EEVblog #992 - Olympus BHM Wafer Inspection Microscope

[EEVblog]

Unboxing an ebay Olympus BHM wafer inspection metallurgical microscope and getting it working after a custom Cree LED light replacement.

[Dave3]

Neat toy.

I would make an eBay claim. Broken light and related components. And given poor state of box, you should claim that the delicate optics need to be recalibrated.

Absolutely unacceptable packaging. At an absolute min, all electronics and equipment must be double boxed and way overfilled with foam.

I wonder if the original bulb provides significantly better performance than an optimized LED contraption would.

[amspire]

I don't know how precisely Olympus make the objective lenses for focus, but to keep focus when you change the lenses, have you tried adjusting for focus with the x40 lens by moving the table, and then adjust for best focus for the the x10 (or lower) by using the dioptre adjustments on the eyepieces and the camera positions? Go back and forward a few times.

Sometimes Olympus microscopes work the other way, so you could try the objective distance at x10 and the dioptre at x40. Probably can end up with at least two objectives in focus this way.

Richard

[Grapsus]

Damn you eBay shipping program !!! Two years ago I won a Fluke 343A voltage reference. It was shipped from the UK to France. The bare device was inside a paper box just rambling around and banging against the ground whenever the box was moved. They have no respect for precision equipment.

[Brumby]

Such packaging for any heavy bit of gear is totally unacceptable - let alone for a microscope such as this!

I noticed impressions on the inside of the top of the box - from the eyepieces it seems.  SERIOUSLY?

I would definitely be making noise about such an abysmal and totally, completely and hopelessly inadequate effort.  I've sent large ceramics interstate without ever getting a chip and while I don't claim my packaging is "professional" it certainly pisses all over that effort - but, then, that wouldn't be difficult.

This deserves a big "NOT ... HAPPY ... JAN!" and I would not be pushing for a claim on the broken lamp housing.  The break does look fresh - and the inadequate packaging offers no defence to damage in transit.

(For those unfamiliar with the phrase: https://en.wikipedia.org/wiki/Not_happy,_Jan! )



Oh, nice microsope 

[Brumby]

While we're on the subject, one has to wonder what other damage may have occurred that Dave is yet to discover.

For example - is the roughness of the stage movement something that existed before shipping?

[rs20]

Your description of what a Plan objective is subtly but importantly incorrect. A Plan objective has a good flatness of field, i.e., the "plane" of perfect focus is (close to) an actual plane. In simpler terms, when looking at a planar object (e.g. a silicon wafer), the periphery of a non-Plan objective would be out of focus, not just distorted. Or, if you adjusted the focus slightly, you could have the periphery in focus but the center would be out of focus. The defining characteristic of a Plan objective is that the entire view is in focus at the same time (again, only when looking at a flat object like a wafer).

An important distinction because a non-Plan objective's image cannot be corrected in post-production like a merely distorted image could be.

Source

Awesome video! Glad it worked so well!

[mdszy]

I can certainly see how he could make a claim for crap packaging... but would eBay/whoever require that it be returned? It seems like filing and claim and having to return it would be more trouble than its worth.

Very cool 3D printed LED mount, though!

[free_electron]

shove one of my wafers under it !  ( the tiny little loose preamps from the silicon fab episode. those are silicon germanium)

[hayatepilot]

That packaging was horrible!

They should have used something like this:

[rs20]

That packaging was horrible!

They should have used something like this:

Very impressive; but the fact I've never encountered such packaging makes me think it's prohibitively expensive. Probably much more practical to just train these people to pack things properly.

[LazyJack]

Dave, I'd hate to disappoint you, but this packaging is by the seller. Ebax global shipping generally do not repackage. They may open it for customs paperwork, but do not repackage. I have purchased several Texktronix CROs from the US, from professional used test equipment dealers and all came in rock solid packaging. In one case the box actually had the seller company's name on it, so it was definitely not repacked.
For these large items, the GSP has unbeatable price and it is fast.
For smaller items, not that cheap, but many times it is the only way to get something overseas.
And yes, i had cases, wheret the seller didn't care, and the stuff was just rattling around in the box.
You should have asked for a partial refund on the broken light.
Can the GSP impove? Yes. But it is unfair to blame lousy packaging by the seller on them.

