Test Equipment Anonymous (TEA) group therapy thread

[mnementh]

             

So... Took it all apart down to the MB; unplugged & reseated everything I could, cleaned a whole herd of dustbunnies out and... not a damned thing different.

Tried C's magical incantations (or at least best I could according to research aboot what keys on a PC KB map over to Mac) and no joy either. If I boot it with no GPU, I hear the Mac welcome tone, and the HDDs hashing, so at this point I think it actually is trying to boot, but has a dead GPU.

Unfortunately, the list of supported GPUs is pretty short. And surprise, surprise, surprise... only GPUs flashed with a Mac-specific FW will have the preboot splash. 

Here's some pretty pics of the insides... I'm gonna toddle off to ded now.

mnem
 

I'm pretty sure, like all the nvidias of that age, is hit by bumpgate; maybe try an "oven reflow"?

Yeah, I actually had considered this; I've been eyeing my cheap hot-air rework station all morning over coffee. So far, the coffee has won out every time. 

Think you can stick any card in it but you just won’t get the boot splash and helpful EFI stuff as these don’t have a PC style BIOS.  If you get boot camp assistant on another windows PC and use that to make a windows install that’ll use a basic VGA driver AFAIK so it’ll come up when windows boots.

Yeah... I was aware of the Mac's lack of evolution past a need for BIOS ; but did not realize that there wouldn't even be a preboot splash screen from the GPU's firmware (ie, some signs of life) until I did some research.

It appears the cMP (original MacPro) had a limited list of cards with support at all; at least going by this webpage: https://forums.macrumors.com/threads/gpu-compatibility-list-for-cmp.2174600/

I'm guessing this is because that was the very beginning of Apple's dalliance with Intel hardware...?

I'll admit it is a steep learning curve here, especially for someone who hasn't Mac'd since the days of PowerPC... even the G5 I used to "have" was a work machine I had no choice aboot using, IYKWIM.

mnem
This literally is what my living room looks like right now... (spits cable out so I can type)    

[Kosmic]

That charger is way less expensive than I'd have thought it would be. Might be worthwhile picking up a couple of them so you have a spare.

The charger is actually in the meter. This is just a AC/DC adapter. Supposed to output 12V 500ma.

[ch_scr]

Little jug ear teaser, more later.
Edit: added inside / backplate view

[mnementh]

so who got the deal of the century? Immaculate looking Tek 2467B for £135:    https://www.ebay.co.uk/itm/Tektronix-2467B-4-Channel-400MHz-Analogue-Oscilloscope/143988361052

Its described as not working. A bit of a gamble at £135, it could have a bust tube for starters. You might have dodged a bullet 

Yeah that's a possibility but worth twice that for the parts value on its own.

Yeah, just the visible model-specific front fascia, bail and front cover will net you a profit. Everything else is just gravy.

And if you managed to get it up & working with just a recapped PSU...   

* insert soundclip of Dave going "wah-wah-wah..." *   A shame we missed it...

mnem
time to feed the kiddles... not sure who-o-o I'll feed them to-o-o...

[THDplusN_bad]

@ch_scr: Nice, nice, nice work - really!

 

[mansaxel]

In other news:

• The DM42 is a very, very desirable instrument. I recommend it.
• No, ABS impulse rings shall not be slotted, not even by corrosion.

I've counted lumps, and 1 r/s is 50 Hz. At about 200Hz the ABS self-test will run and conclude whether one's got a lump of rust or a detectable induction modulation device attached.  With 251mm circumference there's a new lump every ~5mm, and a 500µm crack (they expand because the cracks are due to rust swelling under the ring) is 1/5 of the lug length. (lug and valley are symmetrical) Definitely a bad day at work if you're to calculate speed on that, especially since you can't loiter and make averages for braking purposes. (obvious frequency counter reference here)

The most interesting thing is that I've been unable to auto-close my window while I've been driving with these rings. Now that there are new shiny ones fitted (and the ABS and traction control started working again without hard feelings) the one-click window closure works again. And the motor control warning lamp has gone dark.

