Printing experience: settings, suggestions, etc.

[paddy]

Thank you for the pictures.

Regarding disk, I never used the disk from shop because I never had the "old" HDPE Chassis. I startet Ardumower with an old Tianchen chassis, so I re-used everything expect PCB. Tianchen comes with two cutting motors and small blades (like on common gas mowers). The blades were crap, never got them sharp and loud as hell. So I tried to print some kind of disk to be able to attach the common blades from Automower.

Search here in Forum or on Thingiverse (search here for Ardumower) and you'll find my results. I printed them with PETG and use RAMPA thread inserts. They run without any issue for two seasons. I will re-use the discs or even print a larger version for the new chassis.

If you don't trust them because they spin fast, have a look for Automower spare parts. They use different cutting discs (diameter), and don't cost much (15-20€).

Patrick

 

[romualdas]

Short info regarding aluminum profiles used in the mower.
I've bought 4x500mm 2020 aluminum profiles and at the moment have used only 3 of them.

I've put one on the back of the mower boxes here:

Out of 500mm, there was only 135mm left

The rest 36.5 went into the back side of the motor boxes.
I might make it even shorter, just enough to reach the nuts, but will put through the emergency switch there (if I won't figure how to put it on the top).

The mower box holders are ~98mm, but they don't have to be precise as well.

With the main chassis profile, I went from the motor boxes (the fusion file shows that the extrusion should go into the box) up to he front bracket.
This was approximately 30+cm length:

on the picture is 33 (but it is at an angle), but from those 2x500mm extrusions I've got both main chassis elements and mower motor box supports.

What I did
I've just installed 500mm long extrusions, mounted all parts and then measured how much length is required by the front bracket from the front of the PCB box mounting bracket.

Hope that helps
Attachment: https://forum.ardumower.de/data/media/kunena/attachments/4472/IMG_20190114_225146_LI.jpg/

 

[romualdas]

Some more updates:
- installed button (16mm hole diameter)
- switch (12mm hole diameter)
- and charging port (12mm hole diameter)
on the back of the motor box. I have drilled the holes in the original boxes. To get the button installed, I needed to take some material off from the rear plastic bar that covers aluminum extrusion up to the extrusion in order to screw the ring on the back.
Anyway in the motor boxes there are quite few things that have wires that need to be brought to PCB box: motor power and sensor wires (x2), power wire, button and button led wires, switch wires.
There's only one hole between the motor box and PCB box and it is just enough to push through those wires.
If you would want to put some devices on the back (there is some space there) like GPS, I doubt the wires would go through.
On the other hand, if there's a need, it shouldn't be a problem to drill additional hole between PCB box and motor box.

Regarding mow motors and other sensors that will reside somewhere in the front, I'm not sure. The only hole to get the wires inside is the one on the top of the battery box. I've installed battery box so that the hole is oriented toward motor box, hoping that less dirt will get through it. That hole however is tiny and I could barely get two mower engine power wires. Considering that there will be US sensors, perimeter sensor, bumper sensor(s), etc. All of those will require a way to get them into PCB box.

This is how the wiring looks without PCB box. No wiring for US, perimeter and other sensors is there yet.

When the PCB installed, basically wires from the below can be extracted via gaps on the sides. Not much of space and maybe not very comfortable but it should work.

Attachment: https://forum.ardumower.de/data/media/kunena/attachments/4472/IMG_20190124_233941.jpg/

 

[paddy]

Thanks for sharing your progress.

Mine is mostly complete so I will start assembly this weekend (waiting for some slot nuts).
I was worry if all cables will fit through the hole between PCB- and motorbox.

One thing I wonder is, how to get all the other sensors and Kill-switch connected to PCB. One way might be to use the slots of alu profiles to guide cables. We have to drill holes into the box for sure but I think it could be sealed with hot glue.

Do you plan to cover/coat your chassis, and if so, that kind of material/method to use? There are a few points where dirt and water can reach electronics:
- motor boxes (where they are joined together
- motor box lids
- PCB Box (where they're joined together)
- PCB Box Lid (where they're joined together)

Other parts of chassis will also not be water tight but I assume it will not hurt.

