board runs with a USB-Connection there is no way to run it directly with LinuxCNC. I wanted to know if its possible to reuse the electrical components and it seems so, the stepper motor controllers and the VFD for the spindle. Untested so far.
My plan is to use a MESA i796S board. If you have another suggestion let me know. Unfortunately they are out of stock currently.
  
The stepper motor controllers are 20 pin modules that stick on some kind of motherboard which also hosts the Mach3 controller, namely a Mach3-R-V1.3. With some reverse engineering I'm confident that I could figure out the function of the pins of the stepper motor controllers. The IC of the drivers seems to be a TB6600HQ or a clone. It has a voltage range of 8-42V, supports operating current of 4.5A, about 150w. It says www.shyidiao.com on them.
 

Front side from left to right (with the heat spreader remove from the IC. On the bottom left are two optocouplers):
GND, +5V, Pulse, Direction, Stepper RED, Stepper BLUE, Stepper GREEN, Stepper BLACK, +24V, +24V

 
Back side from left to right (reversed direction from above):
+24V, +24V, Stepper BLACK, Stepper GREEN, Stepper BLUE, Stepper RED, GND, GND, GND, GND

The soldering joints are very likely determining the microstepping of the drivers. If the configuration is from left to right: 1, 1/2, 1/4, 1/8, 1/16 and 1/32 then they are set to a reasonable 1/8.

 
The spindle driver runs independent of the 24V system. It has two ways to receive an input:

1. an analog input with voltage range 0-5V, currently wired to a potentiometer to manually control the spindle
2. three digital inputs, allowing for 8 fixed speeds

The VFD has the marking Inverter_B_v07. Its pins seem to be from top to bottom:
+12V, +5V, i5V (analog Input), SGND (signal ground), FWD (run forward), REV (reverse), BCD2, BCD1, BCD0, ErrB, ErrA

LinuxCNC can handle the TB6600HQ so if its correct they should run with it. Also the spindle can either be run with the analog input or the 3 bit digital one. I still have to decide which one to use. I like the possibility to run the stepper drivers and motors with 36V. That could be a nice upgrade from 24V.
I need a controller board that can handle 4 axis, this VFD, at least one limit switch, touch probe, automatic tool setting and in the future the possibility to attach sensors to close the loop (linear glass scale). The MESA i796S should be able to do that.

A question that is still open to me is how to physically connect the drivers to the MESA board?

1. Buy these 20 Pin sockets and make a board with them? Probably costly and time consuming
2. Solder the connections directly to the pins on the module and attach them somehow. Dirty, easy, quick, downside is, I can't plug them back in the motherboard
3. Solder the connecting wires to the motherboard and plug the drivers into the board. Could be neat because the machine could be back ported to Mach3. Possible unwanted side effects on the board. Of course I would unplug the Mach3 controller chip.

Feel free to correct me or give suggestions etc. I hope the MESA cards are back in stock soon.

Bes

Thanks — that’s a good point. I did see that the 7i96 (non-S) is currently in stock and that it would work for THCAD-2 and encoder input.

That said, for this machine I’m trying to stay fully within the Mesa ecosystem with maximum flexibility and isolation, and to keep things future-proof in a noisy industrial environment. Even if I don’t strictly need 0–10V today, the added analog output and larger FPGA on the 7i96S make it more attractive for my long-term use.

If 7i96S restock is expected in the near term, I’m inclined to wait and standardize on that rather than mixing generations.

If you do not need the 0-10V analog output for VFD, there are in stock Mesa 7i96 without the S, they do have an encoder input:
store.mesanet.com/index.php?route=produc...oduct&product_id=311

No need for a bit file, only if you need non standard firmware.
That is from old Mesa boards that needed the firmware at startup, all new ones have that on board, many of them twice, just in case a bad flash messes up switch to the backup and flash again, correctly if possible.

Thanks Peter — one additional note: I also see that 7i96S and 7I84U are currently showing as out of stock on the Mesa store.

My preference is to stay fully within the Mesa ecosystem (rather than using generic BOBs), so if there are any expected restock timelines for 7i96S and 7I84U, or recommended current-production alternatives that keep a similar architecture, that would be very helpful for planning.

I can wait a bit if needed, but I’d like to avoid purchasing from Europe if these boards are expected to be restocked or revised in the near term.

Thanks Peter — that’s very helpful.

