Modular Electronics

[Folknology]

Yesterday evening a number of us convened on IRC to discuss (among other things) the electronics for our build.

Previous meetings and conversations have led us from a monolithic design to a modular design, I won't repeat that hear as its already documented on the original meeting thread. Yesterday however folks had an opportunity to tour the modular electronics design interactively using online tools and IRC chat, I believe the minutes of which will be published later for those interested. We spent nearly 4 hours going through the various design features and choices that had been made, overall I felt that folk were positive about the design features and that we had accomplished pretty much everything feature wise that had been requested. The electronics focuses around 2 modules a control module (the brains) we named Seguino and the dual stepper modules (we need 2) the brawn. each stepper module can drive 2 steppers and 2 powered PWM (Pulse Width Modulators) loads like heaters for extruder and heatedbed. Thus a basic solution would consist of 1 Seguino boards connected to 2 Dual Stepper boards.

The Seguino board also features a 7 Segment display for showing temperature and possibly diagnostic information, along with a rotary encoder which enables manual setting (or real-time tweaking) of the temperature. The display feature and manual temperature setting didn't bring up a great deal of agreement and there was some doubt about there value. I mention this because Barnaby asked a very good question that amongst all of the other concurrent conversations on IRC didn't really get answered. Barnaby asked 'Sounds good, presumably having three of the 10pin connectors on the Seguino wasn't possible?' The connectors he is referring too are the IDC connectors that connect the Seguino to the Dual Steppers (1 each module), this kept me awake last night and I investigated it this morning.

The answer to Baranaby's question by the way is actually it is possible to have a third IDC connector, this connector could thus provide another 2 stepper motor channels and 2 more PWM channels. This would be useful if you wished to upgrade to 2 extruders for instance. However in order to add the third IDC connector we would have to loose the 7 Segment display and rotary encoder. Thus it would be come a choice of one or the other and what the relative value of one over the other might be. In many ways being true to RepRap principles I would suggest that another IDC connector would be the better option as we are seeing more and more moves in the direction of additional extruders and very little in the way of adding 7 segment displays and rotary encoders, thus indicating that the former be of better functional value. It also occurred to me this morning that the 7 Segment display and rotary encoder could itself be made into a separate optional module itself that could use the same IDC connector as the dual steppers. That would mean one could choose the interactive display module (7 seg) connect it using the third IDC connector and at a later date if desired unplug it and plug in a third stepper module in stead, similar could occur the other way around replacing third stepper module with display module. Going this route would of course be more in tune with our modular approach, thus I would suggest that this is the way we progress as I need to finalise the electronics today in order to get the PCB ordered and tested and running on a Prusa ASAP.

Please let me know your thoughts.

regards

Al

[Drew]

Sorry I could not make the IRC last night, unfortunately Tuesdays are
not the best for me...

But I give a +1 for an extra IDC connector / stepper board. I like the
idea of expanding to multiple heads.

Correct me if im wrong but could we not also do real time temp changes
and read temp values in software?

Ta
Andrew

--
Drew
a3d...@googlemail.co.uk

[Alan Wood]

On Wed, Oct 5, 2011 at 11:29 AM, Drew <a3d...@googlemail.com> wrote:

Sorry I could not make the IRC last night, unfortunately Tuesdays are not the best for me...

But I give a +1 for an extra IDC connector / stepper board.  I like the idea of expanding to multiple heads.

Correct me if im wrong but could we not also do real time temp changes and read temp values in software?

Yes we could read temp via software/firmware , the 7 segment display was for stand alone mode (using the built in SDCard) and or just visual embellishment and tweaking ;-)

[Malcolm Napier (gmail)]

I am in the camp in favour of the additional IDC connector.

 

Just to reprise our experience with RAMPS 1.2 – adding an SD card module and putting the file to be printed on the SD card improved the print quality for reasons that we don’t really understand. We don’t have a display and if we were to have a display, we would also want control buttons – so that the printer can be used completely stand alone – without connection to a PC.

 

As far as I can see it, having the extra IDC connector provide options is the better way forward. We don’t have to use it at this stage (thereby replicating what we currently have - printer controlled and monitored by PC but with print files either on PC or on SD card) but there are the options currently identified and probably some others we haven’t yet thought of.

