Forum Replies Created
-
AuthorPosts
-
cmcgrath5035Moderator
You might find on the wiki some references to networking two tinyGv8s together, it is not clear how well it works.
The likely best answer for you is to study up on tinyG2, as it is rapidly improving and in-use by many. It is still Edgy but if you need >3 independent axis control, likely your best bet.
tinyGv8 has only enough compute horsepower to deal with 4 axis channels.
Networked tinyGs was intended to share a >3 axis Gcode stream across two compute engines, Folks making custom machines (and compiles) made it work, but G2 is likely the easier to go forward path and lower cost to boot, if that matters.cmcgrath5035ModeratorDedicated G2 sub-forum? Maybe in future, for now use this.
You may find this helpful (by downloading it you acknowledge it’s DRAFT status). I’ll clean it up after review by the Synthetos folks and make available in editable (svg) form, so folks who custom compile pinouts can have their own. I got dizzy last week trying to explain this, so made this diagram from the familiar DUE base
Without doing a lot of looking, I’ll make two reasonable assumptions about the Due ports: 1. they probably have ESD protection structures on chip and 2. whatever looks like a pull-up is likely not a high wattage structure.
My suggestion would be to implement limit switch interface similar to what is used on tinygv8. Schematics can be found here (hint: 2.7K and 0.22uF)Also make sure you see that build G2 75.02 now has individual switch settings, min/max *6 axes that need to be configured correctly.
If issues continue with homing, perhaps a post at
Be sure to describe how you have set up the switches and how you builtG2
And remember, ground shield at only one point, probably DUE end for all shields
- This reply was modified 9 years, 11 months ago by cmcgrath5035.
- This reply was modified 9 years, 11 months ago by cmcgrath5035.
cmcgrath5035ModeratorCool!, Nice gallery of shots.
Looks a whole lot better than my version, made on a drill press with a 3″ holesaw and 1/4″ Starboard. Printing for sure on my future agenda.cmcgrath5035ModeratorI sort of thought so, comparing forums.
I have not had the time to dig into the details as you have, but have seen folks have to (try to) replace QFPs in tinyGs due to an oops with wiring, etc. So my current quick solution is to raise the flag high and let them seek their own understanding.
Given the ‘almost same’ UNO vs DUE interfaces, I bet a whole lot of DUEs are being compromised by 5V shields of all types.
Gaups looks cool – now on my list of things for folks to consider.
cmcgrath5035ModeratorI’ll hazard a guess that the Gaups folks will realize quickly that they are uniquely positioned to serve the 4 motor DUE users and make the 5V to 3.3V transition easier in the near future.
Enjoy your machine and please feedback bugs and issues.
Nothing improves a product faster than good user experience.cmcgrath5035ModeratorWhy ask? It is something that might be of interest if the results look ‘good’, whatever that means.
But you already answered my next question, “what does your machine look like?”.So I won’t bother to ask what it will look like, I’ll wait for you to post after you are done and run a job or two and declare success.
cmcgrath5035ModeratorThere is never a ‘clean’ answer to this question.
In tinyG space, lets call it G1, ‘Master’ has become very out of date and I would recommend ‘Edge’. Having said that, I do know that a new ‘Master’ will be posted soon after some cleanups to the current ‘Edge’.In G2 space, the current Master is a relatively new build, 72.01 I believe. Currently Edge is building 75.02.
G2 is still under heavy development. Master and Edge are changing more frequently. Master will always be a bit more mature than Edge.
All that said, as best I can see more folks are using the Edge of the day, perhaps because they are in experimental mode as well designing new machines around the capabilities of the DUE and full 6 axis control.I guess the good news is that the BOSSA loader works well, so switching is not a huge hassle once you have your build (compile) environment tweaked for your system.
Do be aware that custom settings files may not port directly to new builds, Edge added some additional parameters recently that affected my builds.Best solution: Unless you want the most stable environment, I would build a Master that you test and verify it works, then put it in a safe place should you need it. Then build an Edge, if it seems to work for you stick with it, you can ride the crest of the wave.
Also, monitor the chatter here:You will learn a lot form other’s experiences and responses from the developers
cmcgrath5035ModeratorBy the way, look up at my item 5 again
It may be moot now that you have figured out all the re-compiles.That block on the schematic would allow you to drive two 8825 drivers from the same logic thread called motor”y”. Each motor would have it’s own driver. You would reverse the direction on one motor by wiring it ‘reversed’.
You gain nothing, in reality, by maintaing four motor calculations doing only 3 functions, X,Y and Z.
Unless Gaups becomes uber popular, I would doubt Synthetos will add a Platform=Gaups build option, so you will always be recompiling standard builds wilth unique .h files.
The good news is, you have the flexibility, the bad news is, you have to remember to do it and do it. All builds will be named the same (e.g. 75.02)cmcgrath5035ModeratorAlso, could you have a look at this forum post
You seem to have succeeded building on Win 8.1 where he is having issues.
