Gen7T or Toshiba drivers for Gen7?

Quote

I'm wondering if I should add the schottky diode in anyway just for kicks.

It reduces ATmega's supply voltage by 0.3 volts, so it might even do the opposite of what you want. The answer is "It depends ..." and if this answer comes up I usually decide for the simpler, leaner way.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Since I'm a pc tech and I know a heck of a lot more about the atx standard vs the other things in electronics, I'd say the 5Vsb has a tendency in all the psu's I've tested to stay fairly stable around 5V (4.75 - 5.25V is the listed 5% tolerance for the atx spec). PC PSU's are not the most stable of power delivery systems this is why most motherboards there days generate their cpu voltage from a 12V line and the voltage regulation is designed to deal with tolerances by basing the regulation not on the supply voltage but a fixed onboard reference. i.e they take the input they use a small circuit that come hell or highwater will produce the same voltage no matter the input and they go base it from that.
if the people use a fairly decent PSU and you use the Schottky diode it could keep the voltage quite stable I mean if the supply on average is from 4.9 to 5.1 and just dips during motor and heated bed starts the cap should keep the 644P running for quite a bit longer than an ATX12V PSU should need to get back to spec.

Note on Schottky diodes: One thing I noticed when I was using them when I still had a home made solar panel on the roof (as in eBay cells) the larger ones have a tendency (3A plus rated) to only have a forward voltage drop of 0.2V as long as the current is not towards the max cap. but your gonna have to dig through the data sheets to find a suitable one might have to go a bit higher than the 300 mA max were thinking it might use to get an optimal voltage drop for our application.

Just my $0.02
terramir

Well I guess I'll get it all running again now that the mods are about done and do some long prints to test stuff out. If I get comms errors or something I'll try adding the diode.

check my video - sanguinololu with tb6560ahq

[www.youtube.com]

@ codec cool. I love when things start to actually move.

@ everybody else. I'm back to 90%. After the ground fix I got stuff hooked back up and tried printing a bit. Everything seems fine except now I have a strange Z creep. It's obvious when the motors are enabled but not moving. The Z axis will step randomly every few seconds. So I'm in the middle of troubleshooting again.

Tried
desolder and resolder all control lines to Z driver
Use a different cable adapter for the stepper (I have dissimilar plugs on the steppers vs the board. )
run an exacto knife between all related pins to double check for shorts/bridges
Power from USB rather than ATX

When I torque on the Z axis manually while this is going on it seems like the motor is being disabled before it moves but I've checked that connection multiple times and it's fine. Plus the other three motors are solid, and enable is tied across all the drivers.

I'm starting to wonder if I did something naughty by tying enable and reset together, reset is technically supposed to go first but it was only a clock cycle or so, so I figured it might not hurt to do them at the same time. But maybe that's my issue. I'm hoping that's not the case because it would be a pain to fix on my beta board. Any other ideas before I go buy a cutting wheel and do some surgery? I'd hate to do all that work and then find out it was something simpler.

For signals it's pretty simple to delay them:

Signal --100k--+--|z>--
               |
             100nF
               |
              GND

I've read delay time is about t = R * C, so 100e3 * 100e-9 = 0.01 seconds in the drawn case.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Sure, but I'd still have to do surgery so I might as well just separate the signals rather than combine them with a delay. I'll probably end up tying reset to the 5V stby line.

OK, back to 99%. Separated the reset and enable lines, tied reset to 5V stby and wired enable back to the enable pin on the atmega, now the motors are rock solid. So I've got a bit of rerouting to do and then I think it will be ready for beta testers.

Edited 1 time(s). Last edit at 03/13/2012 02:19AM by bryanandaimee.

OK, I think I have all the above mentioned issues fixed on the beta board. I'm going to call it my 0.1 beta release and encourage the adventurous among you to try it out. (You're on your own for PCB creation.) I'll likely get some boards made later for a more significant beta test group, but for now I'll be testing more thoroughly with my own board.

@Traumflug could you take a look and see if it is OK for routing. If it looks like it will work PM me with a price and I'll get a couple from you to test alongside my toner transfer boards. If there are problems let me know and I'll fix them so it will route.

