Welcome to Our Community

Some features disabled for guests. Register Today.

new board?

Discussion in 'Controller Boards' started by phill1981, Oct 13, 2017.

  1. phill1981

    phill1981 New
    Builder

    Joined:
    Oct 13, 2017
    Messages:
    5
    Likes Received:
    0
    hey lovely people. just need some advice on a few things
    so I was looking at the A4988 boards and from what I understand they only need 2 pins from an arduino (with shared power) 1 for rotation and 1 for steps..... so hypothetically on a arduino mega that' possibly 27 (54 digitial pins) a4988's can be controlled right? or am I limited as only 15 provide PWM to 7.5?

    also with the a4988 I see there are 3 pins used to decide step rate. so with these can i use standard pins on the arduino to software select? this way if I set up the machine it will move faster in areas where no material is and move slower and more precise when working. if so can someone show a brief layout of the schematic to use? I was thinking of in gcode to add a function for step rate to enable faster rates. I was also going to use an encoder to go absolutely overboard and ensure no missed steps but thats another design and software issue for later.

    I got most of my info from http://howtomechatronics.com/tutorials/arduino/how-to-control-stepper-motor-with-a4988-driver-and-arduino/ so i wanted to use an external source to check items.
     
  2. Joe Santarsiero

    Joe Santarsiero OB addict
    Staff Member Moderator Builder

    Joined:
    Oct 30, 2014
    Messages:
    823
    Likes Received:
    169
    I wouldn't put those drivers on a nema 23. What are you building?
     
  3. phill1981

    phill1981 New
    Builder

    Joined:
    Oct 13, 2017
    Messages:
    5
    Likes Received:
    0
    trying to make a board for CNC that can be used for 3d printer later. If i know the board myself I can modify as needed.
    without too much debate why not the A4988? and what would be better in that price range?
     
  4. phill1981

    phill1981 New
    Builder

    Joined:
    Oct 13, 2017
    Messages:
    5
    Likes Received:
    0
    sorry miscommunication somewhere, NEMA17 is the motor I intend to start with. hoping to get close to 35volts to make it run smoother.
    the driver boards are varied but most appear to have same pinouts is this the case?
    lastly does anyone had a table showing Polulu or the alternatives and what motors are best for each?
    as this board is built and depending on testing I intend to upload files for people to review and try. I am not reinventing a wheel I am just making a wheel easier to handle lol.
     
  5. Joe Santarsiero

    Joe Santarsiero OB addict
    Staff Member Moderator Builder

    Joined:
    Oct 30, 2014
    Messages:
    823
    Likes Received:
    169
    Okay. Just making sure you weren't planning an OX or similar. Carry on. :)
     
  6. phil from seattle

    phil from seattle Journeyman
    Builder

    Joined:
    Mar 17, 2017
    Messages:
    38
    Likes Received:
    10
    The 4988 is an ok driver but it's got limits that aren't readily obvious from the specs. Probably the biggest trap newbies fall into is believing absolute max specs. Sure, 2A and 35V are the limits. But you would quickly discover that running at 2A and 35V is going to cook your chip. Heat is the enemy and running at the limits produces a lot of heat. Even with a heat sink and forced air cooling, 2A is risky. So, it's not about size but rather power. In engineering, you want to avoid running close to absolute max limits. When I design a circuit, I target no more than 60-70% of absolute max for the chips. Also, power dissipation in the chip is proportional the the voltage, when you double the voltage, you double the watts and double the heat. I run my A4988, NEMA17 based 3D printer at 12V and it's fine. Move up to 24V and the drivers get very hot.

    I don't think there is a table like you are asking for. But, it's not necessary. Take the time to understand what the various drivers can do, particularly current and voltage wise, and then you can match them up with different motors based on specs. Read the data sheets for the driver chips - very instructive. The comment about NEMA23 is because those motors tend to have higher current draw than NEMA17 though you can find some 17s that out draw some 23s so it's not absolute. I would try to avoid higher than 1.5A on a 4988, especially at higher voltages.

    So, if you target both a 3D printer and CNC machine, you will be way over specing for the printer. The CNC machines need a lot more power than 3D printers.
     
    Joe Santarsiero likes this.
  7. Joe Santarsiero

    Joe Santarsiero OB addict
    Staff Member Moderator Builder

    Joined:
    Oct 30, 2014
    Messages:
    823
    Likes Received:
    169
    Thank you Phill for taking the time to clarify on this for the other Phil.
    Yes the 17 and 23 motors do have overlap. As known the numbers are only form factor designations. There is much correlation between current requirment and the form factors ability to accommodate. As a result, we see a trend in current capacity requirements based on commonly sold motors. That, as you pointed out, was where I based my comment about 23's.
    After seeing so many noobs building Nema 23 systems for entry milling, the commonality of selling the aio 3 and 4 axis potted packages around the net over the past few years has really not been sitting right with me. OB Included. An xpro, TG, and Gshield are really not well equipped solutions for what most people are hoping for and I think that a number of sites and users(myself included) have taken a veiled approach to handling this problem.
    Common 23 steppers are specd for higher current demands(we know when a stepper needs the most current(minus current limiting :) ) and these aio boards aren't even there even on a good batch. Will it run? sure. Will it be at its potential? Nope. Hence why there are a lot of questions about tuning and the such around the net. I've broken down over the past year and just started recommending the DQ542MA.
    The package doesn't have a high end capacity of current control, but it's seems ideal for where the entry DIY router/mill expectations sit. It's also a smooth quality driver and seems dialed in for the community in regards to what is needed bare minimum. You're design point about having more on hand is analogous to safety factors in building things submitted to stresses and the strength of the materials used. Same root I'm sure.
    A max/peak current on a drive is never a spec to be considered for what the drive is going to be continuously submitted to. It is a worse case scenario item there for unexpected spikes or other things expected during normal drive operation. This is opposed to how I'm seeing most Aio board applications being used today. I believe that this is what the OB Apex is attempting to tackle.
    It should probably get to the market soon. I think the starting trend now is people going straight for the modular drive systems. A haymaker is still a possibility for the Apex.
    Now, and on the opposite hand, in regards to printing systems and drives, do you think going for seperate components for 3d printing(like Phil is asking about here) is really worth the time and effort? To me it seems like aio tech options are already there to offer low cost printer board options opposed to cobling something together with seperate components like this. An xpro might not make a great wood router, but it should be well qualified to push an engraver with beefy 17's.
    What do you think?