[mnementh]

That charger is way less expensive than I'd have thought it would be. Might be worthwhile picking up a couple of them so you have a spare.

The charger is actually in the meter. This is just a AC/DC adapter. Supposed to output 12V 500ma.

I think he's looking ahead to the day when that cheap cable breaks internally right at the wall-wart, or more likely, right at the plug. By then, they will not be so readily/cheaply available I'm sure. Given the cost of the charger, I'd say his wisdom is self-evident.

OTOH... myself, I'd charge the thing fully and keep it on the bench for a while; see if I actually use it. The flip side of that wisdom is that would also suck to spend ~$50 on chargers only to find a month or three down the road that you don't like the meter and just want to flip it for a quick profit.   

mnem
 

[mnementh]

In other news:

• The DM42 is a very, very desirable instrument. I recommend it.
• No, ABS impulse rings shall not be slotted, not even by corrosion.

I've counted lumps, and 1 r/s is 50 Hz. At about 200Hz the ABS self-test will run and conclude whether one's got a lump of rust or a detectable induction modulation device attached.  With 251mm circumference there's a new lump every ~5mm, and a 500µm crack (they expand because the cracks are due to rust swelling under the ring) is 1/5 of the lug length. (lug and valley are symmetrical) Definitely a bad day at work if you're to calculate speed on that, especially since you can't loiter and make averages for braking purposes. (obvious frequency counter reference here)

The most interesting thing is that I've been unable to auto-close my window while I've been driving with these rings. Now that there are new shiny ones fitted (and the ABS and traction control started working again without hard feelings) the one-click window closure works again. And the motor control warning lamp has gone dark.

DAFUQ...?

So what... you figure both devices do a call on the same subroutine because sector/reluctor counting, and because one has an error condition, neither one works and the windows won't close to partway open automatically like mine? Or did you mean auto-close, as in the windows automatically close when you reach a certain speed?

mnem
 

[Ice-Tea]

Anyone with access to a N8974A, N8975A or compatible noise figure analyzer? Would like to test something, quite willing to pay...

[AVGresponding]

Temperatures of my gear to yours or others to differing rolls of filament make a difference so maybe run some smaller prints and bump the temperatures 5C at a time and look and get an idea of what is best for your roll and printer. Even in the past a filament brand change has made me need to do a tweak to mine.

With the brim normally just break it away and cleanup with a deburring tool is what I do but if it works for you then 

Ironically the flash is too strong to break off...

Looking at the spacers as the next piece prints, they aren't being infilled, which may be a large part of the problem.
I wonder if the nozzle temp can be increased on the fly... brb


EDIT: Yes you can, but too far into this print to risk it. Looking closely at the spacers, they really would benefit from some infilling to improve strength.

[McBryce]

Why did you add an adhesive flash to a huge flat model? It's really only needed for models that have a small footprint.

McBryce.

[AVGresponding]

Why did you add an adhesive flash to a huge flat model? It's really only needed for models that have a small footprint.

McBryce.

I didn't, that's just what it's printing.

[bd139]

China fun. These were only dispatched on Monday which was quite impressive. So £7.62 including delivery from https://www.ebay.co.uk/itm/254721451632

Much like the previous Bliley ones I ordered, the band saw is cheaper than desoldering them. I love how they blasted straight through the unfortunate ICs at the bottom.



Will desolder and test later. Fingers crossed!

Edit: minor rant. Stuff from China is taking 4-6 days at the moment. Stuff from EU, 2-3 weeks and 50% of it doesn't turn up 

[PA0PBZ]

Much like the previous Bliley ones I ordered, the band saw is cheaper than desoldering them. I love how they blasted straight through the unfortunate ICs at the bottom.

I wonder what a vibrating bandsaw does to the intestines

[bd139]

The Bliley ones I bought ages ago appear to have survived that. I suspect the reason they come in packs of two is because they have an expected failure rate and it's better to refund half of it and get some ok feedback than negative for a complete failure.

