Test Equipment Anonymous (TEA) group therapy thread

[beanflying]

IMNSHO, 3d printers are like solderless breadboards. You spend as much time fettling as printing.

If your mode of thinking is like machining, i.e. adding and removing parameterised component shapes, then OpenSCAD is a good starting point. There are free standalone versions, and browser/web version.

This collet, one threaded, one not, took me a couple of hours to knock together.

  You have uttered this crap once to often and it just doesn't hold up with decent quality modern printers!

It is based on observation of the old and new printer at the local hackspace; the latter is a "Original Prusa i3 MK3".

The local "how to" manual lists many ways in which prints can go wrong and twiddles that can and must be made. It appears there is quite a lot of "suck it and see" involved, which I regard as fettling.

There may be fewer issues if only one person ever uses the machine with only one material, but since I use a variety of materials, I'm never going to own a machine. Hence any prints I make personally will have to be on that shared machine.

Used by non owners and the unwashed masses is not any basis for reality! You have made this same sweeping generalization which doesn't apply to 90-95%+ of the one user printers based on this now explanation.

[Addicted2AnalogTek]

Unrelated to 3DP, but isn't someone here in need of one of these?  It just popped up in the related items to one of my ebay searches:

https://www.ebay.com/itm/VINTAGE-AC-POWER-CORD-OVAL-ROUND-3-PRONG-8FT-LONG/274029501541

[tggzzz]

The slicing setup is an interesting point; is it part of the design or part of fettling, or both.

It's the step between the .stl file and the printer. It generates the gcode that tells the machine how to do everything required - all the directional moves, the temps, the speed and velocity it moves at, the initial layer temp, ... all kinds of things. If the settings in the slicing program (I use Cura) aren't the best they can be (you learn what works best and what doesn't) the print can be like that red case and worse. It won't be the fault of the machine at all!

Also until recently I didn't have manual mesh bed leveling enabled until after I flashed the Ender factory firmware with the Marlin firmware (they are where it all comes from). That made a BIG difference in the quality of the first layer, which is of course what it's all built on top of. 

In the slicer I would have to use (PrusaSlicer), it looks like there are ~50 tweakable settings. The local manual indicates you may need to tweak 15-20 settings to get the best results for your design.

[tggzzz]

IMNSHO, 3d printers are like solderless breadboards. You spend as much time fettling as printing.

If your mode of thinking is like machining, i.e. adding and removing parameterised component shapes, then OpenSCAD is a good starting point. There are free standalone versions, and browser/web version.

This collet, one threaded, one not, took me a couple of hours to knock together.

  You have uttered this crap once to often and it just doesn't hold up with decent quality modern printers!

It is based on observation of the old and new printer at the local hackspace; the latter is a "Original Prusa i3 MK3".

The local "how to" manual lists many ways in which prints can go wrong and twiddles that can and must be made. It appears there is quite a lot of "suck it and see" involved, which I regard as fettling.

There may be fewer issues if only one person ever uses the machine with only one material, but since I use a variety of materials, I'm never going to own a machine. Hence any prints I make personally will have to be on that shared machine.

Used by non owners and the unwashed masses is not any basis for reality!

It is the basis for one valid version of reality. I accept you may live in a different, equally valid, version.

You have made this same sweeping generalization which doesn't apply to 90-95%+ of the one user printers based on this now explanation.

You are making equally sweeping generalisations for the 90-95% of people that use shared printers. (But it is a bit difficult to understand exactly what your sentence is trying to convey)

[Neomys Sapiens]

Ooooooooooh, I like it!

Edit, here is some of my rpn stash, I have a 41C collection - printers, wand, plug ins. HPIB IF.
Quite a few 41s with the battery connection roo**d but most work.

The first 28 operator besides myself! 28C or 28S?

Now that HP-IL came up: I already wondered that nobody seems to operate the HP41C? and HP3468A combination. But given the success of HP-IB, I always thought HP-IL would have been better suited to a high-end handheld DMM.

