Just Checking In and Catching Up #44 - Delhi India

[aniljangra]

Originally Posted by jhiggins7

I'll be interested in your observations once you get your spindle working and start using your MechMate for long cutting sessions.

John, the last 3D job is 12x12 inchs approx and took about 50 mins using 4mm ballnose cutting about 0.5" deep in single pass with stepover = d/7 at 150 ipm.

Only Z drive seems getting hot (still touchable) in that an hour time, I think because :
- its the last one in the row, far away from vetilation
- Z motor is 4.6 amp and the coils are connected in series (more inductance)
- and of course its the one doing most work in 3d job

I think with a fan blowing air on drives will be sufficient for any cutting job.
Will keep updating as I go

[jhiggins7]

Very nice work, Anil. Love the lion's head.

I wish I could lend you my Temperature gun. I understand from this Forum that the stepper motors can run safely at 80 degrees C. I think that would feel really hot to the touch...you couldn't hold your hand on the motor.

When you cut the lion's head, what was your driver temperature like?

[aniljangra]

My cam seems to get better pics in daylight, here is again same pic.
first-3d.jpg

John,

Motors run cool with no sign of heating up, all of them only just warm.

While cutting, as mentioned earlier only Z-driver seems to heatup (still can hold hand on it), all other drivers were normal (just warm), Can't give you exact temp figures as not measured, but can say 3 drivers and all motors were warm similar to a Hard Drive running in computer. (may be about 30 deg C)

I'll connect the Z motor as single coil and It'll cool down the driver as other in system. I feel Z is heating up due to very high inductance of Z series connected coils.

John, even with it heating up there was no problem as electronics can run at that temp safely.

[javeria]

congrats again - you seem to be doing it very very well

[jhiggins7]

Thanks, again, Anil.

Really great work. Thanks for the feedback on your motors and drivers. Hope all continues to go well for you.

[KenC]

Nice work Anil.

[aniljangra]

Thanks you all, Irfan, Ken & John.
Now I need nice files to continue cutting

[aniljangra]

Amazing, how tilte of my thread changes automatically

[Claudiu]

good job, Anil.

[aniljangra]

Thank you Claus,

Finally got PCBs done and my BoB and controller for proxies and switches installed on MM. Now I can ref home and use the switches finally

The charge pump on Bob is done using a tiny micro-controller and same way a controller riding on Gantry takes all inputs from proxies and buttons and combine them to BoB output.

The BoB with charge pump and a simple PWM to 0-10v.
Bob.jpg

The Controller for proxies and switches, the LEDs show the status of Proxies, on estop all LEDS turns on to show if any of the estop button pressed.
cntrl1.jpg

Back side of board is all mess with all wires coming to connectors.
cntrl2.jpg

[javeria]

great work

[KenC]

Fantastic! Wish I know how they work.

[vishnu]

Hi Anil,

Thats a nice BOB, keep us updated on the perfomance

Vishnu

[aniljangra]

Thanks you Vishnu/Irfan

Originally Posted by KenC

Fantastic! Wish I know how they work.

Ken, Don't understand what you mean ?

If you want to know how its working for me, then yes these seems to be working fine for me , for now. I'll test it with VFD when I get the spindle and report.

If you want to know how its programmed (ts simple stuff anyway), then for controller on Gantry :
samples all switches continuously and whenever any of "start" buttons is pressed it, it makes "start" output high. same way for "pause" buttons.

Proxy and estop inputs are sampled every millisecond, on estop input the estop line is made low and all LEDs are put on to show the event.

On any of proxy input the proxy output is made low and corresponding LED On to show the status.

(Proxies are not sampled continuously along with buttons, but every millisecond so that the homing sequence is repeatable every time exactly)

On BoB, there are buffers on inputs and outputs with the addition of a simple RC network to convert PWM to 0-10V signal for VFD.

and for Charge pump, there is a tiny micro-controller that counts pulses from CP pin every 10mS (for a frequency in the range of 6KHz to 13Khz), if the pulses are found in the range 10 times successively the outputs are enabled.

Same way on enabled outputs the CP signal has to miss for 10 successive times before the outputs are disabled.

In turn thats 100mS before the outputs are enabled/disabled on valid/invalid signal. I'm sure it won't act on noise with all this

By the way, can mach use same input for z-zero sequence as the other proxy inputs ? If yes, then the z-zero input can be combined as well, also the same 2 buttons when pressed together can initiate the z-zero sequence !!!

[KenC]

Anil, Thank you for your detail explanation of the working principle. I'll have to spend some time understanding it.
Woulkd you kindly elaborate the RC tank convert the PWM analog signal? Does the PWM signal comes in the form of pulse train or bit stream? I am having the idea that the PWM is a Psudo Analog signal...

[aniljangra]

PWM as the name suggest is pulse train of varying pulse widths. For low value the on time is less compared to off time, same way for higher value the on time of pulse is grater than the off time. So if you just average the pulses you will get an analog signal (of course in the range 0-5V).