[BU508A]

This is looking very similiar to the one from Dave:

http://www.ebay.de/itm/332212014062

[nctnico]

Dave, I'd hate to disappoint you, but this packaging is by the seller. Ebax global shipping generally do not repackage. They may open it for customs paperwork, but do not repackage. I have purchased several Texktronix CROs from the US, from professional used test equipment dealers and all came in rock solid packaging. In one case the box actually had the seller company's name on it, so it was definitely not repacked.

Yes same here. Ebay GSP doesn't repack. That would be way too costly and time consuming anyway!

[FloFo]

A few comments from someone doing microscopy (but not waffers) for a living and build his own experimental microscopes in the lab:
1) In the beginning dave said it's an stereoscopic microscope, which is off course not true. The Mantis would be stereoscopic, e.g. having two slightly different angeled beam paths for both eyes to give 3D perception. This microscope however doesen't have that, both eyes see exactly under the same angle (and i guess dave nows this). I assume he meant that it has a binocular tubus (which is better than a monocular tubus, you have to close one eye all the time which can be tiring). Unfortunately a mistake often made.
2) The fact that your Objectives are not parfocal (you have to refocus after switching the lens) is caused by wrong adjustment of the camera distance and eyepice distance. You should follow the following procedure:
a) set the eyepices to the right distance for your eyes
b) read the eyepice distance (in mm) at the scale above the eyepieces. The barrels around the eyepices have the same scale, set it to exactly these values
c) change to the highest magnification, focus on your sample. If you need glasses, do this step with the glasses or cantacts.
d) check if switching the lenses still requires focus adjustment. If so, slightly adjust the barrels of the eyepices till you get rid of this effect
e) when this works, focus on sample with high magnification and adjust camera to be focussed.
This procedure is quite complicated with this type of binocular tube, other manufactureres and later olympus scopes had other constructions that adjust the length of the optical path automatically when you change the width of the eyepices.
3) as mentioned before, it seems the table is slightly tilted. This could have happend in transitt, but can be solved by a skilled repairman (by shimming the table or slightly losening screws holding the table mount to the z  machanism and tapping it back). But the effect seems to be very slightly, so i wouldn't attemp to correct this. Chances are you make it worse.
4) Adapting a LED is a good way to upgrade old lightning. There are a few important notes:
a) you don't need much power (as you experienced). 1W is more than enough for brightfield and polarisation microscopy if adapted correctly (see below). Darkfield might use a bit more power, but dave doesn't have any of the necessary components anyway.
b) You should use a LED with a single emitter, you want to have a homogeneous light source
c) the size of the emitter should be the same as the size of filament in the old bulb
d) the LED surface should be at same position the bulb filament was in (and don't remove any lenses from the lamp housing). Idealy you should have an image of your light source in the back focal plane of the lenses (which you can check with special tools if necessary).
5) The optical quality of the middle class microscopes from india/china is quite good nowadays and can compete with the entry/mid level lenses from the big four (Nikon, Olympus, Zeiss, Leitz/Leica). If you need special lenses, you will end up with the big four (ans paying a few thousand bucks for one lens). The main disadvantage of the cheaper microscopes (like AM Scope) is the not so good mechanics and aviablility of aditional parts. The Olympus scope dave gought can be modifies for transmitted light, can do polarisation with a rotary table, Differential interference contrast (DIC), Phase Contrast, Fluorescence, darkfield, .... These microscopes are like a lego set, you have one frame and can add lots of different parts to create basically every technique you like.

[EEVblog]

Dave, I'd hate to disappoint you, but this packaging is by the seller.

No it wasn't, they had a shipping company pack it and send.