Files and link to the mouser project for parts procurement are up now on the Gen7T page.

bryanandaimee actualy I test TB6560 with Sanguinololu board. I have working, small TB6560 boards, and if everything will work fine, I will start to draw nu pcb (without led's, few resistors, unused jumpers etc)

[www.youtube.com]

Two days ago I actually milled bryandandaimee's Gen7T already but didn't find the time to post about this.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Looks nice, I can't wait to test one out.

Well I've performed my first driver replacement on the beta board $4.20 . Driver replacement was a bit of a pain as you can imagine and I'm starting to think socketing the driver chips might be a good idea except that female pin headers are so darn expensive it would add between $4 and $10 to the cost of materials (25x4 = 100 pins).

My beta board is getting to be a bit of a hairy mess with all the surgery lately, but I think it's about there and I should soon have some boards from Traumflug to test that incorporate all the recent changes plus possibly some commercial boards soon once I get brave enough to send some out for public beta testing.

Also I tested repetier firmware out and it looks like the step pulse width is a bit too short. These drivers need a longer step pulse than pololus do. At least 10us. So back to teacup for some more testing.

Love your design cant wait to build one myself if you end up assembling kits let me know ill buy a beta kit from ya if it dont get around to milling my own board and buying the parts myself first.

As for the female headers why dont you use some of these 20 or 40 pin headers and cut them as required.
[www.taydaelectronics.com] this place has been a great source for my electronic components. The pin spacing is a little off 2.54 vs 2.76 so but you should be able to squeeze that chip in At 13pins per row the maximum mismatch would be about (2.76-2.54)*13 / 2 (assumes center is alligned and the ends are out) so about 1.43mm or roughly half a pin. I think you should be able to just bend the pin a bit to compensate. .

The other alternaive is a riser board to go from 2.76 to 2.54 but thats a lot of extra holes to drill. You could also break the the headers every 6/7 pins and leave a little extra gap to allow for the mismatch making it only a .71 mm off on each side. Its not a superbly pretty solution but it will get the job done on a beta board.

Also found this topic out there on similar issues people have had implementing the Toshiba chip and meeting all their requirements: [dics.voicecontrol.ro]

Also another link with a longer discussion of the power cycling shortcomings of this chipset.
[www.cnczone.com]

Edited 4 time(s). Last edit at 04/08/2012 05:13AM by hotdrop.

Hi All,

I am really loving all this Gen7T progress.

Even if it turns out to be a bit more parts cost than an all in one mass produced SMD controller it will still be a lot more fun for us people who are out of USA/EUR and have a bit of extra time to solder their own without having to first bootstrap a toaster oven.

I was looking at the pin out and the problem with the close spaced pins is not new to industry.

Firstly one could cut any unused pins off with impunity. Then the power and motor pins could be staggered further away from the line of the other pins to grow the pad and track widths and still allow for the minimum track spacing. Adjacent low current pads could be thinned out to reasonable minimums.

The links below (mentioned here before I think) discusses many of the points I went and added afterwards including a brute force remedy to have MOSFET switching on the motor supply. In the case of the Gen7T it could be done under firmware control and be linked to an eStop circuit that will power the motors down and only power them up after all is well.

[dics.voicecontrol.ro]

[www.cnczone.com]

The requirement for using an ATX PSU is obvious to those of us who have old PCs lying around. However there may be a lot of potential users who would not have them. Specifying the board so it can run on just a single 10-24V supply for everything would be a good choice in my book. Having just solar or wind power in 12V storage batteries would make supplying an ATX PSU inefficient. I think the Gen7 through hole designs should give that extra little bit of thought put intoout in the woods operation as well as component sourcing and PCB manufacture.

