use second sanguinolulu to expand extruders

Hi Everyone,
This is my first posting.
I have successfully built a Rostock with traxxas ends and Sanguinolulu 1.3a (with mod for fan control.)
I am using Kisslicer, Repetier host and both Marlin and repetier firmware.
My question is ....can I easily address 2 sanguinolulus using 2 extra drivers ( stepstick) on the second board.
I could not find anyone doing this on the forums.
My first thoughts were to piggy back initially one extra drive board for dual extruder and find a spare i/o to enable it as i had with the fan mod.
Then I thougtht Richraps triple blend head was where I would want to go (eventually). So I thought 2 sanguinolulus was a simple solution with lots of expansion.
But here is the crunch..I don't know how to program this, so need help.
Any suggestions welcome.

RepRapPro does this with two Melzis, which are quite similar to Sanguinololus. You would have to do a bit of research and looking through RepRapPro's firmware, but I think it should be doable with two Sanguinololus. See: [reprappro.com]

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

@New Perfection ,
Thanks for the quick response, just what I needed for further investigation, I have just finshed building the second Sanguinolulu and am rearing to try it. I will post any positive results.

Just an update after looking at all the possibilitys to do this.
1. Email Repetier for help.
2. Appy full Reprap pro Tricolor mendel firmware even though its a rostock delta! To verify i/o link to slave works.
Unless I can find other alternatives I shall try this with dry run.
3. Try to isolate the slave software and turn a mendel into a delta ! or a delta with slave addons.
This is where I lack the programming skills, but if no progress on any front will have learn to write some software.
I will anyway but just a little impatient to get this done now I have thought of it.

If anyone can come up with a comms patch or a few pointers for Rostock varient of Marlin , much appreciated.
Also any other forums discussing this.

I feel alone in trying this, has no one else tried it. I feel a multi feed bowden nozzle is the way forward, although I will accept a working dual as a starting point.

Finally I am going to the TCT show next week and would like to have progressed a little so I can ask sensible questions

Edited 1 time(s). Last edit at 09/18/2013 06:07AM by prof111.

Further update.on stages 1,2,3 1. Had a reply from Repetier, nothing in the pipeline there, recommended going to a Rumba.
2. Applied tricolor mendel software to my delta as a test and got correct response from the second Sanguinolulu , but dreadfully slow all round,(possibly because no sensors connected). I have a fan on/off mod which under normal circumstances is instant.. with extra board with serial connection...It took seconds!!! ( connection to original board)
3. Played around at combining sofwares but out of my depth, too many variables ..need to have some tips from software gurus.
I still believe it is possible to do this, if the firmware can be streamlined. I still think this is a better option than all of the present boards as it would give 5 extruders to do C,M,Y,B,W . and be more cost effective.

Or are there other expansion methods?

*******Please may I have a few opinions on this as buying a Rumba only gives tricolour for the future.


Also do those running tricolor mendel notice a slow down in response from mono? If not then the extra serial may not be a problem, and this is worth persuing.

Edited 1 time(s). Last edit at 09/24/2013 09:30AM by prof111.

I think that the extra serial hop will inevitably introduce some delay, although I wouldn't expect it to take seconds.

While not wanting to discourage you from investigating dual Sango boards, as it does give a cheap and immediate option, I am working on a generic extruder expansion board. I can do 6 extruders fairly easily, possibly 8. Initially it would be 4. The interface will use SPI/I2C type comms, so does not require extra firmware on the slave side, however I am also looking at a "smart" version.

My upcoming RAMPS-FD board will support 3 extruders and should be compatible with Arduino Mega and Due, and a further LPC1768 board I have planned. So that will give a range of 3 to 11 extruders, hopefully that is enough for everyone?!

Of course all of this will require support in firmware on the master side, but I figure the hardware needs to be there first. Volunteer guinea pigs welcome

Here is a render of 4 stepper expansion board, as you can see there should be room for 6 extruders:


RAMPS-FD-EXT by donotdespisethesnake, on Flickr

@bobc
Thankyou for this response.
If i may come back to you more fully after tomorrow when i am at the tct show, hopefuilly picking a few brains!
I was also thinking of addressing the second board drivers directly in some way, I hadn't worked it out though.
Also I haven't yet worked out the necessary bit rates of the interfaces with respect to the amount of interupts and polling.
Question: What are the limits / bottlenecks?

My experience is in the maintenance of Broadcast equi[pment especially cameras where i2c is used fior control, but runs into trouble when gamma requires 32bits to work smoothly. As I now have time to devote to learning the programming aspects, it may take me a while to feel confident to understand the whole picture. It is great fun though and to see a delta printer work from individually sourced bits is quite a buzz!

Hey, have fun at the TCT show! I might get along to the 3D Printshow in London.

> Question: What are the limits / bottlenecks?

That is a very good question. The data rates are quite low, but it would be good to keep step pulses synchronised with low latency.

For temperature readings, cheap ADCs only handle a few samples per second, which should be enough and polled say every 100ms by I2C. I am hoping to use SPI to shift out up to 32 bits which contains data for step and dir, enable and heater output. That will drive 8 extruders with 4 bytes transferred, SPI can run maybe 10Mb/s. Latency should be in order of 10us I hope.

PWM output for heaters is more difficult, and I think would be too much overhead to do via this method. However, a low rate PWM or modified bang-bang approach may be sufficient.

For true PWM control, it would benefit from a microprocessor, which could still be driven by SPI and use a very simple firmware. That would also have the advantage of being able to perform temperature control directly, and operating as a standalone motor controller with suitable firmware.

Hi Everyone,
I went to the TCT show yesterday, and met with amongst others Dr Adrian Bowyer,Nophead, Richrap ,Roland and Tony Locke of Think3dprint3D. All of which spent valued time answering my numerous questions.
My conclusions are to continue a little longer with the 2 Sanguinolulus until a clear solution is available.Everything is in the pipeline.

@ bobc , 1 will come back to you lthis evening about your helpfull suggestions as soon as I can sit down and think everything through and be constructive with my questions.

Edited 1 time(s). Last edit at 09/26/2013 06:24AM by prof111.