[EEVblog]

5) The optical quality of the middle class microscopes from india/china is quite good nowadays and can compete with the entry/mid level lenses from the big four (Nikon, Olympus, Zeiss, Leitz/Leica). If you need special lenses, you will end up with the big four (ans paying a few thousand bucks for one lens). The main disadvantage of the cheaper microscopes (like AM Scope) is the not so good mechanics and aviablility of aditional parts. The Olympus scope dave gought can be modifies for transmitted light, can do polarisation with a rotary table, Differential interference contrast (DIC), Phase Contrast, Fluorescence, darkfield, .... These microscopes are like a lego set, you have one frame and can add lots of different parts to create basically every technique you like.

I'd like to at least get a polarisation filter and some colour filters for it to see what they do.
Also, x60 and x80 lens might be useful to complete the set. x100 will likely need oil to work properly.

[FloFo]

5) The optical quality of the middle class microscopes from india/china is quite good nowadays and can compete with the entry/mid level lenses from the big four (Nikon, Olympus, Zeiss, Leitz/Leica). If you need special lenses, you will end up with the big four (ans paying a few thousand bucks for one lens). The main disadvantage of the cheaper microscopes (like AM Scope) is the not so good mechanics and aviablility of aditional parts. The Olympus scope dave gought can be modifies for transmitted light, can do polarisation with a rotary table, Differential interference contrast (DIC), Phase Contrast, Fluorescence, darkfield, .... These microscopes are like a lego set, you have one frame and can add lots of different parts to create basically every technique you like.

I'd like to at least get a polarisation filter and some colour filters for it to see what they do.
Also, x60 and x80 lens might be useful to complete the set. x100 will likely need oil to work properly.

As i said i'm not an expert with semiconductor microscopy, but i assume the polarisation filter will be useless. For polarisation microscopy you would combine two (one in illumination, one oriented under 90° in detection) to visualize birefringence of samples. A technique often used for mineralogy, but i don't expect birefringece from semiconductors.  From Color filters i don't expect much improvement either, but feel free to experiment with both

Regarding the lensen: Be aware to get the right ones. Most lenses for biology application are calculated for 0,17mm of glass between lens and specimen, and espacially with high magnifications (or more precisely high NA) your image gets bad if you don't have the right thickness of glass (which you normally don't have for semiconductors). So you want to use ones for metallurgical microscopes (cover glass = 0). Your existing ones are MPlans, with the M standing for Metallurgical. You can get metallurgical lenses with magnifications over 100 without oil immersion. I would recommend to stick with Olympus MPlan lenses, mixing lenses from different manufacturers (or same manufacturer and different eras) have some pitfalls, mainly switching for finite to infinite conjugates, mechanical tube length (finite) or focal width of tube length (infinite) and parfocal distances (distance from Lens Thread to sample, nowadays mainly 45mm, but older microscopes can have shorter ones). An overview on Olympus lenses of your microscopes era can be found in this document: http://www.science-info.net/docs/olympus/olympus-micro-optics.pdf

[LazyJack]

How do you know that? GSP works in a way that the seller packs the stuff up and sends it to the GSP center in the US. From that they handle international shipping. This is a convenience to the US sellers that they don't have to deal with customs labelling etc. And a convenience to the interanational buyers, because they can get the stuff at all.
The seller doesn't even know that it is going to be shipped outside of the US. So that is entirely possible that the seller used some crap shipping company to ship the microscope to the GSP center, but it was not arranged by GSP (Pitney Bowes), so it is not fair to blame them. They may get a blame on not noticing the substandard packaging, but after all, it is the sellers responsibility, to package properly and you can very well go for a refund if something is broken.

[Brumby]

Dave - a question:  Where did you get the "Angry Nerds" T-shirt?

I've been instructed to acquire one for someone in the household.

[Towger]

eBay GSP...  It serves a purpose, but I would always try and avoid it.  I have used it twice, the first time it was just being introduced in the UK and the seller (unknown to me) decided to use it to ship a couple of HV 2W resisters from the UK to Ireland.  WTF... A standard envelope with a couple of resistors, they took weeks to arrive via Norway, from memory.