As for the power supply issues with the Toshiba drivers perhaps a note or call to a larger Toshiba office with advice might elicit some gems of wisdom. Tech support sometimes lets good ideas slip. I think relying on any peculiar ATX or USB quirks or functionality for stable power operation is asking for a very hard to support design. The on board regulation of power seems the safest way to go if the drivers will tollerate the logic supply coming up with the motor supply. Some designs (unrelated) I have seen in the past have required rather specifis supply voltage presentation and this is always a pain. Linear Technologies and similar supply low-drop out regulators that can give 5V with very low loss of voltage. This allows the logic power to track the main supply pertty well and lagging by only the drop out voltage during the early supply ramp up. Some devices prefer a slow ramp up (so all caps charge evenly) and others want a fast ramp so reset circuits do not time out before final voltages are reached. A similar arrangement may be possible here (like the Schottky diode ideas) for the motor supply with a big cap after a diode (and resistor) for the motor supply and then a small cap and a LDO regulator to get the logic supply ramping up before the motor supply reches the logic level.

The ground loop issues could be called inevitable and it is good they came out now and are correctly diagnosed and removed before any production boards are made. Ir is also quite acceptable to add top side power bus connections to single sided boards to route power to where it has to go, these are also useful for diagnostic purposes during first testing as they can be left till last and sections of the boards powered up in stages or disabled later for debugging.

Lots of waffle but I do hope this all works out. I would like to try out one of these before long as I am starting to assemble my RepStrap and also have a issue with the SMD stepper drivers.

Regards

Kalle
--
Johannesburg, South Africa

@ hotdrop Thanks for the link to headers. The cheapest I had found to now was about 3 or 4 cents per pin. At 24 cents per 40 pin (and decent shipping from Thailand no less) I think socketing the drivers is a bit of a no brainer.

I hadn't even thought about pin spacing mismatch. I didn't notice. I did socket one of the drivers on my alpha board and it fit just fine with .1 inch female headers installed, so I don't think it will be a problem. And now that I think about it the footprint on the board is .1 inch spacing, so I guess I'm bending the pins slightly anyway.

@KalleP Thanks for the input. From the links you guys provided I'm thinking I could add a FET to isolate the motor supply till the logic supply was up. For back EMF issues when turning the FET off it seems that having the capacitors on the motor side of the FET would help with that. On the other hand just using an ATX supply handles this issue and this is really a pretty ATX centric design. I'm worried that adding the FET might add back EMF issues that will make things worse. I'm not sure on this point.

It may be smarter to make a non ATX design separate from this one. I've been thinking that a sanguinololu derivative would work well for those using the 12V LED driver power supplies. It seems a bit wasteful to spend all that money on ATX connectors and then not use an ATX supply. Of course for the self source crowd that's not an issue as they can just leave them out of the build, but for kit built electronics (likely the majority ?) you could save quite a bit with a purpose built design.

cutting unused pins on the drivers would only really help for one supply line, and it's not really the problematic trace anyway. Not sure what you mean by staggering. Bend the pins out from the center line? Of so, that would seem to make it more difficult to socket but might be worth it. On the other hand, the worst trace is the one that goes between the motor connector pins so changes to the driver footprint wouldn't help much there. So maybe another jumper like you suggested. That does increase the build time though. So as long as the routed boards from Traumflug work OK I'm thinking I'll leave that part as is. My kits will likely be double sided commercially etched boards anyway. So I think it will be fine.

Thanks for all the input. I'm no electrical engineer, just a hobbyist, so any help is greatly appreciated.

I was notified about a simple enhancement to make Pololus fit very low resistance motors better (mine have 1.53 ohms only) and seeing this, it reminded me the TB6560 might benefit from tuning. For example, there are different decay modes and glancing over the data sheet didn't make it obvious to me, which decay mode to choose.

[hydraraptor.blogspot.de]
[hydraraptor.blogspot.co.uk]
[hydraraptor.blogspot.de]

Thanks to nophead for these excellent blog posts.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Yes, and for the initial boards I didn't even include jumpers to set decay modes because I didn't know how to tell what decay mode to set it at. And because the board was getting a bit crowded anyway. I should probably add those in so people can experiment/adjust for their motors.