[Robert763]

@beanflying
I did not mean to offend you. My posts can be terse particuarly when I don't have much time. That is not a reason for you to be rude.
 Your original post incuded the suggestion from mcbryce to replace the socket with a standard one. The S10 plug you linked to is locking and used with a metal bush on the socket. Putting the matching socket on the meter would make it unsafe even when not being charged. I understand that was not your intention, but it is not clear and it is worth remembering that somone may find your post in the future and read it without context.
 
I'm not the "safety police" and have no power to stop people suggesting or doing unsafe acts. I do however have a moral obligation to comment when I see a post suggesting a clearly unsafe act. I am a professional engineer and it also a requirement of the code of conduct to address unsafe acts. If I comment people can make informed decisions.

I also have a different concern that one day a suggestion made on a forum (Cars are an other example) will result in an injury or death, a lawsuit and laws so no-one will dare make any suggestions. Far-fetched? Maybe, but who knows.

[mnementh]

Temperatures of my gear to yours or others to differing rolls of filament make a difference so maybe run some smaller prints and bump the temperatures 5C at a time and look and get an idea of what is best for your roll and printer. Even in the past a filament brand change has made me need to do a tweak to mine.

With the brim normally just break it away and cleanup with a deburring tool is what I do but if it works for you then 

Ironically the flash is too strong to break off...

Looking at the spacers as the next piece prints, they aren't being infilled, which may be a large part of the problem.
I wonder if the nozzle temp can be increased on the fly... brb   EDIT: Yes you can, but too far into this print to risk it. Looking closely at the spacers, they really would benefit from some infilling to improve strength.

This is a place where manually selecting your infill may be necessary. If I were doing this, I'd manually select GRID infill at 1-2mm spacing, then instead of skirt adhesion, I'd choose BRIM, 2-3mm wide, outside only, with MIN DISTANCE set at approx 1 or 2 line width, so 0.4mm or 0.8mm. This still helps adhesion with large prints like this, which are prone to curling/lifting due to walls at the perimeter, while being easy to remove.

Cheers,

mnem

[Cerebus]

I saw a yt vid at some point where they reheated the print in an oven, and it improved the strength considerably. I expect temperature and duration are both critical and found by experiment...

I've seen some numbers on that and the results are impressive - from cobbled part to something that you can reasonably ascribe some actual engineering properties to.

Two things to note for those unfamiliar with the technique, which involves heating the part buried in some supporting medium (usually table salt granules). You get shrinkage, so you have to make allowances in your original part sizing. The surface texture also changes, taking on the texture of whatever granular medium you've used to support it.

The material behaves much as polystyrene would.

What it drips, burning, onto your head?   
(When I were a lad, everybody seemed to have expanded polystyrene ceiling tiles. I still have nightmares about them.)

[SilverSolder]

The first part of the battery store was waiting for me when I got up. Middle section now printing.

Layer adhesion is clearly an issue on tall, thin parts, like the spacers, but I did expect that. I need to read up on the re-heating technique I think.

The electric fretsaw is ideal for trimming the flash as it turns out.

I wouldn't leave a 3DP on when you're asleep

This one looks like it would be described as "not working" on eBay! 

[Specmaster]

That charger is way less expensive than I'd have thought it would be. Might be worthwhile picking up a couple of them so you have a spare.

The charger is actually in the meter. This is just a AC/DC adapter. Supposed to output 12V 500ma.

Even, so it is still far cheaper than I'd have thought, especially as you have zero chance of using any other one you may have because of the special plug. It is a captive "walled garden" if you, and many other companies would seriously have milked that to death.

[Neomys Sapiens]

Another go at the DCR75, this time with a brawny MOSFET (2N7057)

The initial test result:
N-Ch Enhancement mode MOSFET
Red-G Green-D Blue-S
Vgs(on)=2,703V at Id=5,03mA and Ig=1µA
Vgs(off)=2,094V at Id=5,1µA
gm=39,1mA/V at Id=3,1mA to 5,0mA
Rds(on)<1.0O at Id=5,0mA and Vgs=8,0V
with body diode

So far so good. For MOSFETs, it offers only 2 graphs: Id over Ugs with stepped Uds and Id over Uds with stepped Ugs. Results below.

