Hi Strawb - I've watched some of his vids. Very good, but he does not discuss the control side of things. I may have to send him a note. Peter
edit - I downloaded the MG10HL cad model very nice.
That's actually a great idea. Had a look around and they have 4um slabs in thickness that I need. Particularly 1.600x1.000x200mm for EUR 2.500. I can then cut out or have them cut out all the parts that I need from one slab.
2 years ago, when I started this process, I tried building with slabs but I guess didn't have enough experience for that. I just revisited my old gantry frame and everything fits nicely and much cheaper then having plates machined. Now I gotta rebuild the double column with slabs and then decide whether to go double column or the gantry.
About the previous epoxy discussion, would I fill it with steel powder or pure epoxy, when it comes to gluing plates to concrete/granite?
Hi Ard - I wouldn't use a steel filler. I'd use a structural putty. The granite surfaces have to be roughened not polished. Typically sandblasted but an orbital sander to roughen would be OK. Epoxy grips by flowing into the small interstices in the surface and then curing keying the surfaces together. I would use a ~30g glass tissue between the surfaces to ensure you had a bondline if you used a runny epoxy. Tissue can be bought from any composite supplier and is available in 25-100gsm. The best bonds are acheived by using a hot air gun to remove moisture from the surface (or a propane torch is even better) then use a runny epoxy to prime the two surfaces, when that's tacky use an epoxy putty to set the surfaces. An ideal bondline is 0.25mm to 0,5mm thick thats what the tissue is for.
https://trojanfibreglass.com.au/prod...urface-tissue/
Peter
Hi Ard - heres a good intro to bonding prep. Flame treatment is becoming standard practice as prep. Peter
Good to know about the flame method! I've been using the hot air gun the whole time and there's a torch right in the corner of the shop!
Our procedure when I was at the servo company was:
1) Clean with isopropyl alcohol and Chem Wipes
2) Buff with Scotch Brite pad
3) Clean again with isopropyl alcohol and Chem Wipes
4) Heat gun to ensure complete evaporation
5) Wet the bond surface of one part with the mixed epoxy by applying dots and then spreading and working it in with wooden or plastic sticks (popsicle sticks or tongue depressors) or using a neoprene glove and your finger(s)
6) If thick bond line, wet part 2, otherwise ensure complete coverage of the epoxy
7) Install parts together.
8) Use swabs and Chem Wipes to clean excess
9) Cure per schedule
Some but few of the epoxies we used required clamping. Lots of Delryn fixtures since the epoxies almost universally just flake off of Delryn. PTFE is the ultimate but flexible, not great for rigid fixturing.
We were dealing with preventing high energy magnets from throwing off rotors at high speeds in expensive machines and even life critical applications. The above wasn't exactly as specified by the epoxy suppliers, but it resulted in zero bond failures the two years I was there and years preceding me.
EDIT: I should note we designed relief grooves to maintain bond thickness over 70-90% of the bonding faces. Otherwise we used glass sheets or glass bead filler for that purpose, but the built in bond line relief was far more reliable for mass assembly.
Hi Strawb:
1) cleaning with an aromatic solvent chills the surface and water will condense on it even at low RH. Epoxy cannot penetrate a water barrier. Some adhesives are moisture curing so this does not matter eg urethanes and some acrylics
2) buff with scotchbrite is to roughen the surface but can contaminate it as well hence
3) reclean with solvent but then it can get condensate on it
should add the distilled water break test here, then dry surface
4) heat gun to remove condensate. Should look up the saturated vapour chart and you will be able to determine the surface temp needed to flash off vapor. A propane torch is better as the oxides in the flame activate the surface. The surface energy actually increases with the flame. An air gun only removes water.
Rest same - in one project where we bonded a personnel basket to a crane for high voltage overhead power cable work we found that the finger application was best. Don't know why but every time we used paddles or V paddles or straight gun application we got lower test results. Typically we needed 4T at test and some people scored 50kg was a shock! Maybe using a finger makes the person more mindful about what they are doing....Peter
If you do a search for BMW chassis assembly you will see many robots with flames at their ends preping surfaces before bonding parts together. Other words are corona prep plasma prep.
see vid at 8min, but the robots don't have fingers....