The second time was more recent and despite the high costs, it was cheaper then ordering the parts direct from the official US supplier.  Who despite having the parts cheaper than on eBay, were looking 5 times their cost for FedEx shipping.

In both cases they arrived in the original packaging, in fact I think from the USA the parcel was placed in an additional eBay/PB bag.  In the early days of GSP a YouTuber did rant about it and said they were repackaging items to save cost/bulk, so the stories may stem from there.

BTW: The 'Part 2 Cleaning', while referenced is not available yet.

[jadew]

Nice. Looking forward for the comparison between the cheap ones and this one.

[free_electron]

5) The optical quality of the middle class microscopes from india/china is quite good nowadays and can compete with the entry/mid level lenses from the big four (Nikon, Olympus, Zeiss, Leitz/Leica). If you need special lenses, you will end up with the big four (ans paying a few thousand bucks for one lens). The main disadvantage of the cheaper microscopes (like AM Scope) is the not so good mechanics and aviablility of aditional parts. The Olympus scope dave gought can be modifies for transmitted light, can do polarisation with a rotary table, Differential interference contrast (DIC), Phase Contrast, Fluorescence, darkfield, .... These microscopes are like a lego set, you have one frame and can add lots of different parts to create basically every technique you like.

I'd like to at least get a polarisation filter and some colour filters for it to see what they do.
Also, x60 and x80 lens might be useful to complete the set. x100 will likely need oil to work properly.

My probe station had a mitutoyo microscope with a x200 lens ... no oil. MPLAN lens.
for lasercutting a had a x100 uv lens

http://www.mitutoyo.com/wp-content/uploads/2012/11/E4191-378_010611.pdf

i need to check if i still have some LCD fluid and apolarizer. you can do neat things with that.
lcd fluid 'toggles' under temperature. so if you ahe something that heats up inside a chip you make a setup where you can get that spot to trigger without catastrophic failure. put a blob of lcd fluid on the naked silicon. then beam polarized light onto it. turn the polarizer until the image is dark. the heated spot will 'flash' as it heats up because the polarisation of the lcd material changes under heat.

i did lots of troubleshooting like that.

[Brumby]

i need to check if i still have some LCD fluid and apolarizer. you can do neat things with that.
lcd fluid 'toggles' under temperature. so if you ahe something that heats up inside a chip you make a setup where you can get that spot to trigger without catastrophic failure. put a blob of lcd fluid on the naked silicon. then beam polarized light onto it. turn the polarizer until the image is dark. the heated spot will 'flash' as it heats up because the polarisation of the lcd material changes under heat.

i did lots of troubleshooting like that.

That sounds like fun!

[PartialDischarge]

Nice microscope, I have the Olympus model before camera add-ons were available and the optics are very clear

[Zom-B]

I've been using a multi ten-thousand dollar Carl Zeiss metallurgical microscope at work (inspection, photographing and fibbing preparations) and it has 5/10/20/50/100x optics. At 100x you're very close to the wavelength of light, and modern chips with features <500nm get inherently blurry and the only thing that works beyond that is electron microscopy. We almost never use 100x because with 50x you see about the same amount of details, and DOF is better.

The cool thing is, I was assigned to programming the XYZ stage control and camera capture for automated photographing and stitching of entire chips.

Here's a MEMS microphone from an old cellphone:



Pentium II Klamath (350nm process, 1997 (Ehm wow! 20yo)): (Maximum optical detail, DOF animation)

[S13]

those are some nice looking images!
Very interesting to see the MEMS with so much detail.
I own an old Olympus E series microscope, primarily for biological purposes, so ive never been able to view a silicon die up close this nicely.