 

[romualdas]

I'm thinking to use epoxy to seal-off gaps and then primer and paint on top of it. Hope that it will be enough for the gaps between part joins.
I'm not worried as much for the join between motor boxes as the join is pretty tight (parts slide into one another).
The biggest concern for me is are the:

PCB box Lid:
a) it is joined of 2 parts
b) the back of it just enough to cover PCB box (sides are over the PCB box)
c) the closing is not tight, so the dirt will definitely get there

Holes in the PCB box and motor box for the wires:
a) there's a gap between the motor boxes and PCB box
b) while hot-gluing the holes is possible, but then the modules are tied-down and changing the wiring layout gets complicated
c) There's a gap between parts, but gap not wide enough to add any type of connector there.

Motor box covers:
a) they have these wells for dampers, so if the water will get on top of motor box, it will get into the "engine room"
b) sealing edges and wells of the cover completely after installing is probably not the best idea, because maintenance might be required.

Getting the wires into PCB box is a tricky problem. There will be quite a few cables to move in: US sensors, perimeter,bumper, charging (from station) etc.
I'm thinking about getting the cables into some sort of nylon sleeve and then zip tying them to the aluminum frame (or just zip-tying without sleeve).
I like the placement of the single hole that I have in original battery box. It is possible to push the wires from outside to PCB box via that hole.
It is in a fairly closed space between the motor box and the battery, so less chances for the water to get in, but the hole is small (I've managed to push through only mow motor cables, you can see those in the picture). And again this will require sealing off the holes with hot glue.
One possibility would be to use DIN connectors (or anything similar), but they would take a lot of space in the PCB box and space there is at premium.
On the other hand, it would be easy to detach if necessary and the box would be sealed from elements.

So far: holes filled with hot glue sound like the simplest and most affordable option

 

[paddy]

My current plan is to use some kind of small box on front bracket (above front wheel) and to collect all cables coming from sonar, perimeter and others here on a small PCB. From here there will be one shielded cable with lots of wires (Lapp Unitronic or similar) which goes into PCB Box. Alternatively, we might use these small ribbon cables https://en.wikipedia.org/wiki/Ribbon_cable.
To protect the tiny cables from sensor to this box, I plan to use small tubes like silicone tubes.

I have some experience in molding 2K silicone. You can print a mold, mix up silicone and ready is your personal sealing. This might be an solution for the rear dampers so no water can go into motor box. We can also build sealing tubes for cables.

 

[paddy]

What about adding a small box under front part of pcb box? I mean like battery box only much smaller. We can place water proof connectors like sub d or those you mentioned. This way, we don't have trouble with space inside pcb box

 

[romualdas]

I agree regarding the box on top of the front bracket. That would somewhat decouple the shell from the PCB box.
Still, there would be need to bring the cable from the box (~13mm hole) and cables from the mow motors (possibly another 8mm hole) to the PCB box.

IMO box needs some place for connectors, because wires from the sensors that are located on the shell, should be attached to the box after the shell is installed in place.
It means that they have to be easily accessible and easy to connect, when the shell is attached.
I've looked at one type of connector available in the local shop:

There are versions from 2 to 8 pins. You can attach the end even without seeing it (one way only) and the end has this screw cap that holds connector to the socket.
Still, we're talking about 12mm hole for each connector and there must be like 34mm free distance on the outside.

Hopefully, they would fit, if holes are made on the sides of the box.
The perimeter sensor itself could be located inside that box. That would leave us with US and bumperduino (if anyone) on the outside.

P.S. have you decided on the bumper sensor yet?
Attachment: https://forum.ardumower.de/data/media/kunena/attachments/4472/IMG_20190127_155337.jpg/

 

[paddy]

I found similar connectors like you but as sealed, waterproof version (Weipu and others). They are used in industry and automotive but they are expensive (~10€). I like the idea to have the connectors outside the PCB Box and to be able to detach them. I'm still wonder if we can find a less expensive solution. AMP Superseal connectors might be great too, but they commonly use 1,5mm² cables which is way off for sonar or bumper

I still have no solution for bumper. I've ordered some linear hall sensors and some switches to start tests. But first of all, I need to verify if my self-molded dampers will work. They should be easily to move, but should not move when driving hills or shake too much.

 

[romualdas]

Some more updates on my build.

As I've already mentioned, there's not much space to move cables between the motor box and PCB box, Originally, there's one 14mm hole on the motor box and 10mm - on the PCB box.
If you want to have more cables going to motor box, i.e. buttons, emergency switch, charging socket and maybe even some sensors (like GPS), it maybe hard to push through all of them.
Also, when they exit in one place, it becomes hard to disperse them and the space on the bottom is required for capacitors and other parts on the bottom of the PCB.

