NanoEFI September 2020 Updates: More Funding, Harness and Wire Equipment

Hi Everyone! :wave:

Since our last update in July, I’ve kept up the focus on building funding and the equipment necessary to support the project through BETA. Lots of ducks to get in a row, both business and personal.

As it stands, I have a couple more (outside) jobs to wrap up. Then I expect we’ll have built enough of a runway to allow up to three or four months of dedicated full-time effort working on NanoEFI. If all goes well, stays on budget (and I manage to get enough sleep in the process) we should see the first test kits become available before the end of the year. Still, lots to do before then :crossed_fingers:

At the moment we’re doing well on budget. Before I go any further, let’s give our Patreon contributors a massive THANK YOU :clap: for their continued support. In fact, an important piece of equipment we’ll need for BETA arrived just a few days ago!

Here’s our latest score! I present to you our new automated wire cutter and stripper! :tada: :confetti_ball:

s-l1600

Kickin’ tires and Cuttin’ Wires :us:
One of my favorite past gigs was the design and production of motorsports harnesses. Mostly engine swap stuff. Here’s an example of my work :heart:

This was by hand and a lot went into doing it right. Since then I’ve still been drooling over wire processing machines, and I’m absolutely thrilled at the idea of speeding up some of the more repetitive aspects of harness building.

Running a tight ship :ship:
The harness is a major chunk of the cost of our EFI kits, and a risk for the project if not done right. Rather than outsourcing our harnesses starting out, I’ll be assembling harnesses myself to ensure that quality control is :100:, and that the cost :moneybag: isn’t inflated on this relatively small BETA run. Also, it’s important that our limited funds aren’t risked on what could very likely end up being a bad first batch if the new design is outsourced right away.

Besides, personally I find there to be a therapeutic quality to wiring and I’m looking forward to hopping back into it for while. To that end, and to make sure we get enough kits produced for you all while maintaining a sustainable level of effort for me, it’s time to introduce you to our latest hired help :robot:

Initial Impressions

This thing shoots out wire fast! On shorter lengths, press start and you’ll have a hand full of completed pieces stripped and shot out in what seems like an instant :exploding_head:… Accuracy and overall consistency appears to be on point. It’s built like a tank and looks like it’ll take care of us for many years.

However, there are some concerns. The roller movements have no ease-in effect. So the wire spool often sees a lot of force abruptly, almost to the point of feeling like it’ll yank the spool off the wall before it has a chance to overcome inertia to begin rotating. I’ll need to create a buffered pulley system with some shock absorption to keep all the whirly bits happy long term.

Also, the keypad+LCD interface isn’t as intuitive as it really could be. Actually… I’d say that it’s cripplingly bad. Even more so with the limited instructions provided. I’m doing well figuring out how the machine operates through a lot of trial and error. Still though (and despite the rough edges), the absolute speed and consistency is already more than making up for the learning curve.

The manual did suggest there to be an RS-485 interface for issuing instructions to the machine from an external source. This is great, but only three instructions were documented: STOP, E-STOP, and what I think is START, but nothing on program entry or selection :man_facepalming:… I’ll see about finding or reverse-engineering the rest of the list one way or another. This would go a long way to making model-specific harness variations doable. Allowing for a separate visual GUI to help take care of calculations and entering programs, well… programmatically. Very similar to Airtune or Octoprint.

I’ll have more to show soon, stay tuned. Thanks everyone! :v:

3 Likes

Wow, awesome new toy you have there, didn’t know such a thing existed.
As for reducing the initial shock on the rolls of wire, first thing that came to mind was having an idler pulley, like the auto tension type on your car engine, but that may end up needing individual pulleys and a tensioner on each roll to prevent overrun.
Being an automotive elec, your previous work is exceptional, and seeing quality like that is reassuring for the installer and customer alike.

1 Like

@alby1976 beat me to praise. Well stated.
If you manage to figure out some commands on the rs485, you could just add a motor to the spool itself to give it a push on start.
What’s the model of the new gear? Perhaps we can crowdsource some more serial documentation.

1 Like

Thank you!! :pray:

That’d be great, I’d love the help. The model is SWT508-SD, branded under several names. I’ll be digging into it over the weekend, finding a PDF of the original Chinese manual is probably the best bet. There may be additional information that wasn’t included in the English translated version.

I’m thinking an array of three printed idlers loaded with soft extension springs - between the spool and machine. They’d be positioned in a cage, the center idler opposing the other two. As the machine pulls the wire taught, all three idlers are pulled towards the center. Dovetail rails printed into the cage should hold up well enough for now to keep the costs down. We’ll see!

Or perhaps a strain gauge measuring tension on the wire may do the job. A low KV brushless motor and nice fat retired server power supply should provide enough boost to take the edge off. This would probably be Plan C.

I’m thinking Plan B would be writing a serial scanner to hit the RS-485 port with incrementing command codes. Send “0x09” and see if it ACKs, then “0x0A”, etc. We’ll just hope there isn’t a hidden self-destruct code :sweat_smile:

Here’s our teaser from the manual.

Does the controller not have a setting for ramp time? They are PWM controlled servos aren’t they?
Basically the same as a 3d printer.

That controller seems to be closed loop. I don’t think there is a way to have a complete external I/o. perhaps find another controller that still lets you “copy” the source coding but gives you more I/o to control.
you can make a leading assembly that’s spring loaded but you would have to create a brake or a friction clutch for the wire spools themselves.
Without keeping tension on the wire spool the entire strand has nothing holding it back. I work on conveyers and sheet metal strip lines. We apply a kind of riding air brake to the beginning arbor to maintain tension until the next set of powered rollers.

No such setting unfortunately. They’re almost surely big heavy steppers rather than servos. I probably won’t crack into the motoring side of this egg. It works well as it is, just needs a little buffering upstream.

If it turns out that the serial interface really is limited to only three commands, tapping into the keypad directly with an ESP32 would be the next lowest hanging fruit. I think “WiFi Everything” is going to become my slogan at some point :smile:

I did some digging and i was unable to come up with a manual that even had as much as you’ve got pictures of. That is some pretty serious engrish though.

Looks like that might be the best bet. The text seems to indicate that there’s an ack of some sort via the screen.

01,66,99 = E-stop
01,77,88 = stop?
01,88,77 = stop also?

It says 3 bytes, if that’s not lost in translation, that means the representations here are in decimal (confusingly…)
That’ll make the hex codes
01,42,63
01,4D,58
01,58,4D

Not especially more illuminating.
Binary pattern is:
00000001,01000010,01100011
00000001,01001101,01011000
00000001,01011000,01001101

Of course, none of this will be a surprise to you, @TravisNano, but perhaps it’ll help some others make sense of it and spark some inspiration.

My only other thought is:
Possible that these are front panel key codes corresponding to the physical buttons? This seems unnecessarily complex, but perhaps the front panel is a serial HMI which could be listened to over its communication to the main board to decipher control codes.

Probably not, but we’re all shooting in the dark at the moment.