I can’t really help him directly.cmcgrath5035ModeratorFrankly, G2 has many new details, so we are trying to figure out what to ask for.
Item 6 is a 10K pull up to VDD on the GAUPS, feeding both the 8825 drivers and DUE port D8. Vdd=5v, correct? I am not sure DUE port D8 will like that. A pullup to 3.3V would be safer.
Item 8 you answer is unclear, have you loaded your binary to the DUE yet? What does it reports as $fb (firmware build)? I was just on another forum where the member reported building 75.02 less than 12hours ago. But yes, this is a fast moving world.
Pololu DRV8825 should be great – I have been looking at them as well. The GAUPS reference had a pointer to older drivers which were not familiar.
I’ll reinforce my Item 6 concern – Gaups uses Vdd=+5V to power all 4 drivers. I am probably being a bit paranoid, but if +5v should get back to the Due from any leads on the 8825 that might have an active or passive pullup, it could degrade or kill your DUE. I have been slowly thinking thru a custom shield similar to what you are doing and how to power these functions is still on my to-do list.
cmcgrath5035ModeratorI normally build directly on Linux, so have no direct experience with Win.
Can we assume you followed this procedure, exactly?
You will note there that Win8.1 “could be problematic” due to enhanced driver signing demands from the OS. It has been my general experience that Win does a poor job of revealing critical pop-ups, such as requests for authorization, blah, blah; they seem to get burried at the worst time.
I sort of doubt that drivers would impact the build phase, but will come into play when interfacing to the atmel-ice for flashingI am no make guru, but the messaging indicates that it can’t find some expected component(s) in the build hierarchy.
One could wonder how much exercise AS6.2 has had on Win8.1 as well, it was for sure buggy when released by Atmel (form browsing web on similar issues)Suggestions
1. If you are not sure, blow it all away and follow the instructions in the above link again, keep an eye out for complaints from Win8.1.
2. I’ll ask the developers to take a peek at this, but can pretty much guarantee they have little, if any, experience with Win8. The predominately build in MacOS, and Win7VMs when they must.
3. Cross your fingers that anther Win8 user stops by and can offer assistance.Unix ‘make’ (Linux, MacOS) is an incredibly fussy process and does not tolerate faults well for somewhat obvious reasons. You don’t want 1000s of lines of machine generated code making assumptions for you. Unfortunately, one little syntax issue can dump a whole load of confusing error messages on your desk.
cmcgrath5035ModeratorCool. Can you post a photo of a finished item? I have been curious as to how well the surface comes out for typical machines. How long to print?
cmcgrath5035ModeratorJust some comments and questions, this will likely need to be looked at by the developers. A quick look on my part.
1. Rather than posting items like the schematic , I would suggest you post a link to a copy in a Dropbox or similar account. The Schematic is somewhat difficult to read within this forum tool. See2. I am not familiar with the specified drivers. Are you sure DUE can supply adequate +5V current to run them
3. I found a URL for Gaups here; is this the best reference?
4. Are you sure the specified drivers are compatible with 3.3v input?
5. It would appear that in the upper right of your schematic, there is a way to derive the Amotor step and dir from the Y leads, making Gaups pin compatible with Gshield. Is there a reason why you don’t want to use that feature? Just curious.
6. I would carefully review the pullup to +5 (I think) on Gaups pin D8. That might kill your DUE (3.3v logic)
7. Perhaps post (a link) your proposed modifications to gShield-pinout.h, as a commented copy from git-hub, as that will be speaking the developers language most clearly. Make your proposed modifications clear and search-able, something like //Gaups_mod ……..
8. For safety sake, include info on the OS you are using for G2 builds. There are some issues at the moment with the Linux cross compiler that is pulled in by the Makefile.
- This reply was modified 9 years, 11 months ago by cmcgrath5035.
- This reply was modified 9 years, 11 months ago by cmcgrath5035.
cmcgrath5035ModeratorI was hoping someone would comment, I have no useful experience with this status message.
Still occurring?
All Gcodes or just some?Your tinyG FW is not the most current (438.02 is most current), but can’t say that would resolve.
Perhaps post your settings and the (or one of) offending Gcode files.
Suggestion: follow procedure here:cmcgrath5035ModeratorYes, what you want to do has been done by many, but not me so details would be third hand.
If useful, I suggest you review tinyG schematics here:and the go to this Forum at level:
and use the search capability , searching on Serial or…
Remember, tinyG is 3.3V logic, so direct wiring to UNO may not be appropriate (5V logic), but DUE probably OK (3.3V logic).
As you will see in the schematics, the Tx Rx you see result in a “wired OR” with the ftdi output and those terminals.
You can (carefully) use both channels simultaneously in parallel (a debugging hint) -
AuthorPosts