Interesting posts - I use these Toshiba chips in my home made CNC machine (eBay boards) with no issues.
I have a suggestion for the FET driving issue - why not drive an opto-coupler from the ATMEGA and source the opto transistor from 12V so you have plenty of FET Gate voltage and signal isolation as well?
I just ordered a Stepstick board as I was going to build a 1.4 board once you have finalised the design but the Toshiba chips have no problem driving my current Nema23 motors and I would like a little headroom.
Once I am confident in the electronics and software chain I will start my Reprap build (sort of like the Makerbot Replicator style).
I have prototyped a heater head using muffler putty/ gasket cement and a theaded rod and it seems to work but as it's not in a machine yet who knows?
What is the best heated extruder design? (I can't Reprap parts yet but have a small lathe and the CNC mill).
Love this whole idea and can't wait to get one built.
Regards
Denis

Got three routed boards from Traumflug. (Thanks a bunch!) so some more testing is coming soon. I think once I try out the first routed one I'll build the other two and find some beta testers, and then maybe get some kits together for a more serious beta test group. So progress continues, even if it's not at the speed of net.

Me me me wanna betatest
wanted to actually ask you how it's going right about now since I fried 4 of 7 (total ten I have but untill yesterday the8th was not soldered) pololus I h8 those things
so I wanted to know the progress of this because I would like a board like this myself . Ya know it's cheaper to replace these well and I found out that you can get these on eBay from china for like 3.30 ea including shipping or if ya dun want to wait there availible from mouser for like 3.36 ea plus shipping if you buy 10
I am so fed up with the pololu's
terramir

I read this thread on cnczone about chips thb6040 that didnt suffer power on/off sequence problem, but the chips it self cost $20 (aliexpress.com) hmm no feasible but dunno thb6128 ($4) can do the same as thb6064. Maybe this chips become another option to tb6560 or allegro or L297/8.

Edited 2 time(s). Last edit at 05/10/2012 06:28AM by dida_r.

Haven't been around for a week or so because I've been studying for my medical physics boards. I'll get back to this stuff in a week or two after I'm finished with my exam.

Hi everyone! Thought I'd throw my 2 cents in as well since this Toshiba chip looks to be the next evolution of the RepRap electronics

@hotdrop Isn't the pin spacing 2.54mm? The data sheet suggests the distance from the edge of the chip to the first pin is 2.76, then the pin spacing is 1.27 to the next pin but in the next row. (See picture)

@bryanandaimee I have a number of good quotes from suppliers, and I can get the parts for this very cheaply, but since it is primarily your effort that has gone into this project, I thought I should check with you first if you would be ok with me distributing these (and I'm not sure how the CC license applies if I'm doing it as non-profit anyway). As a guide, in terms of raw parts cost they would total less than half of Mousers cost (about $25 actually), with the catch being they need to be bought in rather large quantities to get this cheap. But this is including ATMega and all 4 TB6560s so not too bad. Also, in the Mouser project you've got 100pF caps and no 10uF caps, were they a straight swap and which are you using in your test boards?

@everyone Anyone looking to buy this from the mouser project should make sure they have the right capacitors for the Temperature sensors, as these appear to be missing/swapped from the Mouser project. I've also made an Eagle version of bryanandaimee's schematic but with a slightly different version of the PCB for those that want it (its attached). Any suggestions/constructive criticism are more than welcome.

Someone who knows the wiki: Should I be putting this stuff up on the Gen7T page, starting my own page, or neither?

Looks like some great development so far, looking forward to what this will become

Chris

Which "mouser project" are you talking about and since when does Gen7 use Eagle?

For the current state, it probably doesn't matter wether the chip costs $2 or $5. The next challenge is to get this chip running reliable and find a proper configuration for RepRap uses. For example, some people talk about the TB6560 having a pretty limited step rate, like 15 kHz.

"It works for me" isn't sufficient, it should work for everybody ;-)

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Re: Wiki, If the layout and BOM are the same, such that the build instructions and opperating instructions would be the same, I wouldn't mind the help in keeping the Gen7T wiki current. If your eagle version has swapped jumpers or anything that might confuse builders then you might want a different page to keep things clear. You can make as many wiki pages as you like.