To get 0-10v output you need to amplify it. I used a transistor to shift the level of pulse to 10v and then averaged.

[KenC]

Now I can see much more clearer. Thanks

I'm so happy!

[aniljangra]

As usual, again showing a PCB done on MM
This time made few adapters for surface mount parts for my electronic prototype work.

This pic is interesting because of the sizes involved on smds parts, this pcb is milled using 0.2mm bit and note the track width is only 0.1mm. (calculate that in inches )
smd-pcb.jpg

Some tracks are exact width but some are a little wider, Is that backlash ??
Also noticed that the pattern is consistent on all of them

[Gerald D]

Ouch! that thing is only 1" square! Okay, while the tracks are supposed to be 0.1mm [0.004"] wide, some are twice as wide as others, range probably about 0.05mm [0.002"] to 0.15mm [0.006"], which means you are holding an accuracy of +/- 0.05mm [0.002"]. That is too technical, and it tells us nothing . . . .

Did you always go in the same direction around each pad/track? If you go clockwise around, you will get wider tracks than if you went anti-clockwise, because of the flex in the cutting bit.

[jhiggins7]

Anil,

Well, .1 mm is about 4 mils (.004 inches). Not bad. Nice work.

I think I see the pattern...the lines are thicker on the left of the adapter picture...is that what you are seeing? I have no idea why it happens.


Were all the adapters you cut usable? I mean, were any of the signal lines too thin or "open"?

[Gerald D]

John, you posted the same time as me. I wasn't aware that 0.004" was called 4 mils.....I would have said 4 thou. ??

[jhiggins7]

1 mil = .001 inches or 1 thou. See here. I'm no machinist, so I just used the term as something I'm familiar with.

I'm also aware that a "extra fine" PCB trace width is one or two mils. So I was letting Anil know that the achievement of 4 mils using a CNC router was rather remarkable.

[aniljangra]

Originally Posted by Gerald D

Did you always go in the same direction around each pad/track? If you go clockwise around, you will get wider tracks than if you went anti-clockwise, because of the flex in the cutting bit.

Hmm, I did not think about the flex in bit, I reran the file again (in slow motion ) and noticed that all but 3 tracks on top left are milled clockwise !! But don't think its the cause in my case.

Sorry I forgot to mention that I used 30 deg V bit and closely zooming in the picture found that the depth is increasing from left to right, causing the isolation widening and track shrinking. (try zooming the pic to see what i mean)

(It must be level of the table or PCB that is fixed with double tape, even may be the level of my rails that's causing depth variation).

John, tested the continuity and the PCBs are very much usable

[aniljangra]

Originally Posted by jhiggins7

1 mil = .001 inches or 1 thou. See here. I'm no machinist, so I just used the term as something I'm familiar with.

I'm also aware that a "extra fine" PCB trace width is one or two mils. So I was letting Anil know that the achievement of 4 mils using a CNC router was rather remarkable.

Thank you John,
This measurement stuff is confusing in PCBs even more as the track and components are usually measured in mils (fraction of inch) and thickness of copper on PCB in microns (fraction of mm)

[javeria]

what ipm did you get this done pal

[Gerald D]

Thanks John, I hadn't taken notice of "mils" before.

Anil, my first thought on the varying widths was V-cutter depth, but I read your post twice: "this pcb is milled using 0.2mm bit" Well, I had to believe you had a very thin flexible bit!

[aniljangra]

Very slow Irfan

Feed rate was 20 ipm but it was scaring to see the rapid moves in between with small bit.

[jhiggins7]

I zoomed the picture. I can't see the depth of the groves. But your explanation of the problem cause makes a lot of sense.

Keep us posted. This is a very interesting development.

[aniljangra]

Originally Posted by Gerald D

Anil, my first thought on the varying widths was V-cutter depth, but I read your post twice: "this pcb is milled using 0.2mm bit" Well, I had to believe you had a very thin flexible bit!

Correct Gerald, I too gone back and read it again, the smallest I have till now is 0.6mm used for rubout sometime. (to mill away the excess copper between track)

[riesvantwisk]

Anil,

for me this looks like an interference pattern like a moiré pattern, I have seen this a lot when I was working in the pre-press industry. This happens when the required design doesn't match your machines capabilities in such a way that sometimes the machine decides to add a extra step, or removes an extra step.

You can test this by creating a pattern of lines witj increasing distance.

So make a line of 4 mill thick and 10mm length (sorry I don't think in mils anymore... metric rules) spacing 2 mil apart and stack 20 orso on top of each other.
Then next to that use teh same 4 mil line, 10mm wide but with 3 mil spacing.
Then next to that use the same lines, but with 4 mil spacing. If you print this a pattern should appear of thicker and thinner lines and even if you stick enough on top you can even see a wave ....

Was I clear?

I think this is the problem....

Ries