There is a very good reason to post issues on github even if you think a project is abandoned, because then the issue is recorded in a public place for people to be aware of, comment on, and also fork and create a fixed version. Open Source is about sharing.
Did anyone actually try asking me rather than pass around rumor??
Also the rumors of Due being dead are simply not true… The Due was never made by arduino.cc in the first place, it was made by SmartProjects who rebranded as http://arduino.org. There is no reason to believe that http://arduino.org are going to stop selling it. So the news on the Due is that nothing has changed.
Funny, I’ve been at the forefront of developing 32 bit controllers when everyone was saying “don’t need it, 8 bit is fine”, now that 32 bit M3 are becoming popular, it is now considered “old and too slow”. How things change
I don’t think things have changed that much to make an M3 too slow, but I guess people need a way to sell new products.
RAMPS-FD does not depend on Due, it can be used with any baseboard with a Mega shield footprint. As I mentioned earlier, that extends up to 1GHz Cortex A5.
On a side note, is there always this amount of negativity here?
@bob_cousins I wouldn’t say Ryan is being negative so much as expressing his opinion.
As for Smartprojects, the way Martino split the company and basically tried to appropriate the Arduino goodwill and basically hijack the trademark leaves a bad taste in my mouth. He went behind his partners backs and registered the trademark for Aruduino in Italy. Poor form. They needed to scale production outside Italy and as Martino owned the fab they were using he didn’t want to use any other houses. To keep the business to himself he tried to appropriate what the team had built.
I see the SAMA5 boards with an R3 footprint a great host for this sheild. Now if the firmware could catch up…
@bob_cousins Please don’t get me wrong, I very much appreciate what you’re doing and am 100% in favor of M3-based boards over Atmega-based boards. No negativity intended. 8bit controllers are ridiculously obsolete – they’re a major bottleneck in both firmware development and high-performance printer control. Any 32bit board is a big improvement and will meet 99.X% of users’ needs.
Here’s my take:
The AVR boards are completely maxed out with execution-optimized firmware like Repetier and Sailfish. Major new features cause performance losses in other areas. There is no room for growth. These are terminally mature and should cease major development work.
The M3 chips in use today (on Due, Duet, Alligator, Smoothieboard, etc) are maxed out with the current generation of non-execution-optimized firmware. When I say non-execution-optimized, I mean the math and algorithms aren’t tuned for maximum execution speed on the chips. Marlin4Due and Repetier-Due are full of AVR-optimized legacy code and don’t take advantage of some M3 capabilities. They are easily maxed out by feature-rich printers at high print speeds. Smoothieware is highly abstracted with OO code for legibility and extensibility, and is basically out of RAM. RepRapFirmware has added highly clock-cycle-intensive features like non-segmented delta kinematics, and is maxed out at high print speeds. Most users don’t see these issues yet because their printer hardware isn’t performing at a high enough level to push the processors. But in the last couple years, printer hardware and advanced features have reached the point that the MCU is becoming the bottleneck even with the 32bit M3 line. Higher performance will require hardware-specific algorithm optimization that kills portability. We are already seeing this discouraging firmware devs from implementing cool new ideas. The Smoothieware and RepRapFirmware guys are both focused on their respective v2 hardware now.
We can get vastly more processing power at similar or only slightly higher cost, by switching to SBC-class options like BBBs. 1GHz and two PRUs? Yes please! That’s the kind of controller hardware that gives firmware devs serious headroom to stop worrying about clock cycles for years to come and focus on making printers run better. I’m still agnostic between approaches – there’s pros and cons to Fastbot, Redeem, and MachineKit – but that’s the direction that I personally think the community should be moving.
@ThantiK to my experience repetier runs still much smoother than smoothieware, also temperature control on smoothie is kind of garbage compared to repetier. (no dead time control, no fast bang bang). I’m looking forward to this board.
@bob_cousins Using “NG” for a new name (e.g., RAMPS-NG) may prove confusing given that David Crocker (dc42) announced a few months back the Duet-NG electronics (Cortex M4 based Duet; an Atmel SAM4E8E).
I think Ryan is right, the BBB is less than 2x price of Due, yet vastly more powerful, and you get a full OS and rich application environment in Linux. There is now a good range of printer and CNC capes. I see no point in reinventing an old wheel, I would rather do something that adds something new.
RADDS is a perfectly viable alternative to RAMPS-FD, plus there is Duet, upcoming DuetNG and maybe others that are compatible with Due. Then there is a bunch of Smoothie boards to choose from.
I think RAMPS-FD has served its purpose, but I might do one last round of bug fixing and call it a day.
I think the nature of DIY printing has matured a bit, people are now looking for plug and play solutions, with richer features.
An area of 3DP which needs some careful work is safety. E.g., standalone safety cutoff hardware/electronics which can be added to existing printers. And a forum to discuss it in might be https://groups.google.com/forum/#!forum/3dp-ideas . Just a suggestion in the very unlikely event that you are devoid of ideas of where to direct your energy.
In RAMPS fd v2.1 circuit, the values of resistors R701, R702, R703, R704, R901, R902, 220 ohms are stated. Are these values correct? Isn’t 220 ohms too much for Gate MOSFETs?
My next question is, can I use irl3713 or FDP038AN06A0 instead of irlb8743? If so, how much should I change the gate resistance values?
This four-year-old thread from Google+ on a different topic about an abandoned design is probably not the best place to get answers to your question generally. But 220 ohms isn’t necessarily too much for a MOSFET gate.
Clearly the replicape didn’t take off. Mine died before I could even get my printer dialed in due to a known frequent bug so I cut my losses there. On the other hand, 32-bit boards are now so incredibly cheap I can’t imagine building a new 8-bit board.
@bobc if you login using google, and with the same google account you used for Google+, it will associate you with the old ID and if you ask I can even merge it all together under one account for you, just DM me a request after doing that login.
You have designed this board. The answer to this question does not take much time. Please tell, can I use IRL3713 instead of IRLB8743? If yes, do I need to change the gate resistance?
All specifications are the same and only Qg is different.
IRL3713 — Qg Total Gate Charge Vgs4.5v ––– 75 -110nc
IRLB8743 — Qg Total Gate Charge Vgs4.5v –––– 36 -54nc
Many weeks of work went into the design, checking datasheets, doing calculations, building prototypes, verifying the performance with oscope, going back round the loop. I am a software designer, hardware design is not my main skill.
The gate resistor was determined by testing a prototype, but I don’t think the value is critical.
I really can’t give an instant answer off the top of my head, sorry to disappoint you. If you think the specs are the same, why not just try it? If you are a beginner to electronics, I suggest asking for help on a site like eevblog.
