Please don’t take this as slander. there are too many issues that need resolved with this machine to be sold as anything more than a prototype. they also market as it does all these things making people like myself think it does way more than it does do. and then doing a quick search and watching their very short vague videos you feel that it does do these things. its not until after you own the machine and you do the hours of research it takes to know what this machine really is. though it may be my fault for not reading and doing research on a product before spending $3000 on a machine that i cannot get to work properly. due to stock hardware issues that if i didn’t know something about already i would have never figured out. then getting into their troubleshooting guide is lackluster to say the least. they say yea just try something similar to this and it should work out for you without properly explaining what they are talking about.and after all that you learn that the software is too convoluted and needs to be streamlined and changed almost completely. but at this point i feel i am rambling. i do think that this has major potential if they get the right people working on this to turn it in the right direction. here is to hoping my issues will be resolved.
I was able to take a look at the video and your post on the Forum. I would like to help you point by point:
1.Adjust the tension on your belts. If 0% tension is completely loose and 100% tension is completely tight, you should adjust your belts so that they are at approximately 60% to 75% tension. (We will add this to the troubleshooting guide)
2.Only move one axis at a time. If you are having issues with motor stalls please change your sequences so that only one axis (X-Axis, Y-Axis, and Z-Axis) moves at a time. When you move more than one axis at the same time time, it can create vibration feedback and this is picked up by the encoders as resistance and it can cause a stall on your device. (We will add this to the troubleshooting guide)
There is a small bump at the midpoint transition between the two track sections. This bump needs to be sanded away with some sand paper. This is why your bot stalled at the midpoint. The transition between tracks need to be smooth to the touch on the top and on the bottom of the tracks. (This item is already in the instructions)
Adjust your Max speeds so that they are lower than the current values. When you adjust your speeds lower, you are lowering the probability that the motors will stall out. When you adjust your speeds lower, you will need to increase your timeout parameter accordingly.
In the Genesis v1.4 model the stepper drivers are a source of reliability issues for us. In the Genesis v1.5 (The next model) we have updated the stepper drivers so that they are significantly more reliable. I realize that doesn’t fix your issue but this item has been addressed on the v1.5 systems.
You may also consider replacing all of your A4988 drivers so that they are all fresh new drivers.
We are still working on streamlining our software and the final version of the Weeding sequence is still under development. We are working hard to complete the weeding functions of our software.
We have found a product called Jig-a-Loo Graphite Lubricant and we suggest that you use this to significantly reduce friction on the FarmBot Z-Axis lead screw. You should apply enough graphite lubricant so that the hue of the screw should look significantly darker than the original silver metallic color. You should use enough graphite lubricant so that the screw should appear black or dark grey. This will allow the screw to operate without friction and vibrations.
Here is a YouTube video that shows how the product is used and the typical applications.
Please let me know if I can be of further assistance with respect to the FarmBot.
No offense, and please take the following comments as ways to improve and not pure critics, but seing this reply I can’t help but feeling that the mechanical conception of this device is problematic to say the least.
1-It should accomodate with non perfectly linear axis. You can’t realistically ask people to have a perfectly flat rail on their gardens, there will be dirt, there will be leaves, even maybe snails in the way. It’s a garden. The gantry should be able to accomodate for that. Why don’t you use bigger wheels, like roller blade ones? Would solve the issue for cheap. That would also solve the top speed issue, since steppers loose torque at high speed and stall, which they wont if they were to drive larger wheels (through a belt and pulley reduction of course).
2-Grease shouldn’t be necessary on the lead screw. It’s of course better to use some, at least for corrosion resistance, but it should still move freely without it. I have lots of CNC and 3D printers using lead screws, including some tall ones. None of them are lubricated, all of them work fine. If it binds, then the problem is elsewhere, in the conception.
3-A4988 are bad drivers. Almost any other driver will be better than these.
4-There is no excusable reason for not being able to move several axis at the same time. Any CNC machine is able to do this. I can’t say for sure, but that sounds like a firmware issue. Wonder why you don’t use Marlin as a firmware, it’s been there for years and works very reliably. Plus I’m sure the Marlin guys would be happy to help you guys make it work, they are very open when it comes to help adapting their firmware to new uses.
Interesting idea I have also thought of a guide, like a wind screen wiper at the front and back of the gantry plate like the fender on a train, to push leaves and dirt off the tracks.
I used my Farmbot for years without grease on the lead screw, I now put oil on it as it does run smoother, probably needs a clean. I own a few different 3D printers and I regularly oil and grease them as the manual says you should. On one it really helps to keep it maintained.
What is a better alternative and why? Interested to know.
That is strange, I have always been able to move all three axis at the same time with the firmware. So I agree it should work. I have had a few vibration issues though as @Marc says but the firmware allows you to attempt a move 3 times which usually sorts those out. It doesn’t have to be as smooth as a CNC cutter or 3D printer to get the job done (although that would be nice).
I have a Genesis v1.2+
I am interested in better motor drivers. I would love to see someone post a paper about how to swap out the motor drivers in a RAMPS card and set them up to work well with the FarmBot. @Marc is there some user made documentation on this?
I have observed that the X-axis is more prone to stalling if the machine head is near Ymax or Ymin. If I am traversing a long distance in X, I move Y to the midpoint. When there is a large offset in the mass moment of inertia about the Z axis that seems to make the x drives struggle. Just an unproven observation.
For lubricating Z-axis, I have used this https://www.3inone.com/products/silicone/ It is fast drying and leaves a slick film. (I never really had a problem with it in the first place)
What do you expect from a better driver? What about the trinamic drop in driver 2208? Stepper driver change for no-noise bot
Those are a bit more expensive, but they are really worth it as they reduce the noise drastically and the movements feel much smoother.
In my opinion based on experience, a4988 drivers are really the worst ones you can get these days and I probably won’t ever order those again.
I would say that a much better alternative here would be the DRV8825. Not necessarily the best motor driver, but a good compromize, since they are cheap, reliable, and they offer decent current capabilities so your motor gets more power and has less chances to stall.
You can go for better drivers like trinamics and such, but that’s, in my opinion, overkill for such an application since their main advantage is quietness. A Farmbot is supposed to be outside, where noise doesn’t matter all that much. Given the price difference I’d go for the DRV8825 ones and keep expensive drivers for 3D printers.
@Jamesh This system is clearly not turn-key, which is what I think you were saying. For some people that is exactly what we are expecting, a system that has baseline functionality but is accessible enough to improve. I don’t feel there was an intentional misrepresentation but perhaps a misunderstanding of how people outside the robotics industry might see its state. Could you imagine dropping $30000 for the latest “low-cost” production robot only to find similar issues as the FarmBot? Sad to say that I have experienced that with several of them and many had a much higher price tag with a much smaller feature set.
I always recommend grease on a lead screw as they are typically made tight to reduce backlash. That said I think they should be shrouded as well with accordion style conduit to keep the grit out. At a minimum there should be wipers top and bottom.
My weak understanding of EE and these components is that stepper drivers are not meant to be used over such a long cable. I would expect this to exceed the current control capabilities of most drivers. The architectures that I am familiar with use a power bus and communications bus to a driver collocated with the motor.