I've decided to make one more path. Just drilled the 10mm hole symmetrical to the original one on the motor box and PCB box.
This will give enough space to run cables more freely and enough space to have all the necessary equipment.

Because I want to use two mow motors and the PCB 1.3 by default doesn't support that, I've slightly rebuilt the circuitry.
As it is discussed here , it is not safe to connect to motors in parallel. So,I've disconnected output M1OUT from controller to PCB and wired it with small PCB with protection circuit (same as on the PCB) to protect controller from currents induced by motor.

The connection looks like this:

Not perfect, but works with my setup.
The protection circuit can be placed under the PCB and it fits, when placed in the same area as battery.
After wiring both motors to the board, it seems to be working. Both mow motors were spinning and if any of them is stuck - the mowing is shut down.
Also, the current/power usage is reported for both motors.
Attachment: https://forum.ardumower.de/data/media/kunena/attachments/4472/IMG_20190130_003702.jpg/

 

[bernard]

With Pcb1.3 ,I have burn a driver motor in the past but it was at pwm 255, very high grass and 2 or 3 stop on sense security and not with motor from shop.
Maybe it is better to do some change inside Azurit to don't stop the mowing motor immediatly, but reduce the speed before (same in Roland Raindancer firmware)
Except in emergency when imu detect a lift of the mower.
Cool the circuit but what is exactly the utility of this protection board ??.
In the past I use pcb1.2 without protection board,and never issue or burn driver.
By.

 

[romualdas]

I agree regarding switching off gradually. It should probably help with the currents.
However I see at least couple of cases, when it won't:
- "emergency" switch off because of mower is lifted
- Mow motor has hit an obstacle and is stuck (e.g. mole rush, fallen tree branch, ...). In this case both code and driver itself can switch off motor in order to protect driver from high current.

From what I have read in the forums (don't have the links to the posts now), there were cases, when the motors were burned with 1.2 without protection.
Also, did some reading here and here (page 2, 3).
The protection board that is made for PCB 1.2 is not compatible with PCB 1.3, but the schematics used in PCB 1.3 is fairly simple.
What I wanted to do is something similar to the drive motor connection, where each motor is fed via protect circuit shown in the PCB 1.3 schematics page 3.
In the end, after the rework, M2OUT works through the circuit that is on board and M1OUT - via the custom made.
This is with the hope that it will protect the driver from the high currents that are flowing back from motor, when it is switched off.
Also, I've enabled the current sense feedback from the motor driver M1FB to Arduino A6 via JP3.
With that, it is technically possible to slow down (reduce PWM), when current on ANY mow motor increases, because the motor gets stuck or hits something and has trouble running.

A side note regarding currents on the driver. I noticed that the measured current is different for each motor even though the motors are the same.
So, probably it is worth to do some calibration on the motor driver feedback input. Just to make sure that it is not running at a higher currents than it is supposed to run.
Last year I was running my mower with both mow and drive motor drivers set to max 255 PWM (single mow motor though).
With the heatsinks attached to driver chips, I haven't experienced any issues except one: from time to time - mower would switch off the drive motors after hitting obstacle.
Interestingly enough, it would still show the state "FORWARD", but only switching mower off/on would help.
AFAIK driver itself has a feature, when it turns off the motors after the over-current is detected.
I'm thinking if this turning off during the peak currents flowing through the driver/engine could still cause driver to blow or at least enter faulty state.
I guess reducing the PWM value is a thing to consider as well.

 

[romualdas]

Some updates here

Here's my take on where to put the GPS module on the printed chassis:

There's one thing that has to be addressed - 5 pin connector won't fit through any of the holes. And even 4 pin connector won't fit through PCB box hole as it is smaller (it fits through motor box hole).

Of course, it is possible to drag the cable first and then crimp one of the connectors, but then, it would be impossible to completely disconnect PCB box from motor box without the need to break the cable. Other option is to have some connectors like CH4 to splice wires from both ends in the PCB box.

Also, here's my take on Patricks version Hall bumper sensor .

I'm not as good at soldering, but somehow it works. I've used US2881EUA and even though they have minimum supply voltage of 3.5, they seem to work with 3.3 just fine. Based on the magnet polarity the output can be set to LOW or HIGH in normal state.
Attachment: https://forum.ardumower.de/data/media/kunena/attachments/4472/IMG_20190223_234628_HDR.jpg/

 

[paddy]

Hi,

nice to see hall sensor PCB does work for you too. I still have no enclosure for the sensor, how do you plan to attach them?

Cable management is a weak point on this chassis. For all cables coming out of PCB Box (Sonar, Perimeter, Mow Motors, Charge contacts) I use these things https://www.conrad.de/de/kabelverschraubungen-o0809042.html to get the cables out. I just drilled some holes at bottom of PCB Box (on the front part) and use these cables https://www.xbay.de/itm/Wasserdicht...g-wasserdicht-Auto-Motorrad-Boot/352155799306 (ebay link doesn't work in forum but you should be able to oversome this issue)
For charging contact, I use AMP Superseal connectors because they can handle higher current.

This way, I get a (maybe) water proof cable connection and can still detach upper chassis from lower body.

 

[romualdas]

Regarding the sensor attachment - I don't know yet. Most probably I'll wait for the sensor enclosure you're developing.
Another option, would be to put them on the back of the mower, but as you've mentioned - it might be not good enough to detect collisions when mower turns (it shouldn't be a bit problem in my case environment though).

Regarding the wiring, I'd like to see how your approach looks like.
In my head, I'd like to avoid any holes downwards. The holes in that case need to be 100% water tight as the water from grass may still enter into PCB box.
Also, the grass clippings will definitely stick to any parts and make it hard to clean/unscrew them if necessary for maintenance.
I like the idea of having holes in the back of PCB box and then, sealing them with either epoxy, hot glue or silicon as there's not much space between motor box and PCB box.
That means, however that in case of a maintenance, the holes must be resealed.

 

[paddy]

I agree with you, holes downwards are not great at all, but the screws have sealing and I will add a small cover to protect it further (and to ease cleaning). Maybe next time I will use D-Sub connectors with lots of pins (you can purchase sealed one) to have just one cable going out of PCB Box. Then build a small box on top of front bracket to split to sensors. Maybe it can be solved without this additional box but I don't know how to seal an cable when you split it into individual, smaller ones (imagine you have a cable with 20 wires soming out and you need to split them like 3x4 wires for sonar, 1x3 wires for perimeter, 1x4 wires for bumper and so on)

The current design has some drawbacks in my opinion but I will not be able to fix them this season. So currently I just try to bring it alive and observe it during summer. After season, I will need to re-print some parts.

Weak points I see:
- cable management (cables between motor box and PCB Box, cables for sensors)
- Motor box lid (traction motor), water can enter easily, it is attached with tiny screws and might break easily,
- nose need re-work to add clearance for bulk wheel and maybe it is better to use two wheels at front
- PCB Box is not water tight (where front and back part gets screwed together)
- hinges at PCB Box are weak
- PCB Box lid need rework (cable management, mounting points to assemble sensors like compass or PI Zero
- not much clearance between blade protection and ground
- max cutting height is approx 5cm
- dirt everywhere

Do you see additional points? Let us collect them and see how to solve for next season.

In my opinion we shouldn't stick to these kind of Alu profiles. Instead, we should use 20x20 squared pipe profiles (I mean these ones https://metall-fachhandel.de/produkt/aluminium-vierkantrohre/) and guide cables inside.

 

[romualdas]

Perfect list IMO. I would add just one more thing:

- Almost no space/place to put additional modules (e.g. GPS, IMU, etc.)

My current;y planned layout for the extra modules that I'm using:

- Bluetooth module barely fits, when placed at an angle;
- For now, I've decided to move GPS to the motor box;
- I've placed IMU right on the PCB box;
- Custom Motor driver protection board for the second mowing motor seems to be fitting under the PCB, just enough

The rest of the sensors (perimeter, ultrasonic) are outside anyway, so it comes to the cable management issue, where IMO managing cables inside the PCB box is complicated too.

 

[jussip]

Also, I've enabled the current sense feedback from the motor driver M1FB to Arduino A6 via JP3.
With that, it is technically possible to slow down (reduce PWM), when current on ANY mow motor increases, because the motor gets stuck or hits something and has trouble running.

Did you add new sensor to code "SEN_MOTOR_MOW2" or something like that? Or is there code already for second mow motor?

 

[romualdas]

I've found some code that was commented out, but not enough to handle everything.
I've separated the changes that I've made in separate commit here:
https://github.com/rstonkus/ardumower/commit/8a82400d82b51a6a3f42e0032fe982fc7794871a

 

[jussip]

Thanks, I wrote little more code to pfod, ros, rmcs and consoleui I think.. Now it compiles without errors and theres both motors showing in arduremote, calibration works also I think..