[Neomys Sapiens]

nobody talks about the instrument what is sitting on the HP counter...
This is a nixies device:

Before the guests can play with the old instruments there we want to play with them,
and for this job we need that instrument. It is a 5 digit Nixie radiation detector, readout in µSievert 
The probe is made by Thomson in france, a large black beta/gamma sensitive Geiger-Mueller-tube, high sensitive. Typical used in weather stations.
This tube are connected to around 500 Volt in series with a high M resistor, delivering the typical tick..tick...tick what will be counted.
The highest range of the instrument you will not read bec you lost your Hair, teaths, and your eyes are glowing in the dark 

Oh come on, nobody objected to a purrrrrrr sound when it came from the cat, so why that panic when the GM counter purrrs?

Contrary to public opinion, the human body is equipped with a radiation warning: you should be able to taste the petechia from your mouth memranes before damage becomes irreversible.

[xrunner]

In the slicer I would have to use (PrusaSlicer), it looks like there are ~50 tweakable settings. The local manual indicates you may need to tweak 15-20 settings to get the best results for your design.

Yes and in mine (Cura) there are many more well over 100 that could be tweaked, BUT ... (there's always a but)

MOST of them will not need to be tweaked because there is a default setup for many printers you can select. Now then, the ones you will end up tweaking will make or break you print and your satisfaction in the hobby. Here's a few that I've tweaked -

Bed temp
Initial layer speed
Print speed after initial layer
Nozzle temp
Jerk setting
Print quality (layer height)
Whether to use a skirt or brim for initial layer stability
Which side you want to print facing down (it may very well not be the way it comes out of the CAD program unless you thought about it during the design process)
Whether to use extra support in the printing process to hold up overhangs (it tears off after the job is done)

Also, you can preview the slicing operation in the slicer by dragging a control. It lets you see how the machine will actually move during the print which as you may know is layer by layer up from the bottom, i.e. it never goes back down a few layers - always up. Sometimes even looking at that movement will alert you to a possible small design change you might want to make in the CAD process to make it come out best.

Probably a few more but it's not really that bad. Like I said, go to the library and you are at the mercy of the person preparing the file and setting up the printer and maintaining it. They may or may not have much more knowledge than you do. At home, you are the master and nobody will care more about your project than you do.


It's like any other hobby or profession - you have to get over the learning curve, after that hump it isn't all that bad.

[bd139]

Yeah usually iterative process. Plan for me is:

1. Load the defaults into the slicer. I have a recommended slicer and settings for ender 3.
2. Print something simple.
3. Google what I am not happy about.
4. Frob settings to try and make it go away.
5. Go to 2 unless I'm happy
6. Print something better.
7. Evaluate it, go to 3.

I have about 20 things I'm going to print already. This might turn into an unwelcome distraction 

Of course after test prints, a fuckofftopus is being done

Have ordered some white PLA, grey ABS to play with. I will be using PLA until I get the hang of it, then trying ABS as I can use acetone vapour smoothing on that. So fine layers, then acetone smoothing for one thing which needs to look decent with little finishing.

[capt bullshot]

The first 28 operator besides myself! 28C or 28S?

First? Bought mine around 1988 - still works
Always liked RPN and still do, though most often I use the ubiquitous Casio 991

[mnementh]

<-------  Is over here casually trying to ignore the 3DP talk while my friend sends me a text that reads:  "Would you be interested in going in half on a 3D printer? I want to make some R/C parts and I know it'd come in handy for you, too."

 

Let us know which way you are facing when you stop spinning.

Let us know how you ensure that each individual manufactured item has appropriate mechanical properties to ensure it doesn't fall apart sooner than desirable. That's an issue I'm facing with my collet.

FDM is not the right tool for this job; tensile strength of even well-bonded layers is relatively poor. FDM's strength is typically more compression. SLA MIGHT hold, at least as well as OEM, but farming that out to a contractor will probably cost anywhere from $20-50 $5ish...? Holy shit that has dropped! depending on operator and material.

For your application, I'd mod a RC model airplane prop adapter collet and have done with it: Find one that matches the taper of the hub and the motor shaft size; this will require some trial & error at a hobby shop. The prop stud will be too short, so you'll have to drill & tap the center of the stud for a m2-m3 screw, then use a tiny fender washer over the hole in the middle of the squirrelcage. You may need to turn the prop stud down or drill the center of the squirrelcage out to match the prop stud.