 

Regards,

Malcolm

[Hugo Mills]

On Wed, Oct 05, 2011 at 03:15:32AM -0700, Folknology wrote:
> Yesterday evening a number of us convened on IRC to discuss (among other
> things) the electronics for our build.

Thanks for leading us through the details of the two boards. A
tough job with everyone pitching in at random moments with awkward
questions.

[snip]

> The Seguino board also features a 7 Segment display for showing temperature
> and possibly diagnostic information, along with a rotary encoder which
> enables manual setting (or real-time tweaking) of the temperature

[snip]

> actually it is possible to
> have a third IDC connector, this connector could thus provide another 2
> stepper motor channels and 2 more PWM channels. This would be useful if you
> wished to upgrade to 2 extruders for instance. However in order to add the
> third IDC connector we would have to loose the 7 Segment display and rotary
> encoder. Thus it would be come a choice of one or the other and what the
> relative value of one over the other might be. In many ways being true to
> RepRap principles I would suggest that another IDC connector would be the
> better option as we are seeing more and more moves in the direction of
> additional extruders and very little in the way of adding 7 segment displays
> and rotary encoders, thus indicating that the former be of better functional
> value. It also occurred to me this morning that the 7 Segment display and
> rotary encoder could itself be made into a separate optional module itself
> that could use the same IDC connector as the dual steppers.

This sounds like a good modification. I was sceptical about the
utility of the 3x7-seg display. I think having the extra header would
be a better idea (and we can supply more detailed feedback with an
alnum/pixel display attached via the I2C/SPI bus, or through software
and a serial connection to the PC as Andrew suggests).

Hugo.

--
=== Hugo Mills: hugo@... carfax.org.uk | darksatanic.net | lug.org.uk ===
PGP key: 515C238D from wwwkeys.eu.pgp.net or http://www.carfax.org.uk
--- As long as you're getting different error messages, ---
you're making progress.

[Alan Wood]

Ok I am moving ahead with this final change in order to expedite the initial PCB production to facilitate testing

Here are the new project links and naming:

Open Motion Controller (OMC) Module formerly Seguino - http://solderpad.com/folknology/open-motion-controller/

Open Dual Stepper - http://solderpad.com/folknology/dual-stepper-motor-module/

(PS you don't need username/passwords to view on solderpad now)

Please give me any feedback today (I will finish routing OMC by tomorrow) as my intention is sending gerbers to china pm tomorrow.

regards

Al

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iD8DBQFOjDn3IKyzvlFcI40RAi23AJ9NQCa0m0tj1xfVdMo3FpY2ZeIrSwCfTd3N
AhrE7rCmTJGOV9hRQ6La4KM=
=BztA
-----END PGP SIGNATURE-----

[Hannah Napier]

Hi Alan,

Just been writing up the minutes from last night on the IRC and had a realisation. The thermistor inputs are on the main controller board and there are only two. You have been busy building in flexibility to expand to dual extruders, but with the current thermistor set up one will have to forgo a heated bed in order to take advantage of this. Would it not make more sense to have thermistor inputs grouped with the heaters on the dual stepper motor boards?

Hannah

[Alan Wood]

On Wed, Oct 5, 2011 at 2:39 PM, Hannah Napier <hmna...@gmail.com> wrote:

Hi Alan,

Just been writing up the minutes from last night on the IRC and had a realisation. The thermistor inputs are on the main controller board and there are only two. You have been busy building in flexibility to expand to dual extruders, but with the current thermistor set up one will have to forgo a heated bed in order to take advantage of this. Would it not make more sense to have thermistor inputs grouped with the heaters on the dual stepper motor boards?

Hannah

Excellent point, Thank you, possible solutions:

1) used Thermocouples and I2c interfaces (more expensive)

2) Use I2c temp sensors less expensive than above

3) Re-examine pin usage to try to salvage another suitable pin

Will investigate imediately

regards

Al

[Alan Wood]

On Wed, Oct 5, 2011 at 3:56 PM, Alan Wood <folkn...@gmail.com> wrote:

On Wed, Oct 5, 2011 at 2:39 PM, Hannah Napier <hmna...@gmail.com> wrote:

Hi Alan,

Just been writing up the minutes from last night on the IRC and had a realisation. The thermistor inputs are on the main controller board and there are only two. You have been busy building in flexibility to expand to dual extruders, but with the current thermistor set up one will have to forgo a heated bed in order to take advantage of this. Would it not make more sense to have thermistor inputs grouped with the heaters on the dual stepper motor boards?