That's excellent info.
Hi All - I've been digging around for robot control info. There are a few open source codes that look good. Unfortunately they don't work for the ballscrew design only the direct drives. So I can 1) make a model or machine and map the motion then use the software and write a post processor that converts the code or 2) find a smart person to write a Jacobian for it. Here's a robotics course if anyones interested. Peter
https://petercorke.com/resources/robot-academy/
lasers do a great job on steel, grit blasting is ho hum now
Here's the basket, the jib is fibreglass and the assembly has to resist 250,000 volts. Its all bonded together. Toughest problem I've had in 20 years. It had to pass fire tests, electrical tests and structural tests including a 25 year life cyclic test. Max service load 750kg design load 3000kg. Endurance test at 150% load for 250,000 cycles then tried to break it. The ram was not long enough to break it, gave up at 4500kg.... lots of steel brackets and bolts broke along the way. I think the test ran for just over a week. Peter
G'day All - A little bit about epoxy for those EG people out there. I've been using a very low activity epoxy for painting my washroom. Its very thin (for infusion) but has a gel time of hours so I have had to accelerate it by preheating it so I can get 2 coats in a day. Here is a shot of the resin at 25C after being stirred with a tongue depressor. Its full of bubbles. If you use a normal laminating epoxy it will be even worse as that type of epoxy is really thick, full of thixotropes so it won't run much. You really don't have much chance of getting air out. Now look at the clear image. This is the same resin heated to 45deg C and it has clarified by itself. This could be poured into a dry mix in a mould and it would permeate through very well. Its gel time could be controlled to 1-2hrs lots of time to seep. The epoxy is Sicomin SR8100... Peter
How much does the Sicomin SR8100 cost?
7xCNC.com - CNC info for the minilathe (7x10, 7x12, 7x14, 7x16)
Hi Pippin - Don't know waiting for the bill. Its distributed by Lavender-CE in Oz.
Welcome to Lavender Composites » Lavender CE Pty Ltd Send them a pigeon.. Peter
While looking for spindles I found the company that makes them for HSD and hiteco. I looked particularly at their SAB150 spindle. They sell it for 3,000-4,000. The same exact spindle HSD sells for 10,000. I assume other spindles are also at manufacturers price. Maybe you can find something better then Gpenny at a good price.
https://www.olispeed.com/
Hi All - A question for the electronics guys out there. Small cheap routers often connect two motors from a single driver to cut cost. I have played with this but they didn't run fast. Mid band resonance I think kicked in. So the driver becomes confused with the two motors. My question is can you run two drivers from a single BOB output? With the lifting gantry design I need more axes so this would solve part of that problem. Next time I have a test board set up I will test this but someone may know the answer already... Peter
What signal type are you using from the BOB to the driver board?
Alright, let me put it differently with a bit more answer. If you are using a PWM output signal, which I assume, then the question is going to be whether or not the BOB can handle twice the current on the output signals. The current is very low, this is a 5V signal only used for data purposes and not load power,
That said, it's good to make sure that the BOB has had this in mind. The attached, for example, is from a Mach4 BOB. It specifically allows for two motor drives under 'Motor3'. The software doesn't see them as two motors, just as a single motor. The wiring for this setup allows the controller to reverse the signal on the motors.
If you are just using the single output, then at least here they just blanket state yes... https://buildyourcnc.com/FAQ/14260
Hi Strawb - This is the BOB I usually use and the driver. no spec on the 5V amp out capacity. But being a control signal it will be very low? Your link seems to think that this is fine in two spots so next time I have a board set up I'll test the theory. Ta Peter
It's 99% safe to assume that the 5V PWM circuit in the BOB is protected via a resistor against over current even against a full short.
In a really quality manual, they would specify the protection method (ie, pull-down or pull-up or other) including isolation etc. Same on the motor driver side.
EDIT: Assuming said current limiter is in place, the worst case is that the driver inputs are too low of resistance and they end up seeing less than 5v due to resistor divider effects, possibly resulting in a loss of signal to the drivers.