Or, replace the collet with a whole grub-screw type prop adapter and drill the center of the squirrel-cage out to match the stud on the prop adapter. You may need to turn the prop stud down or drill the center of the squirrelcage out to match the prop stud.



Remember that this application is not high-precision; I literally made a serviceable replacement fan by epoxying the squirrelcage to the hub of a computer fan. Vibration was still minimal even with a good several grams out of balance.

mnem
moo...?

[med6753]

Holy shit....will you guys SLOW DOWN.  I can't keep up with the popcorn. 


 
 

[mnementh]

No.

mnem
 

[mnementh]

Yeah usually iterative process. Plan for me is:

1. Load the defaults into the slicer. I have a recommended slicer and settings for ender 3.
2. Print something simple.
3. Google what I am not happy about.
4. Frob settings to try and make it go away.
5. Go to 2 unless I'm happy
6. Print something better.
7. Evaluate it, go to 3.

I have about 20 things I'm going to print already. This might turn into an unwelcome distraction 

Of course after test prints, a fuckofftopus is being done

Have ordered some white PLA, grey ABS to play with. I will be using PLA until I get the hang of it, then trying ABS as I can use acetone vapour smoothing on that. So fine layers, then acetone smoothing for one thing which needs to look decent with little finishing.

Welcome home, buddy! I'll do my best to misguide you the same way I learned!

Honestly, you're looking at the process with the correct, pragmatic view. You'll get a few wasted prints (like I did with neoBracket trying to get printing on THE SIDE to turn out good) but even THOSE will eventually get turned into something useful.

Be mindful that PLA and ABS are literally 2 different processes... you need to manage environmental temp much more carefully with ABS, and it stinks out loud, so is NOT "living-room-friendly" even for bachelor-types. When you start doing ABS, you'll want to move it to the garage, close to the big doors so the stink can get out. Also be prepared to either make an enclosure or arrange the printer so you can drop a big cardboard box over the whole thing while it works.

You may find that filling/sanding/painting a single "example" PLA part is considerably less assache than trying to get good prints with ABS just so you can vapor-polish.

Cheers,

mnem
*enabling*

[med6753]

Something in my 7854 has decided to take an occasional dump to ground, and I've only managed to notice because the case shocked me earlier. As far as I can tell, there's no other faults occurring when this happens. It's only a mild shock as one would expect from an otherwise functioning device.  The fault exists regardless of what plugins I've got installed, so it's definitely in the mainframe itself. It does take around 30 minutes of operation before the fault manifests itself, and it's doing so fairly consistently.

I've never had to track down a "mild" fault such as this.  Anyone have any suggestions on how to tackle tracking down the bugger?

The power supply voltages and ripple were good last I checked, but I'll be checking them again at startup and once the fault manifests.

As for tracking down the fault to a particular board, I'm considering using a 7A22 in another mainframe with the negative input probe attached to the ground at the mains input and move the positive input probe around each board's ground test points to narrow the fault down to the board, then continue from there to find the bastard.

Does anyone have any ideas?

Am I reading this right? You're getting a shock off a supposedly grounded chassis? Correct??   

If yes you need to determine why your scope isn't grounded. Is there a problem chassis to ground prong on the power plug? Or do you have bigger problems with the lack of a ground on your mains wiring? You need to figure this out first.

I think, but not sure, that the EMI filter has capacitors in it tied between hot and ground. If you don't have the chassis grounded these caps could cause the chassis to elevate to a fairly high potential, sometimes as high as 80 to 90 volts. The current is real small but you might feel it. And if the caps are going leaky then the current will go up. And also this leakage current could damage other equipment.

Be safe. If you are getting a shock off the chassis of your equipment something is seriously wrong and you need to shut it down and determine ASAP what's going on. 

I agree 100%, somewhere you have lost the ground connection. In what manner are feeling this leakage, is it from the case of the scope or is it via the ground lead on the probe when connecting to the DUT?

If it’s the latter then you have still got a problem with the scope, possibly a bad connection between the BNC socket and scope chassis if so then the problem exists on the DUT.