Re: 10 uf caps. Yes, they are missing, I'm still finding mistakes in the mouser project every once in a while. I'll fix that. (Traumflug: I posted a Mouser project on the Gen7T page as an initial BOM untill I can get a proper BOM written into the wiki)

Re: Spacing, Pin to pin spacing the long way seems fine to me, I have discovered that the footprint I'm using is a little close between the two rows such that using sockets is a bit of a wrestle. The two rows of pin sockets press against each other and they have to be forced together before soldering.

Re: Distributing, I was never really planning on becoming a large scale distributor of kits. If you feel like you want to jump in I'm fine with that, as long as you realize (and inform your customers) that this is a beta board. AFAIK only two boards have been made and tested. Both are in my basement right now. I'm half way through three more that will constitute the first public beta test set.

Re: License, My understanding is that the Open Source licenses technically apply to the plans only, so restrictions for NC licenses etc. don't apply to actual boards made using said plans. I know there is some debate here.

I don't know what your timeline/plan might be but I would suggest something like this:

1. add in decay mode jumpers ( It's on my to do list)
2. possibly update the atmega side of the layout to match with the new version of Gen7 (Kill 3 of the end stop connectors, add in the socket for daughter boards, etc. )
3. Build a few boards and test
4. Document, Document, Document
5. Sell to your hearts content

Quote

5. Sell to your hearts content

Bryan, as you probably can imagine, I'm not exactly enthusiastic about such a recommendation. What you did is copy a design under a NC license and remove the copyright strings. This isn't accepted practice anywhere, not even under the GPL. So far I considered this to be an accidental slip, as the NC license string is still in place, on the board's design file as well as on the wiki page.

Making a commercial board from a plan is a commercial use of this plan, so the license applies to manufacturing, too. If it weren't, one could make copies of music CDs without ever worrying about copyright.

If you want to have a design under a different license, please do your own design and give it a sufficiently distinct name. This also excludes such actions:

Quote

update the atmega side of the layout to match with the new version of Gen7

That said, commercial distribution of standard Gen7 is forthcoming. Next to paoparts.com (and me myself, of course), two additional shops have accepted to share a few euros/dollars per sold Gen7. Even if this would spread to other parts of a printer, I'd see no problem if a $600 kit becomes a $700 kit and these additional $100 were split up among the participating RepRap developers. It isn't like shop owners have to pay that from their own pocket.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Quote

If it weren't, one could make copies of music CDs without ever worrying about copyright.

Works of art are subject to copyright, functional designs are not. I don't know if a PCB design could be described as a work of art even though it is often called artwork, especially if it voronoi as it can be mechanically generated from the schematic and the placement, which would be hard to call art.

---

[www.hydraraptor.blogspot.com]

Sorry, maybe I should have been a bit less bold. I should have said something more like

5. Work out some mutually agreeable arangement with the designers. Then sell to your hearts content.

I have tried to do my homework on the subject of licenses and my understanding is similar to Nophead's. Patents are for functional designs, Trademarks are for names like Gen7, and Copyright is for plans, software, etc. And while many open hardware licenses are trying to stretch copyright law to cover functional designs I just don't see good information that shows these types of licenses actually affording the protection you are talking about.

If that were true then Intel would no longer need to go through the expensive patent process for new chip designs. I'm sure many companies would jump on the NC copyright license if it actually allowed them to skip the patent process and still legally require others to pay royalty for producing their design.

I think the main protection that open source licenses afford is from companies patenting your design or a derivative and making it closed source/commercial. At least that is my understanding of the original intent.

That being said I'm sure all involved are happy to work out something mutually agreeable. What would you suggest as a royalty structure? Would the manufacturers of Gen7T be requested to pay you, or me or both?

Thanks for the opinions on copyright. It's true, this stuff isn't enforceable in court, so trust in people playing nice is involved.

Quote

Would the manufacturers of Gen7T be requested to pay you, or me or both?

Good question. Both, probably, while one of us can forward the share to the other, so a vendor needs to do only one payment.

And yes, I'd love to see a better solution. So far the discussion on how to reward developers in the RepRap community doesn't really happen, resulting in lots of wheel-reinventing (to open up a new "brand") and crappy, unsupported stuff being sold.

A simple alternative solution would be to make a SanguinololuT, as the Sanguinololu PCB is better suited for industry fabrication anyways.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY