[Vidar]

Attachment 364910 I am about to install 3 NPN, normally open proximity sensors using the adaptors I just milled om my X6-2200L. Since the proximity switches send out the VCC (6-36V) on the ouput I will need to pull down the voltage with pull-down resistors to match the input on the USB board.

I plan to use the same connector as used for the old microswitches and just rewire a bit and provide voltage from a small powersupply through the plug. The power supply for the sensors will be fitted in the cabinet. Hopefully I can make space for it there, as it seems like there is plenty of room. The input to the old microswitches lives at 3.65 V, provided by the USB board and shorted to USB-ground when activated. I will need to pull down the voltage on the output of the proximity sensors to a reasonable level to not kill the USB input ports. Would it be OK to pull the voltage down to 5V, or do I need to pull it all the way down to 3.65?

[extent]

that doesn't make any sense, a pulldown pulls to ground, not to 3.3v or 5.v. On the controller, the limit switches are likely wired directly to the input pin on the microcontroller with a 3.3v pullup and I wouldn't bet as to if it's 5v tolerant or not. What I did for level matching is put the sensor behind another transistor. The transistor pulls the sensor line low when activated, and you can adjust the load resistor to whatever voltage your sensor is.

[A_Camera]

Attachment 364910 I am about to install 3 NPN, normally open proximity sensors using the adaptors I just milled om my X6-2200L. Since the proximity switches send out the VCC (6-36V) on the ouput I will need to pull down the voltage with pull-down resistors to match the input on the USB board.

I plan to use the same connector as used for the old microswitches and just rewire a bit and provide voltage from a small powersupply through the plug. The power supply for the sensors will be fitted in the cabinet. Hopefully I can make space for it there, as it seems like there is plenty of room. The input to the old microswitches lives at 3.65 V, provided by the USB board and shorted to USB-ground when activated. I will need to pull down the voltage on the output of the proximity sensors to a reasonable level to not kill the USB input ports. Would it be OK to pull the voltage down to 5V, or do I need to pull it all the way down to 3.65?

Proximity switches should NOT "send out" any voltage at all, but they need a supply voltage to function. Are you sure you are not confusing the voltage with the supply voltage?

Anyway, I don't understand why people use proximity switches for limit. Good old NC mechanical switches ALWAYS work 100% reliably. Proximity switches on the other hand... you never know until you need them. So many things can go wrong with them...

[dharmic]

Level conversion, not pull down.

Just dragged my YongNuo bob off the shelf for a closer look. Inputs go through a choke (EMI suppression), then a reversed M7 rectifier diode, then straight into the cathode of a Sanyo PC817 optocoupler input diode whose anode goes to the rail via a 560Ω resistor.

The reverse voltage on the M7 means that, for a "high" condition, you can bang a thousand volts into the input and the opto won't be touched. I wouldn't though, the board isn't laid out for HV isolation! But it's fine for any low voltage (6, 12, 24, 32, 48, 60, 70V etc).

To reliably trigger the opto, you'll need to get current flow of a few mA. From the 5V rail, we lose 1.2V typical forward voltage in the opto and another 0.8 through the M7 so only have 3V to play with. That 560Ω resistor means 5mA when the output is shorted to ground, which doesn't leave a great deal of room at the bottom - you're going to want to pull down the output fairly hard when it's not active. But a 100Ω resistor will run at 1.5W on a 12V high input so it's going to be hot.

The YongNuo inputs are designed to be triggered by open collector, so the other option (and probably the better one) is to insert an active component eg npn or n-channel FET inverter to drive the YongNuo input low when active, ie be that open collector.

tl;dr?

The higher voltage isn't going to be a problem, the problem will be getting the card to register an "off" condition.

Two options:
1. Simples.
- Use 12V power supply for the sensor.
- Plug the output of the prox sensor straight into the input of the card, and run a 560Ω 1/4W resistor from the input to ground.
- If this doesn't work, try a 200Ω 1W and then a 100Ω 2W resistor, noting that these may get quite warm when the prox sensor output is active.

If this doesn't work,
2. Proper.
- Buy or build an open collector circuit to drive the input.

[madeinoz67]

Many thanks,

looks like a reasonably simple solution I that I will adopt myself.

Thanks again for sharing

cheers, it's going to be quite a long build that one! Generally fairly happy with what has been done so far though.

it is, if memory serves, a 12mm slab which initially I just temporarily fixed to the bed before then using my machine to drill and counterbore some m6 bolt holes that align with the extrusion channels. I then manually added another row behind the reach of the machine.

I then just used t-nuts as usual and bolted it semi-permanently to the extrusion bed. Having sunk the screws down to a good depth, you are then free to take a nice large bit (I used 12mm) and pocket out a bed as big as the machine can manage. I then chose to mount a ground aluminium tooling plate as my bed surface, but that part really depends on what you want to do. For the new machine I'm tempted to do similar but have a sacrificial slab of SRBP as the replaceable main bed rather than a tooling plate. I don't currently do as much vice work as I thought I would.

[shane-w]

Hi, I am not sure it is against the rules to post an item for sale in the regular forum. If it is I am sorry and please remove my post.

I have a Omio X8-2200L-USB that I am selling in Grand Rapids MI. I am not interested in shipping as it comes with a large cart.

Here is a link the the Craigslist ad: https://grandrapids.craigslist.org/tls/6182291307.html

[dharmic]

Haven't cut aluminium for a while but had a job come up last week: a run of wrist rest pads for control panels for a robotic broadcast camera system that I designed a couple of years ago. Interesting job, but they decided they wanted something soft for the operators' hands.

Silicone in an open top mould, easy. Started off with high density fibreboard and it was promising - really nice finish etc, but I cocked up the surface coating and the first shot pulled up the nice surface with it. Oops. Decided to go for an aluminium mould instead for longevity. Adjusted tool cutter depths, widths, feeds and speeds for the new material. Mostly roughed out using a 2 flute 6mm HSS endmill running 11000rpm, 1200mm/min (about 50ipm) at 9mm DOC and 0.3mm WOC. Then the outside cleaned up and rounded out with a 12mm carbide ball mill, 12000rpm, 1200mm/min, 9mm deep by 0.5mm step over. All lubricated with a pulsed mist of ethanol and water and air stream clearing chips to prevent re-cutting.

The roughing seemed to go well but oh my god the howl of the ball mill, especially when it got into the corners. Chatter town, machine not stiff enough etc. And when I cleaned it all up and sandblasted for a matte finish, I noticed a trench where the toolpath for the roughing finished one lap (it was 'spiralling' out from the centre at full depth) and stepped out for the next one. Suggesting that the head was nodding back on the move and dropping back down when things went slow enough for the tool to cut in.

Ah well, nothing an hour with a scraper couldn't fix. And I guess I've learned that it isn't the steppers stalling that limits the feed of the machine, it's the flex. Be interesting to see if the new dowel/thread solid bed I'm making up at the moment stabilises some of the flex and chatter but it won't do jack for that nodding. Toolpaths will be adjusted accordingly in the future and I'll have to look at filling in the pool and getting another shed to make room for an old bridgy or similar

[extent]

I think there's so much flex in the gantry that just the bed alone won't make a huge difference. It really doesn't like taking really big axial cuts, even with low engagement. I've kind of settled on around 4mm axial for passes, doing 0.5 stepover for 'roughing' and 0.25 for finishing, but running at 5000mm/min with 6mm tools. Keeping light passes for anything beyond really roughing reduces the amount of stock left if you try to do a skim pass after. If I load up with 1 and 2mm stepovers I can rerun the same profile and get a real chip off it even the 2nd time, so seems better to me to just keep all the forces down lower in the first place.

ZrN coating did improve surface finish for me, even with those light cuts. Bright carbide seems to want to leave a bit of a wavy vertical scallop in sidewalls even when it's not chattering, and ZrN ends up nice and straight which I can only imagine is a good sign.

[RCaffin]

Photo would be nice!

Cheers
Roger

[dharmic]

Of?

:edit: have a video

https://youtu.be/coak0B_ICJY

[RCaffin]

Photos of wobblies and improvements due to the Zrn coating.

Cheers
Roger

[dharmic]

On another front, spent the day with a quality oldschool machinist (50 years in the game) and made some serious progress on the new bed. Bandsawed roughly to size, edges cleaned up on a manual mill, onto his Bridgeport Interact for surfacing on one side, mounting holes to match the Omio frame drilled and counterbored, a 32x32mm grid of holes drilled, half of them reamed to 6mm for dowels and the other half opened out to 6.8mm ready for tomorrow: peining the backs of the dowel holes so they won't fall through, threading the bigger holes to 8mm, bandsaw to remove the machining tabs then back in the manual mill to finish off the short edges.

Phew! But I'm excited, I gotta say. A day spent playing with the bigger machines has left me with some serious toy envy and the sight of a guy who knows his stuff cold dancing across the conversational interface on the Bridgeport was something wonderful to behold.

[extent]

This wasn't an aggressive cut but it's pretty representative

Attachment 365856

[dharmic]

That's pretty noticeable, especially if it's the only thing changed between the two photos.

Finished my new bed today. Alternating dowel and threaded holes for precise positioning of stuff and solid clamping. Surfaced pretty flat (to within a few tenths of a thou) compared to the original bed which had bumps and jumps of over a millimetre on it. Deflection when I put my weight on it has gone from over 1mm on the old bed down to 0.05mm with this new lump of metal.

Need to take it out and shim it to allow for inconsistencies in the frame, about 0.1mm out from one side to the other and 0.25mm front to back but watching the runout gauge on the traverse it seems pretty linear so I reckon some judicious use of beer can shim stock will sort it nicely

Can't wait to throw some ali in there and see if the extra stiffness and mass helps keep the vibration and chatter down on heavier cuts.

[A_Camera]

Looks very nice. I am pretty sure it will make a huge difference.

[Mmpie]

Great work there Dharmic. One day I'm sure I'll follow suit, for now the only piece of metal I've cut so far is a tin lid for the coolant jar, oh and one of my clamps when the machine jammed. I think I'm confident to have a go at milling some aluminium projects now, need to start collecting suitable tooling.

What are the chances of being able to drill an acceptable 8mm hole through 36mm aluminium? Slim I guess. And although I know it's hopeless trying to cut steel, is there any chance of drilling 6mm and 8mm holes in 6 or 8mm steel plate? And countersinking them.

Speed isn't an issue so even if it takes all day if it's possible I'll be pleased


Sent from my iPhone using Tapatalk

[dharmic]

Thanks Mmpie. I've had decent results even with the original bed but the opportunity came up to grab an "offcut" of material cheap and get some free time on a machine big enough to make the thing, so I jumped at it.

8mm * 36mm no problems at all: I'd first centre drill the start point, then run a peck cycle of about 6000RPM, 150mm/min feed, 3mm per peck and back out all the way to clear swarf and allow the drill tip to get re-lubricated each time.

Steel is doable but you're going to want a carbide bit and not do a lot with it. The killer on these machines with steel is that the spindles get gutless and stall easy AND wear out crazy fast (bearings, overheating etc) when you run them under about 5000RPM. So you can do it but you have to be careful.

[Mmpie]

Sounds like an excuse to fit a milling table to my pillar drill... when asking that question it never crossed my mind about the low rpm's needed for steel, I was thinking about rigidity.

Thanks for the tip with the aluminium. A good carbine spot drill I guess? I've only milked foam, plastic and wood so far, when I've needed a hole in anything thicker than about 10mm I've just milled them using a smaller bit as the cheapo HSS drill bits wobble and howl like a banshee creating a worse racket than fingernails on a chalk board...


Sent from my iPhone using Tapatalk

[extent]

ya, you really want carbide drills just so you can use as high an rpm as possible. I had all kinds of problems stalling 1/4" HSS jobber drills, just got lazy and interpolated all the holes for my fixture plate. A nice carbide screw machine drill wont flex all over the place and in smaller sizes like 3mm can run at your full 24k rpm.

[dharmic]

I've done a heap of holes using standard HSS jobber bits up around 6000RPM no probs, just keep the buggers sharp and well lubricated. If you can afford/justify getting solid carbide bits for particular sizes that you use a lot, though, it makes life a lot less nervous

Alternately find a lower speed spindle arrangement that you can throw into the 80mm clamp and swap over... hmmm, got me thinking now