    Joe S
     
    Peter Van Der Walt likes this.
  8. phil from seattle

    phil from seattle Journeyman
    Builder

    Joined:
    Mar 17, 2017
    Messages:
    38
    Likes Received:
    10
    That's a lot to digest, Joe. I think we see things fairly similarly.

    A 3D printer makes very low demands on power. Speed is more important than power. That's why you see a lot of belt driven systems. I just don't see a lot of overlap. A4988s and similar chips are just fine for 3D printers. The trend there is moving to 32 bit controllers.

    Yeah, I use separate drivers on my little CNC machine - TH6600 based though the DQ542MA is pretty nice. I agree that all-in-one boards are a problem waiting to happen. Blow a driver and toss the board. With the pololu form factor, you can at least replace the driver. But even for 3D printers it's an issue. Short a driver output under power and, poof, blue smoke escapes. I've seen several manuals that stress that point - clearly they've had issues.

    One other point that I didn't mention before. 3D printers and CNC have a number of requirements that make the controllers different. 3D printing needs several power controllers for heat (bed, extruders), fans. CNC machines have need for spindle and coolant control. Does a single SKU makes sense? I'm not so sure.

    I'm a little skeptical that Apex will happen soon. It's been 5+ months now since they did that video. They keep talking about saying something when they are up to testing but not much has been said yet. I think once we start seeing more info out of the team, it will be a lot easier to understand delivery dates. I like what I heard initially, though.
     
  9. phill1981

    phill1981 New
    Builder

    Joined:
    Oct 13, 2017
    Messages:
    5
    Likes Received:
    0
    so I see on the a4988 for example there is a hybrid option to change step rates, also piggy-backing will compensate for Amperage issues. I think the issue I am finding is there are 3 things needed to ensure are variable speed up to high level, decent torque for cnc and precision
    so precision is stepper motor, torque and speed are dependant on motor and driver... but for entry level I was thinking smashing out a few of these boards and nema 17's and see what restrictions we have.
    so back to my first first post... what would the circuit diagram look like if I wanted to code in an electronic way to change step fractions?
    and the step and direction pins do they need PWM or would a standard I/o do?
     
  10. phil from seattle

    phil from seattle Journeyman
    Builder

    Joined:
    Mar 17, 2017
    Messages:
    38
    Likes Received:
    10
    Standard I/O will work, look at the datasheet to understand M1-M3. Either pull up or pull down (I don't recall). The pololu interface should be a model for you. You need to understand the driver software to know whether to use PWM or not. And, on the microstep issue, it's generally determined during setup and then not touched again.

    To be honest, what you are trying to do has been done. Take a look at the RAMPS 1.4 board - it's open source and has pololu driver sockets. It's kind of the standard for low end 3D printers. There's even a variant of GRBL that supposedly works for it. You should also look at the bigfoot driver interface. It hasn't caught on but could be a reasonable approach - more real-estate for the driver, the pololu FF is pretty restricting. Ideally, stepper outputs shouldn't be routed back through the interconnect - you really want to minimize physical connections, especially for higher current applications.

    I think taking a cue from the Apex ideas would be a very good thing. Modular, supports high current drivers, zero wiring interconnects (this is harder than one would think which is what I suspect is holding them up). It takes a systems approach. I've designed a number of GRBL break out boards but there are a million of those out there. Though, I have a couple of Nano based designs that I like. Still, way late to the game. I'm waiting to see what happens with the ARM based port of GRBL before designing any more. The AVR/Arduino approach is pretty much out of steam.
     
    #10 phil from seattle, Oct 15, 2017 at 12:37 PM
    Last edited: Oct 15, 2017 at 12:43 PM
  11. phill1981

    phill1981 New
    Builder

    Joined:
    Oct 13, 2017
    Messages:
    5
    Likes Received:
    0
    so the reason I am modifying the board is ramps is poorly designed. the i2c function is taken up for no reason, minimum breakout etc etc etc...
    i was redesigning to make a motor shield so that something like a pi takes gcode and then translates to something very easy for up to 7 motors on a mega to do. was also considering making readers like encoders at the end to ensure accuracy (there is no debate on this steppers miss steps you NEED to know if they do, and lead screws act weirdly if poorly maintained... whole heap of reasons to check the actual movement)
    by upgrading the processor to a pi I get more options and faster items. At current boards seem to have an all in one thing happening and I want to get boards to do what they are best at and nothing more.

    what details on the arm grbl do you have?
     
  12. phil from seattle

    phil from seattle Journeyman
    Builder

    Joined:
    Mar 17, 2017
    Messages:
    38
    Likes Received:
    10

Share This Page