If it’s not the filter caps going bad in connection poor grounding, then you need several meters connected the various supply rails with a min/max recoding ability to which rail is being dumped to the chassis periodically and go on a hunt to track down the part on that rail. This testing could be applied to both scope and the DUT but only the DUT would only be suspect if the shock is only felt on the probe lead.

It's definitely the scope itself, as it happens only when the scope is not connected to any other devices. When it happens, if I connect it up to a FG, it grounds through the BNC cable and I can no longer get a shock.  To clarify, when I touch it it's not anything more than a very mild tingling.  Before I power it on again, I'm going to be going through and checking the continuity/resistance of the grounding internally on the scope, as I'd like to find the issue without powering it on again.  What's got me thrown off, though, is that the beast's power supply is a SMPS and there's absolutely no other indication of any other faults, no SMPS shut down, no occasional non-starts, nothing. This all leads me to think that the ground is floating 100% of the time.  Also, I don't have a thorough understanding of SMPS, which may be evident here. 

Either way, getting to the bottom of this issue will be the very next thing I do.

Measure chassis to ground prong on the power cord. You should have continuity.

[med6753]

No.

mnem
 

Well...FU too and the horse you rode in on.  I'm going back out to the garage and further defile the Type 114 with another coat of paint. 

[tggzzz]

Like I said, go to the library and you are at the mercy of the person preparing the file and setting up the printer and maintaining it. They may or may not have much more knowledge than you do. At home, you are the master and nobody will care more about your project than you do.

It's like any other hobby or profession - you have to get over the learning curve, after that hump it isn't all that bad.

Just so.

I may need to use PETG rather than PLA, and not ABS since layer adhesion could be a problem. Rather than use nylon, I'd probably get Shapeways to make a sintered nylon part.

For my collets it looks like I'll need to play with
- layer height, first/rest
- vertical shell perimeters
- horizontal shells
- ensure vertical shell thickness
- avoid crossing perimeters
- fill density
- fill pattern
- top/bottom fill pattern
- brim width
- extrusion width
- solid infill every N layers
but avoid all the support material tweaks, advanced settings, and a few others.

[med6753]

Like I said, go to the library and you are at the mercy of the person preparing the file and setting up the printer and maintaining it. They may or may not have much more knowledge than you do. At home, you are the master and nobody will care more about your project than you do.

It's like any other hobby or profession - you have to get over the learning curve, after that hump it isn't all that bad.

Just so.

I may need to use PETG rather than PLA, and not ABS since layer adhesion could be a problem. Rather than use nylon, I'd probably get Shapeways to make a sintered nylon part.

For my collets it looks like I'll need to play with
- layer height, first/rest
- vertical shell perimeters
- horizontal shells
- ensure vertical shell thickness
- avoid crossing perimeters
- fill density
- fill pattern
- top/bottom fill pattern
- brim width
- extrusion width
- solid infill every N layers
but avoid all the support material tweaks, advanced settings, and a few others.

But why a threaded one? The fan shaft isn't threaded.

[tggzzz]

<-------  Is over here casually trying to ignore the 3DP talk while my friend sends me a text that reads:  "Would you be interested in going in half on a 3D printer? I want to make some R/C parts and I know it'd come in handy for you, too."

 

Let us know which way you are facing when you stop spinning.

Let us know how you ensure that each individual manufactured item has appropriate mechanical properties to ensure it doesn't fall apart sooner than desirable. That's an issue I'm facing with my collet.

FDM is not the right tool for this job; tensile strength of even well-bonded layers is relatively poor. FDM's strength is typically more compression. SLA MIGHT hold, at least as well as OEM, but farming that out to a contractor will probably cost anywhere from $20-50 depending on operator and material.

I strongly suspect you are right about FDM's strength, but I'll use it as a means of putting the new printer through its paces.

The SLA is solid rather than "hollow", and at £4 for 4 collets delivered (DirtyPCBs), I'm prepared to take a chance.

If that doesn't work, there's always brass or steel or aluminium for ~$17 + postage (Shapeways).

