Today I did a quick stress test of the machine for run time. And accuracy. For carving a board game. Making Settlers of Catan. I’ve been running for 6+ hours now. And all is working great. Spindle is running around 94 degrees and holding steady there. The spindle has been running at 24k RPM’s the whole time. All’s working great! And. As for noise. 82.4 DB while running with the vacuum. So relatively quiet considering what’s running. (Daughter is an audiologist, so that’s an important thing to her!) I’m cutting with an 1/8” ball nose bit, so lots of cutting to get the carving done.
Just moved the vacuum out of the work area. To the garage stairs, and the sound level dropped to about 76db. Can still hear the vacuum but much quieter. Sitting in the next room, can barely hear the machine. I’m going to really like this as I learn. And the learning is just starting.
Ok.. So I am at a loss today.. I shut the system down completely yesterday after finishing up some minor cuts. Today, I turned everything back on, and the Blackbox is not booting correctly. I normally see the USB light on the controller flashing away, but not today. I have double checked all the wiring, and dont see anything out of place. This is what I am getting on the system when I try to connect.. One thing that did happen was the Openbuilds Controller updated to the latest version today.. So I wonder if that is the cause.. I read a thread that had this exact same error, and they thought it was a EPROM corruption in the data for GRBL, and to use the $RST=* to reset the settings to default, but HOW can I do this, if the system will not boot and allow me to connect? Seems that I will be doing some digging tonight, as I wanted to do some cutting tonight..
While waiting on a replacment BB Brains board, I decided to build an Arduino setup. I will build this as a backup to the BB system, and use external TB6600 Steppers Drivers. I will build this as a direct Plug and Play replacment of the BB, so I will be adding the 3.81mm Connectors to directly plug the wires from everything into the new control box. I will also be trying to increase the DC from 24V to 36V and see if the steppers become more reliable for higher speed moves. Might as well play, while I cant do anything else.. If any of you are from the West coast, your will understand this next comment.. A trip to Frys is in my near future! (Luckily, there is one here in Illinois about 15 mins from my house. Unluckily, its only 15 mins from my house.)
Machine results look good. If you're gonna go 36V, wouldn't it make sense to just use proper DM542 drivers, build it right instead of skimping on cheap variable-quality drivers (that often don't play well with NEMA 23s even at moderate current ratings, as we've seen here in the past), and then use the lower-voltage BlackBox (you can't put that 36V into it) as the lower-performance-but-faster-assembly backup? Sure, it's $180 instead of $40, but while it probably won't be as drastic as the move from 12V to 24V on NEMA17s, it should be a pretty noticeable improvement in all-around performance.
I will look at those. I was say the the tb6600’s because. I have them here already. Any recommendations on what other parts I should do? As you can see, I’ll buy the better parts to get better results. But being a newbie. Im still learning. At the end of the day, I want to be able to use this for limited production parts for my company, thin aluminum parts for computer cabinet assembly. Thanks!
I personally wouldn't use "TB6600" drivers (whether they actually contain Toshiba TB6600 ICs is usually debatable) for anything other than 24V, sub-1.5A, NEMA 17 projects, where cost is more of a factor than anything else- and even that's debatable at this point, I'd strongly consider Trinamic TMC2xxx boards- higher performance at a similar price point. For NEMA 23 projects, well, a) I now exclusively use closed-loop steppers for NEMA23/24/34, because the price really isn't much different to open-loop and the performance/safety is better, and b) if I were to go open-loop on a machine that I want to use for reasonable-performance commercial production work, I'd pay the $45-60 per driver that high quality DM/DQ Leadshine-type 32-bit drivers cost and use as high a voltage as possible. Not sure what other parts you might need at the minute without seeing more pictures of what's going on, but for now I'd probably focus more on actually using it and gaining experience than tweaking it until the end of time (ask me how I know )
Umm.. I dont know what your talking about.. This project started as a straight LEAD 1510, converted to a 1515, changed to include Linear Rails, upgraded the spindle, ..... Sorry.. I don't know what you mean by your comments.. I have been looking at closed loop motors, but that comes later. I just want to make sure that if there is a problem, like I have now with the Blackbox, it doesn't keep me from working on the machine.. Have a redundant system of some type, for now.