That aligns well with my use case. Since the IL-1000 is only ~10 Hz bandwidth and Y motion is intentionally slow, trading loop bandwidth for higher effective resolution via low-pass filtering sounds ideal for stable stand-off control.

Given that, THCAD-2 looks like the right solution for the current phase, especially with it coming back in stock shortly.

For any future higher-speed sensor upgrades, I’ll revisit true ADC options like the 7i87 if needed — but for now THCAD-2 + filtering seems well matched to the application.

In addition, I will need extra isolated digital I/O for valves, interlocks, and machine signals. I was looking at the 7I84U isolated remote field I/O card (sourcing/sinking outputs), but it currently shows as out of stock. My preference is to stay within the Mesa ecosystem rather than using generic BOBs, to keep things robust and well-supported in an electrically noisy environment.

If you have any guidance on availability of the 7I84U, or alternative Mesa smart-serial I/O options that integrate well with 7i96S for isolated field I/O, I’d really appreciate your recommendation.

Thanks again for the detailed clarification.

The THCAD2 should be back in stock in the US by the end of the week.

Our other analog input card (the 7I87 6 channel +- 10V card) will still be about 4 weeks
but you don't need 6 channels...

For the THCAD2, the update rate is the servo thread rate (giving a loop bandwidth
of about 300 Hz at ~10 bit resolution) You can get better resolution by low pass
filtering (in hal) the frequency signal at the expense of loop bandwidth of course.

 If your height sensor only has 10 Hz bandwidth, you could easily increase the resolution
(to 12 bits or so) by this filtering since it will not affect system loop bandwidth significantly.

Again THANKS a lot
My issue isn’t with Linux itself, but with the documentation.

I guarantee you that I am starting on a very small scale. My gantry weighs 600 kg without the spindle.

I will start tomorrow with mesa ct but just a small question:
When I look through the files immediately after the installation i see no bitfile I can relate to 7i97 card. Is that how it is supposed to be? Does the bitfile exist at all as a separate file and is it generated automatically when the config file is written by mesaCT?

Thanks for the feedback — appreciated.

To clarify, my primary focus right now is using the Keyence IL-1000 for slow stand-off correction. Based on the discussion, I agree that THCAD-2 into a 7i96S encoder input should be sufficient in terms of raw speed and isolation for this use case.

My main concern is less about servo-thread rate and more about effective sensor update rate, frequency-to-velocity measurement noise, and required filtering/deadband for stable Y correction in a noisy industrial environment.

I’ve contacted Mesa support regarding isolated analog input options to complement a 7i96S-based system, but haven’t received a response yet. At the moment, many Mesa ADC-related boards appear to be out of stock in the US, and I’m considering purchasing from Europe. However, I’m a bit concerned about committing to hardware if Mesa is planning to release newer or replacement analog/ADC solutions in the near future.

If you (or others) have practical recommendations for current Mesa ADC hardware that works well with 7i96S, or insight into Mesa’s roadmap for analog input, I’d really appreciate hearing them.

LK-G3000 is only a possible future upgrade — the current design focus is on making IL-1000 work robustly and predictably.

Yeah, that is very old info, for some reason google insists on finding docs and info for version 2.4 but not for 2.8 or 2.9, and that is aggravating, extremely, but is beyond our reach to fix it.
We do remove outdated info from the forum whenever we bump into it.
-
Ping OK is the only thing you need, make a config and start wiring stuff.
-
You might need mesaflash ONLY if you need firmware for:
-you want to use a Mesa 7i74 or 7i85 or add a 7i76 or similar
-you request special firmware that is not standard
-
One major tip for wiring, do NOT wire everything and power the machine ON.
Follow the link i posted, wire stuff one by one as mentioned there, check and recheck, and only when making sure all is good, move on to wiring and testing the next thing on that list.

I would just try either MesaCT  or the sample configuration.

As others have mentioned, you don't need mesaflash, unless you need
a configuration with a parallel expansion port connected daughter card,
or perhaps want MODBUS support.

Unfortunately the mesaflash version included with 2.9.4 is quite ancient
and does not recognize the 7I97T.

As long as:

mesaflash --device ethernet --addr 10.10.10.10

replies with the card name (whether it recognizes it or not) you have a connection.

The ping  test also works, except if you set the host IP address to the FPGA cards address
(a common mistake) but if mesaflash finds a card, unsupported or not, you have connectivity.

Thanks. That was the issue. I did the script SrL had above, and lost the customized grub notes. thanks