Hannah

Excellent point, Thank you, possible solutions:

1) used Thermocouples and I2c interfaces (more expensive)

2) Use I2c temp sensors less expensive than above

3) Re-examine pin usage to try to salvage another suitable pin

Will investigate imediately

regards

Al

 

OK I have fixed it (and updated the schematic) I removed the status control pin as this effectively doubles up the enable status anyhow. I then moved the sdcard select to that pin so I could recycle the old CS pin a third thermistor input pin (TE).

[mikethebee]

I have posted an edited version of the chat at
https://docs.google.com/document/d/1Kg9Qtj9KEQmZYpe9BxurQSZ-fafDjny74T9WsRYqszg/edit?hl=en_US

The edit groups the conversations into subjects, removes a bit of the
chatter, but shouldn't lose anything significant I hope. I know Hannah
is going to do a proper summary, but I thought it may be helpful to
some to have a log style record., let me know if you disagree.

BTW: should we be adding ourselves to the order list?

Mike.

[Hannah Napier]

Am I right in thinking the situation now is three thermistor inputs on the "motherboard"?

Hannah

[Alan Wood]

Yes now we have 3 and I have now updated the schematics and connectors for the project so you can take a look directly at the new layout etc..

Does this close the issue..

regards

Al

[Hannah Napier]

Sure that's great for what we want right now. Let's hope we don't start wanting triple head extrusion plus heated bed any time soon :) (I have just seen that on the BitsFromBytes 3D Touch...)

Hannah

[Alan Wood]

On Wed, Oct 5, 2011 at 4:31 PM, Hannah Napier <hmna...@gmail.com> wrote:

Sure that's great for what we want right now. Let's hope we don't start wanting triple head extrusion plus heated bed any time soon :) (I have just seen that on the BitsFromBytes 3D Touch...)

Hannah

Ooh shiny! got a link for that I've been thinking about an CMY triple head for ages

[Hannah Napier]

http://www.bitsfrombytes.com/catalog/3dtouch and a video: http://www.youtube.com/watch?v=9xgSzZXGoD8

Hannah

[Alan Wood]

Wow look at the extruder on that!

[Alan Wood]

All is not lost..

You know it would still be possible to do a triple headed extruder by switching the heatedbed to I2C temp sensor

regards

Al

[Alan Wood]

Ok the design is routed and now passes all the DRCs, I am just double checking everything before I send the gerbers, it would be worth a last minute check by those who feel they might spot something (e.g.support team).

Changes include:

1) Added optional 20Pin PSU connector we don't need this but might be useful for testing and experimentation.

2) I2C is now a secondary function and is superseded by the use of the third dual IDC channel. This enables us to add another two steppers and PWM channels later if required. We could also use this as a display extension as discussed before.

Any feedback let me know asap..

regards

Al

[Drew]

In my bleary eyed state it looks good to me...

-- 
Drew
a3d...@googlemail.co.uk

[Alan Wood]

Thanks Drew,

Please do look it over if you have chance.

This morning I had a brain wave about the potential I2C/Gdrive conflict and have solved it by integrating the I2C and  Gdrive Jumper enable into a single set of connectors so you go one way or another. with the jumpers it uses Gdrive without it becomes an I2C header - Very pleased with that design win means you can sacrifice Gdrive for an I2C peripheral if it suits.

regards

Al

[Drew]

Hi, yep i like that. but what does Gdrive stand for?  I can guess all the others. :)

Andrew

-- 
Drew
a3d...@googlemail.co.uk

[Alan Wood]

On Fri, Oct 7, 2011 at 1:04 PM, Drew <a3d...@googlemail.com> wrote:

Hi, yep i like that. but what does Gdrive stand for?  I can guess all the others. :)

Andrew

I have named the stepper channels on the Open Motion Controller to make it easier when it comes to firmware (we may still change this scheme). for the steppers we have X,Y,Z for the axis and E for the extruder. With the third IDC connector we added another two stepper channels which I have named F and G - Hence Gdrive or Gstepper or GChannel.

regards

Al

[Drew]

Fair enough, just thought it might have a meaning like the others. :) I like it.