Hahaha, right?! Closed-loop steppers work exactly the same, system-wise, as normal steppers. They close the loop at the driver, not the control (unless you want to get into LinuxCNC and even then I don't bother). Not a major upheaval to install them. Yeah I'm not a fan of all-in-one systems for this very reason. Redundancy with a separate driver system is a) a spare driver, b) a spare Arduino. They can sit wrapped on the shelf until the end of time, they'll be fine. Make sure you use a terminal shield with the Arduino, and then any failure is just an unplug-replug operation and you're back up and running within 5 minutes- either reflash and reconfig grbl, or set the microstep and current switches on the driver. Then you just buy another of whatever just died and put it on the spare parts shelf, no big deal.
New brain board installed. Now I’m not getting the steppers to lock in place when not in motion. Time to dig some more but seems the enable is not being set. on the old board, when I powered on, you could hear the steppers lock in place. I’ll look in the morning.
We need more pictures.. and videos.. . btw, thanks for the message, I have not really been online, and my upgrade is not going to happen anytime soon till I gain employment (pandemic layoff )
Problem solved. Looks like a GRBL setting got changed when I did the reconfiguration of the new board. Peter Van Der Walt to the rescue! I will post some pictures and videos when I run the next job. Probably tomorrow. Need to make a game board for a Christmas present. I think it will take like 7.5-8 hrs of machine time to make it with all the 3D carving.
Just got home. Setup machine. Home it, mounted the board to it and sent the drill code to it to create the screw down holes, and now my y/a axis is missing steps... time to investigate this a bit more. Something had to change in the GRBL setting between the old to the new and I’ll figure it out in the morning. —Edit— Started playing with GRBL settings, and I think I see a difference. Seems the default acceleration was increased to 350mm/sec from something slower. I think it was 250 in the old OB Controller settings, but I didn’t backup my setting before it died so my bad. Slowed the acceleration down to 200mm/sec and all seems to be better for now, but I’m going to have to play a bit before I trust starting a 8 hr job and going to bed. So tomorrow it is.
So, Sanity check. I ordered today: 1 Meanwell RSP-500-48 Power Supply (Y/A Axis Power Supply.) 1 Meanwell LRS-350-48 Power Supply (X/Z Axis Power Supply) 1 9V DC Power Supply fot the Arduino 1 Eaton 30A 220V Contactor W 110V coil (Ill use one leg of the 220 to power the coil and the LED of the switch) 1 Eaton 30A 220V Breaker 1 Power Switch with LED Back Light. 5 DM542T Stepping Drivers (4 for use, 1 for spare) Misc Fork Spade connectors and Ferrule connectors for wires Various 16/18AWG Wires for connections Terminal Blocks for Wire Connections PWM to 0-10V Converter. (For Spindle Speed Control) GX16-4 Wire Connectors GX16-3 Wire Connectors GX16-2 Wire Connectors DIN Rail to mount everything I have both a Arduino Uno and Mega with Screw Terminal breakout modules and enclosures.. If I can get the Mega working with 4/5/6 axis, I want to add a rotating axis for Lathe spindle. I think this is everything I need for a complete controller setup. I think this takes care of everything... Am I missing anything?
Someone has a secret credit card his wife does not know about.... Why not use a Mesa card? What are you using for a spindle?
I have a 2.2kw water cooled spindle. This will be good enough for now. And when I get to more involved projects for work, maybe I’ll get a different one. But for now this will work just fine. as for a Mesa card. I’m trying to make it completely self enclosed, and will have a micro PC in the control box as well. I have a very special micro that doesn’t have a home, and will work great for this.
They have expansion slots. They are PCIe. But there is no way to get a real card in it other then the manufacturers cards. This will take a card about half as long and about 30% as high as a normal card. It has usb-c on it, so I could get an external PCIe enclosure and do it that way. But then it defeats the micro PC. I have the arduino’s already, so I’ll start with that and if later want to move to something else, I already have the rest, so not a big deal. but. Looking. They have an Ethernet version. That would work. So now. Do I do this or just wait. I’ll wait.
