Mightyboard Rev E / Customising Sailfish

[Tim Deards]

Hi All

New group member here in the UK who would really appreciate some expertise/advice with the Mightyboard Rev E / Sailfish.

 

I would really appreciate some advice on configuring / compiling a custom configuration for Sailfish 7.7 and flashing to a CTC printer.

I have 2 CTC clones one 4 years old and the other 3 years old which have been happily running Sailfish 7.7 updated through RepG from the repository.

 

I have just got a third CTC clone advertised as not working.

The printer is about a year old and came with two CTC Rev E Mightyboards but both have different issues:-

The first one is pretty dead with no display on power up which will take a bit more prodding with the fluke and oscilloscope but ultimately I will have to get it repaired if I want this one working 

I am guessing the 1280 is badly blown. Something major must have happened here as all 5 botsteps supplied with the printer are also blown

 

The second board powers up fine is running Sailfish 7.7 but only the digipot on X axis is working and there are no step pulses / direction high low / or enable low on the Y-axis botstep headers.

Apart from this and a blown 47uF 50V electrolytic capacitor on the A Extruder, everything else works on the board and communicates to RepG / Makerbot Desktop over USB.

I think that this one is salvageable I am looking to see how much of the below is possible to get this CTC up and running on a single B Extruder.

1:- I can live without the "Lower Stepper VRef's while heating" as the CTC comes with a 360W 24Vdc supply so can do without the digipots unless these are mission critical in other scenarios in Sailfish. 

I have several A4988 stepsticks which I can modify with the header pins the other way and MS3 linked high for 16x microstepping with ItripMax set from the onboard pots once I have checked the sense resistor values.

 

2:- The schematic shows several unassigned "hacking" pins available from the 1280 broken out to pads on the PCB some of which are connected to debug LEDs.

I have visually checked the CTC board and the traces all seem to be there from the 1280.

I am hoping to reassign the unused "hacking" pins in the software to provide the signals that are not working then solder in pin headers on the pads.

On the A4988 for the Y axis I will solder the affected signals pins upwards so I can connect to them with female jumpers and hot glue.

I have seen where the Stepper Pins are defined in the Configuration.h (I was on the Sailfish V7.10 github site any opinions on this version?) 

Is there anything else I would need to consider such as reserved pin function or assigning analog pin to digital etc?

3. Probably the biggest question for me, as I am not a software guy is how to compile the Sailfish package using my modified configuration.h file and how to go about flashing Sailfish 7.7 to the printer?

As at the top I upgraded both my existing machines using RepG.

Any help or advice is greatly appreciated in advance 

Thanks

Tim Deards

[Rich Webb]

There's a VirtualBox package with the specific build environment and instructions. It's on DropBox, links here.

[TobyCWood]

Not sure where you are starting from... perhaps from experience with Marlin? Sailfish has been pretty fallow as of late and all new versions (if any) are coming from companies like CTC and they do not share their info. If you search for "Sailfish" on thingiverse you will find all the urls you need to get the last source, but support on a FW that has kinda reached the end my be hard to get. These days many have moved on to Duet/RepRapFW and Smoothie.

Also... keep in mind that these CTC boards are clones of the mighty board and are not the mighty board. Actually this may not be a bad thing, but if you are using open source mighty board docs for a CTC clone they may not be accurate. Again the Chinese clone makers took the open source info changed it then typically are reluctant to share what they did.

[Tim Deards]

Hi both and thanks for your replies

I am battling a bit to get the VM software packages to work together ( 64 bit Win10 M/C's but with no Hyper-V meant 64 bit Ubuntu would not work) but some progress is being made.

First 32 bit download of Ubuntu would not unpack with Vagrant so downloading and trying again.

Printer wise I have only had CTC's since starting and hopefully will have 3 all printing ABS / PETG 

I also have a Cetus just for pretty things in PLA (Outstanding quality prints from this).

I am from a PLC / Automation Electrical and Electronic Engineering background and the first thing we look for with blown I/O is spare I/O already fitted that we can re-assign to the functions in software.

Obviously in industrial settings we use OEM software for this so it is relatively simple to "Find and Replace" the addresses or tags which need to change.

I use the final version of MB Desktop and spent a fair bit of time modifying the "JSON and Schema" files so there is more functionality in slicing and I can adjust a fair bit more than standard from the custom profile settings pages rather than the text editor. So I have looked at Cura / Flashprint but MB Desktop just works so its good enough. 

So I am trying to do the same here by reassigning some stepper function functions in Sailfish to the spare 1280 I/O on the Mightyboard with a few blown O/P's on the Y axis to get the latest machine running.

I might even reassign the B Axis I/O complete to the Y Axis as one of the CTC's is now single extruder.

If I can't get this working and done then more than likely this single extruder printer will get a 32 Bit Lerdge X Board as only 66GBP with 2208 Trinamics, (just need to change to thermistor for hotend and get the bed module). 

These seem to work pretty well out of the box from what I can see and all set-up is done from the touchscreen. 

Hopefully I can export Gcode from MB Desktop that will allow it to work.

 

The Mightyboard out of the single extruder will then replace the U/S one and go to the new printer for more trials and tribulations with dual extrusion.

Many thanks

Tim Deards

[TobyCWood]

or...

go buy a Duetwifi. Whats more valuable the cost of the board ($140) or your time?

With the Duetwifi you get all the options for temp sensor, you get much improved FW(Sailfish has an unfixed issue which will not likely ever get fixed), way easier to set config, way easier to add any option you may want, full 32 bit words, super smooth fast curves in the tool paths, quiet drivers, std gcode and a pretty slick web interface. When any of my Mightyboards die they will be replaced with Duetwifis. I won't waste my time bothering to fix any of them. I replaced 1 Mightboard clone with a DuetWifi already. It's a no brainer, way better solution hands down.