-- 
Drew
a3d...@googlemail.co.uk

[cynar]

Just a question regarding the board design, does it use a thermistor
or a thermocouple?
I know a lot of the other designs are migrating towards them.
If so, these are well worth looking at. They are pre assembled, so no
faffing with small leads in the hot end. (I had problems with shorts,
etc on my thermistor initially).
http://www.dealextreme.com/p/high-temperature-probe-sensor-350-c-max-68458
About £1.50 each (even less in bulk).

[Malcolm Napier]

> etc on my thermistor initially).http://www.dealextreme.com/p/high-temperature-probe-sensor-350-c-max-...

> About £1.50 each (even less in bulk).

We have a thermocouple that looks very similar to that. It was
provided with our multimeter. We used it to sense check the thermistor
but never thought of fitting it to the hot end.

Will using a thermocouple require different electronics? (I am not
sure how a thermocouple works.)

Regards,
Malcolm

[alan cocks]

thanks al. sounds good to me. i am away just now. i support al's suggestions. and apprrciate the effort being given. alan cocks

[Alan Wood]

Yup the electronics are different. I use thermocouples for my Temperature based PID stuff. Thermocouples (of which there are several varieties) are made by welding two different metals together.When heated they generate very small voltages (micro volts) in proportion to the temperature enabling them to be used for temp measurement. In order to be useful however they have to be linearised as they have curved responses, in addition to taking into account local temperatures for the cold junction (where the thermocouple connects to the electronics). A circuit can be formed using something like an LT1025 to deal with cold junction and temp curves, this is then used in conjunction with a precision low drift low bias/offset opamp with a gain of 200-300 and low frequency roll off. The signal is then piped into the ADC on the MCU in the normal manner. Thus even for one channel it adds at least 2 chips which can be expensive. This isn't the only option as other solution exists, but in all cases there is an increased cost and increased BOM.

In summary they are better and can be more accurate but they need more parts and often mean greater expense.

Also the cheap ones tend to be to fragile (they have brittle glass coatings) and easily break, it's normally worth paying the extra for K-types coated in a more flexible PTFE or similar depending on temp range and use.

regards

Al

 

Regards,
Malcolm

[cynar]

A thermistor is a resister that changes resistance with changing
temperature. They are simple and cheap, but don't always behave
linearly or consistently between devices, particularly at the top of
there temperature range. A thermocouple is a device that generates a
small current based on temperature difference. They tend to be smaller
(quicker reaction) and work to a higher temperature. They are also a
lot more linear and consistent in their response.

In effect thermocouples are a little more complex operationally, but
give far more consistent results at the temperature range we are
working at, with less load on the processor. They also happen to be
pre-assembled into probes already.

[Alan Wood]

On Mon, Oct 10, 2011 at 9:38 PM, cynar <cynarbl...@gmail.com> wrote:

A thermistor is a resister that changes resistance with changing
temperature. They are simple and cheap, but don't always behave
linearly or consistently between devices, particularly at the top of
there temperature range. A thermocouple is a device that generates a
small current based on temperature difference. They tend to be smaller
(quicker reaction) and work to a higher temperature. They are also a
lot more linear and consistent in their response.

In effect thermocouples are a little more complex operationally, but
give far more consistent results at the temperature range we are
working at, with less load on the processor. They also happen to be
pre-assembled into probes already.

As I stated thermocouples are more accurate, the processor overhead is merely a lookup table which the compiler turns into indirect addressing so not really an overhead (apart from the array size which is small) but the table lookup is a piecewise graph that has been balanced with the temperature losses on the extruder and the inaccuracy of the thermistor added together (it doesn't actually measure the molten chamber to be fair).

If we wish to improve accuracy (using a better extruder design, thermocouple in contact with chamber inner boundary) we can add extruder head electronics to adapt from thermistors to thermocouples fairly simply with a small pcb on the extruder mount, populated with something like the following:

LT1025 ~ £6.00

OP07/27 ~ £1.00

T-Type Thermocouple PTFE insulation ~ £5.50

PCB + Passives ~ £2.50

~£15 per extruder or heated-bed as an option for those who want it.

regards

Al