For your application, I'd mod a RC model airplane prop adapter collet and have done with it: Find one that matches the taper of the hub and the motor shaft size; this will require some trial & error at a hobby shop. The prop stud will be too short, so you'll have to drill & tap the center of the stud for a m2-m3 screw, then use a tiny fender washer over the hole in the middle of the squirrelcage. You may need to turn the prop stud down or drill the center of the squirrelcage out to match the prop stud.

Or, replace the collet with a whole grub-screw type prop adapter and drill the center of the squirrel-cage out to match the stud on the prop adapter. You may need to turn the prop stud down or drill the center of the squirrelcage out to match the prop stud.



Remember that this application is not high-precision; I literally made a serviceable replacement fan by epoxying the squirrelcage to the hub of a computer fan. Vibration was still minimal even with a good several grams out of balance.[/b][/i]

I've got a recommended prop adaptor collet on the way, but I'd like to try to avoid butchering the fan if I can.

[Neomys Sapiens]

The first 28 operator besides myself! 28C or 28S?

First? Bought mine around 1988 - still works
Always liked RPN and still do, though most often I use the ubiquitous Casio 991

First to post here about it. At least as far as I knew.
1988? 1989 for one of mine, on the very same day as I ordered my Fluke 87. I found the second in a discarded desk much later. A former coworker was always when seeing mine stating that he got the memory-challenged 28C just before the 28S came out. Well, can't help there!

[bitseeker]

A tribute to bitseeker -

LOL! Thanks to all of you who make the TEA house a great place to be: xrunner, Cubdriver, med6753, mnementh, bd139, Kosmic, tggzzz, Specmaster, tautech, nixiefreqq, beanflying, Neomys Sapiens, wch, and more... It certainly wouldn't have eaten the forum if I was just talking to myself.

As you may have noticed, I'm playing catch up on the thread (a victim of my own success ). Currently on page 1572, I'll apologize in advance if I miss some things in an effort to get current before 2020.

[tggzzz]

But why a threaded one? The fan shaft isn't threaded.

Correct, but I'll reuse the metal part that connects the collet to the fan. That has a thread on both sides, one for the collet, one for a nut.

If my collet's thread doesn't work well enough, then I'll use force for an interference fit. (£4 gets me 3 spare SLA collets )

[xrunner]

As you may have noticed, I'm playing catch up on the thread (a victim of my own success ). Currently on page 1572, I'll apologize in advance if I miss some things in an effort to get current before 2020.

No problem. I have no idea how we ended up on the topics for today from wherever you left off, it's a TEA sort of thing you know. 

[tggzzz]

The first 28 operator besides myself! 28C or 28S?

First? Bought mine around 1988 - still works
Always liked RPN and still do, though most often I use the ubiquitous Casio 991

First to post here about it. At least as far as I knew.
1988? 1989 for one of mine, on the very same day as I ordered my Fluke 87. I found the second in a discarded desk much later. A former coworker was always when seeing mine stating that he got the memory-challenged 28C just before the 28S came out. Well, can't help there!

My first RPN calculator was a Sinclair Scientific in 1974, which managed to squeeze everything (keyboard, display, arithmetic, transcendental) into 320 instructions. See them simulated here: http://files.righto.com/calculator/sinclair_scientific_simulator.html

My second was an HP35 bolted to the university library desk.

My third was a 50th anniversary edition HP35S, in 1989 of course. Plus a spare.

Then I bought an HP35, for old times sake.

All work, provided they have batteries in them

[bitseeker]

As you may have noticed, I'm playing catch up on the thread (a victim of my own success ). Currently on page 1572, I'll apologize in advance if I miss some things in an effort to get current before 2020.

No problem. I have no idea how we ended up on the topics for today from wherever you left off, it's a TEA sort of thing you know. 

Yup. It ebbs and flows of its own accord. Perhaps I will tackle this differently. I can use two tabs: one for reading the backlog and one starting from here.

Oh, and welcome to page 1600!

[tggzzz]

LOL! Thanks to all of you who make the TEA house a great place to be: xrunner, Cubdriver, med6753, mnementh, bd139, Kosmic, tggzzz, Specmaster, tautech, nixiefreqq, beanflying, Neomys Sapiens, wch, and more... It certainly wouldn't have eaten the forum if I was just talking to myself.

Sometimes I want to thank you, sometimes to damn you for "wasting" so much of my life