I have to say that it is a great semiconductor tester for repairs and sorting of parts, which it has in common with it's precedessor. It is not a proper semiconductor curve tracer per se. Among the greatest setbacks I've experienced is the missing possibility to de-clutter those graphs exactly where they are most interesting. Also, that you can't store the picture directly. Further, there is some ambiguous or even outright contradictory information in the manual. For example, the range for Ugs is given as 0..8V in one place and 0...10V at another.  Next, I'll have to look at the data behind the graphs, which can be exported into Excel. Maybe here lies the possibility to get more out of it. Overall, it is a nice device, but I am not yet sure if it can provide useful data beyond the quicktest.

[McBryce]

Why did you add an adhesive flash to a huge flat model? It's really only needed for models that have a small footprint.

McBryce.

I didn't, that's just what it's printing.

?? You choose whether it has that or not. It's an option in the slicer program.

McBryce.

[Specmaster]

China fun. These were only dispatched on Monday which was quite impressive. So £7.62 including delivery from https://www.ebay.co.uk/itm/254721451632

Much like the previous Bliley ones I ordered, the band saw is cheaper than desoldering them. I love how they blasted straight through the unfortunate ICs at the bottom.



Will desolder and test later. Fingers crossed!

Edit: minor rant. Stuff from China is taking 4-6 days at the moment. Stuff from EU, 2-3 weeks and 50% of it doesn't turn up 

Who'd have thought that Brexit would have such a negative impact  Mind you unless they get the Suez Canal unblocked real quickly the EU delivery might look more attractive 

[AVGresponding]

This is a place where manually selecting your infill may be necessary. If I were doing this, I'd manually select GRID infill at 1-2mm spacing, then instead of skirt adhesion, I'd choose BRIM, 2-3mm wide, outside only, with MIN DISTANCE set at approx 1 or 2 line width, so 0.4mm or 0.8mm. This still helps adhesion with large prints like this, which are prone to curling/lifting due to walls at the perimeter, while being easy to remove.

Cheers,

mnem

We have different machines, I don't have those infill options, only infill overlap and fill density % settings.

I've seen some numbers on that and the results are impressive - from cobbled part to something that you can reasonably ascribe some actual engineering properties to.

Two things to note for those unfamiliar with the technique, which involves heating the part buried in some supporting medium (usually table salt granules). You get shrinkage, so you have to make allowances in your original part sizing. The surface texture also changes, taking on the texture of whatever granular medium you've used to support it.

Yeah it's one to experiment with.

The material behaves much as polystyrene would.

What it drips, burning, onto your head?   
(When I were a lad, everybody seemed to have expanded polystyrene ceiling tiles. I still have nightmares about them.)

I was thinking more along the lines of solid polystyrene or acrylic.

Yeah, my dad had a hot-wire and would make fancy ones.

Why did you add an adhesive flash to a huge flat model? It's really only needed for models that have a small footprint.

McBryce.

I didn't, that's just what it's printing.

?? You choose whether it has that or not. It's an option in the slicer program.

McBryce.

Yeah I spotted it, the current print is running without.





EDIT: Sorry I don't have an HP 3457 to make a page header pic   

[mansaxel]

So what... you figure both devices do a call on the same subroutine because sector/reluctor counting, and because one has an error condition, neither one works and the windows won't close to partway open automatically like mine? Or did you mean auto-close, as in the windows automatically close when you reach a certain speed?

mnem
 

I guess that the ABS errors must have overloaded the fast CANbus (there are several, and ABS sits on the fast one) and ran a denial-of-service attack on everything else. Sadly, the motor warning light came back after I'd posted, so it's back to the original plan of swapping inlet cam timing solenoid.