Oh, gotcha, yeah, a 6i25 would struggle to get in there, even as a low-profile card. But yeah, there's a 7i76E that's Ethernet. Not quite as easy to deal with as the pure PCIe/parallel systems, from what I've read, but doesn't seem super difficult to get on with. I almost went with it for M4, but I'm gonna stick with 6i25/7i76 and I think add a 7i73 control interface card for some convenient IO nearer the operator's position Other than the ease of destroying them with ESD, I love Mesa cards. Super high performance, great support on the LinuxCNC forums, and they just work. I'd still wait on it though in your position- once you're ready to jump up to 4-axis, that'd be the time to switch over.
I have not heard that the Mesa cards had an ESD problem.. but at the currents and voltages in use in a CNC machine, it is not surprising. The nice thing about the Mesa cards, is that you won't be limited with future expansion. Get the basic card, like the ethernet one you mentioned, and you are off to the races, want to add more, get daughter boards, need more inputs or outputs, get another daughter board. Your software won't change, the configurations will. With some of the smaller all in one solutions (which I would include any RPI or Arduino with appropriate shields or hats), when it comes time to expand, you essentially have to replace it all (well the compute components). I have a bunch of RPI and arduino boards, but I am not using them for CNC. I did have cncjs running on the RPI with a POE hat for a while, but when my storage array took a dump due to a failed firmware upgrade.. I lost it (I boot my RPI's via iSCSI) I eventually plan on switching to LinuxCNC when I can finally do the major upgrade (right now I am doing minor ones, like the spindle I replaced with the digital interfaced AMB 1400W 3500 to 29000k spindle) I was also looking into Ethercat, but that is pricey, and I am finding it hard to design the system, as US suppliers are slim.
I killed a card once by touching it (took a couple days to actually diagnose it too, I had no idea it happened), and another one once by moving the machine a foot without casters. They can be pretty sensitive to things hobbyists wouldn't typically think much about, and at $100 a time, it adds up. I've never had them have any problems with normal operational environments though, they're pretty resilient electrically. But yeah, just buying the 7i76E gives you all the RS-422 serial IO of the 7i76 PLUS the two 25-pin parallel headers of the 6i25, you get virtually unlimited expandability for moving into toolchanging or other automation like palletization, complex control panels, etc. LinuxCNC, especially with Mesa cards where you can use HostMot2, is... Suprisingly easy? I haven't really found a point (other than lack of information and poor documentation, which is definitely frustrating and causes delays, most people just seem to go straight to the LinuxCNC forums and ask for the answer, which I don't like) where I'm particularly taxed by it more than at some points with grbl so far. I'm sure it could be made more complicated by expansive HAL work or esoteric display environments, but with normal HAL work (glass scales for screw mapping, jog pendant) and a basic AXIS environment, it's been fairly plug and play so far. I briefly looked into EtherCat, but yeah, seems expensive and rare to be jumping on at this point.
What spindle did you end up going with? I just bought the stock 1515 kit. Was thinking that the 1.5kw spindle with a 65mm body would be more appropriate for the rigidity of the machine. Are you controlling the spindle through GRBL? I've tried searching on this forum for some info on spindle recommendations and setup but haven't had a lot of luck. I build my own CNC (Kronos Robotics KRMx02) quite a few years about but I am at a caveman level compared to what you guys are doing and talking about.
I am using a kit 2.2kw spindle. It is a Chinese spindle, but it works. And yes, it is controlled through GRBL. It’s water cooled with air assist. And it’s all controlled through GRBL using IoT outlets for the water pump, compressor, and the dust collector. I’m using the coolant relay to turn multiple IoT outlets on and off being fed from different circuits. So, start of all g-code turns on the coolant, then the spindle to the speed needed, waits 8 seconds, and proceeds. One thing I will be adding shortly is a flow monitor on the cool and and hook it to the door sensor to pause everything and turn off the spindle in case of coolant flow problems. Thats the next thing to add, after I build the new controller this week. Everything should be here by Thursday for that.
I opted for a 1.5kw air cooled spindle. I also got it from China. Check on AliExpress for the spindle size that you want. This is the kit that I bought.
@Netechsys Do you feel that the 2.2kw spindle is heavy for the machine? I am leaning towards the 1.5kw but would like to have the capability to run larger bits and there really isn't any cost difference. Of course paying attention to feeds and speeds. Do you have any pictures/documentation on how you setup your spindle to run everything the way you did? How did you mount the spindle? @JustinTime Thank you for the quick response. That looks like a great option. Did you make an interface plate to mount the spindle?