[Tim Deards]

Hi

Sorry for delay in replying been busy getting the modded Sailfish build to upload and work.

Essentially remapped the Y Axis to use the EX B Socket.

Modded using Ubuntu VM and built with Scons but transferred the .hex to windows to use the RepG Sailfish version AVRdude and command prompt to flash to board over USB.

Had tried getting RepG to look at local firmware.xml but no joy.

After flashing the hex the start up tune played so didn't manage to brick it.

Tested jogging Y axis motor when plugged in to EX B driver and OK. XZ and A axes quickly tested OK

Just have to replace 47UF electrolytic Cap on EX A VMOT supply and reverse some A4988 drivers as only have one spare Botstep17 and should be ready to try a print.

I scored the not working CTC for GBP65 with 7 rolls of PLA and a second blown mightyboard. so essentially the printer was free and if I can get it working for nada even better.

It would be nice to have some more modern boards on all three CTC's but the cost to upgrade to Duet 3d would be in the region of 200GBP/unit with screens and my finances are really short.

A Lerdge-X is 41GBP/unit but would need heatbed expansion modules but I have the A4988 drivers and all the set-up is done from the front screen. 

The CTC printers are remote from the house so normal workstream is via SD Card and noise is not an issue.

I get pretty reasonable print quality in ABS running max to 90mm/sec on infill and raft upper layers although long curves are a bit segmented. 

The only bug I have found is that if you are preheating and go to filament loading it seems that both extruder and platform are ramped at the same time which could be an overcurrent issue for units with external power supplies.

What is the bug that you refer to?

Once again thanks for your replies

Tim Deards

On Tuesday, 13 August 2019 05:01:17 UTC+1, TobyCWood wrote:

or...

go buy a Duetwifi. Whats more valuable the cost of the board ($140) or your time?

With the Duetwifi you get all the options for temp sensor, you get much improved FW(Sailfish has an unfixed issue which will not likely ever get fixed), way easier to set config, way easier to add any option you may want, full 32 bit words, super smooth fast curves in the tool paths, quiet drivers, std gcode and a pretty slick web interface. When any of my Mightyboards die they will be replaced with Duetwifis. I won't waste my time bothering to fix any of them. I replaced 1 Mightboard clone with a DuetWifi already. It's a no brainer, way better solution hands down.

On Monday, August 12, 2019 at 4:37:10 PM UTC-7, Tim Deards wrote:

Hi both and thanks for your replies

I am battling a bit to get the VM software packages to work together ( 64 bit Win10 M/C's but with no Hyper-V meant 64 bit Ubuntu would not work) but some progress is being made.

First 32 bit download of Ubuntu would not unpack with Vagrant so downloading and trying again.

Printer wise I have only had CTC's since starting and hopefully will have 3 all printing ABS / PETG 

I also have a Cetus just for pretty things in PLA (Outstanding quality prints from this).

I am from a PLC / Automation Electrical and Electronic Engineering background and the first thing we look for with blown I/O is spare I/O already fitted that we can re-assign to the functions in software.

Obviously in industrial settings we use OEM software for this so it is relatively simple to "Find and Replace" the addresses or tags which need to change.

I use the final version of MB Desktop and spent a fair bit of time modifying the "JSON and Schema" files so there is more functionality in slicing and I can adjust a fair bit more than standard from the custom profile settings pages rather than the text editor. So I have looked at Cura / Flashprint but MB Desktop just works so its good enough. 

So I am trying to do the same here by reassigning some stepper function functions in Sailfish to the spare 1280 I/O on the Mightyboard with a few blown O/P's on the Y axis to get the latest machine running.

I might even reassign the B Axis I/O complete to the Y Axis as one of the CTC's is now single extruder.

If I can't get this working and done then more than likely this single extruder printer will get a 32 Bit Lerdge X Board as only 66GBP with 2208 Trinamics, (just need to change to thermistor for hotend and get the bed module). 

These seem to work pretty well out of the box from what I can see and all set-up is done from the touchscreen. 

Hopefully I can export Gcode from MB Desktop that will allow it to work.

 

The Mightyboard out of the single extruder will then replace the U/S one and go to the new printer for more trials and tribulations with dual extrusion.

Many thanks

Tim Deards

On Sunday, 11 August 2019 20:59:40 UTC+1, Tim Deards wrote:

Hi All

New group member here in the UK wh

The Cetus is in the home office on the desk.o would really appreciate some expertise/advice with the Mightyboard Rev E / Sailfish.

[Dan Newman]

> Had tried getting RepG to look at local firmware.xml but no joy.

You can build your own RepG if you have infinite free time:
https://github.com/jetty840/ReplicatorG . I really don't recommend
trying that.

> After flashing the hex the start up tune played so didn't manage to
> brick it.

Always a good sign, not bricking things.

> Just have to replace 47UF electrolytic Cap on EX A VMOT supply and

> I get pretty reasonable print quality in ABS running max to 90mm/sec on
> infill and raft upper layers although long curves are a bit segmented.

> The only bug I have found is that if you are preheating and go to
> filament loading it seems that both extruder and platform are ramped at
> the same time which could be an overcurrent issue for units with
> external power supplies.

That's not a bug per se: that's a compile time setting. The "heater
on steroids" is saying to do the HBP and extruders in parallel as
opposed to serially (i.e., first HBP, then the extruders, both extruders
at the same time if you command them both to temp).

> What is the bug that you refer to?

I believe TobyCWood was referring to some of the motion control issues
introduced by a contribution midway through the aborted 7.8 alpha test.
Hence my suggestion to go back in time a bit.

Dan

[TobyCWood]

Yup. It gets noisy and rough at higher speeds.

Sure is nice to see DAN again!!!