To Dave: Swapping out the light bulb is probably the best thing you could have done anyway, broken or not. The crappy standard light bulbs are expensive to replace, dont have very accurate color representation and they warm up badly over time. So perhaps your courier did you a favor in this regard
The warming up from the light bulb is especially annoying if you are camera capturing a screen over a longer period of time, for instance when doing time lapses. The slow building heat from the bulb causes the frame of the microscope to warp excessively, making your picture go out of focus over time. When using a LED this problem is virtually eliminated.   

[FloFo]

Hi Dave,

after watching your latest videos on eevblog2, i think the tension adjustment for the coarse focus is set too low (resulting in the focus drift). Additionally, the too low table can be caused by a stage hight locking feature (could have been used by the previous owner to prevent someone crushing a lens into a sample).
How to use/adjust these features is shown on page 15 (or 16 by pdf numbering) of this BHM manual.

If you want to calibrate the magnification, you can get cheap stage micrometers for approx. 10-20 bucks on ebay (or sometimes amazon). Be shure to get one with 0,01mm divisions (there are some with only 0,1mm for stereoscopic microscopes).

Regarding your sloppy table, this seems not normal. Often there is some kind of friction adjustment on the coaxial knobs on the table, but i don't know olympus construction. With a good adjusted table, you can easily adjust samples with a 100 times lens with some practice, while being able to move larger distances quite fast if necessary

[EEVblog]

If you want to calibrate the magnification, you can get cheap stage micrometers for approx. 10-20 bucks on ebay (or sometimes amazon). Be shure to get one with 0,01mm divisions (there are some with only 0,1mm for stereoscopic microscopes).

Done:

[FloFo]

Magnification, not z-Scale
A Stage micrometer is a slide with a micrometer scale (e.g. this one on ebay) wich is used to calibrate the magnification of every lens (espacially with a camera, but sometimes also with eyepices with an integrated scale). The Magnification printed on the lens itself is not exactly but rounded to nominal values, e.g. i have a Zeiss 100x lens which i measured at 103,9x and a 63x measured at 64,1x. Additionally, camera adapters often have magnification (with rounding to nice values) too, and if you use the zoom of your camera, you can only guesstimate your magnification unless you do such a calibration.

[Fungus]

 

[Brumby]

Yeah .... I noticed that too.  That was obscene. (No other word for it.)

I had the same reaction.

[amspire]

I think I will have to add a micrometer to my microscope - looks like a great idea.

I think 0.01mm is enough resolution for my needs, so I am going for one of these:

https://www.aliexpress.com/item/Brand-New-Car-Tyre-Digital-Tread-Brake-Pad-Shoe-Gauge-Depth-Tester-Guage-Black-H7/32803113797.html

Only cost about $5.

I probably have to add a spring to push the tip downwards, but this kind of calliper sensor does usually have an internal digital output (a 24 bit easy-to-decode packet).

[S13]

After seeing the cleaning and servicing video im wondering if the horrible shakiness of the table is caused by the unstable stack of post-it notes between the XY-table and the wafer? Because normally, these olympus microscopes are rock stable thanks to their heavy solid metal frame.

Ideally you want your sample (the wafer) to be pinned down firmly on the XY-table. Please give that a try!

Glass slide preparations, as used for biological samples for example, are pinned down on the XY table by a spring loaded lever, and doing so myself ive never encountered the horrible shakiness that im seeing in the video. 

[FloFo]

Part of the problem ist the "havy" camera at a high point, connected only through the "thin" back of the microscope. I assume the effekt will be much better if you remove the camera and watch it through the eyepiece. This arrangement can work as a nice acoustic antena
I often work with samples just lying losely on the stage (but without PostIts) at magnifications up to 630x (63x times lens, 10 times eyepices) without any problems, using a Zeiss Axioplan on an table with a 20cm granite top (that's overkill, we just happend to have this one left over from something more sensitive). I think the main culprit is the table itself though, a microscope table should be very rigid and heavy (40mm particle board on a steel frame) and the legs should be properly adjusted to prevent wobbling.

Regarding you current source for the LED: good choise to use a linear regulator, with LEDs you can get problems with interference between shutter speed and switching frequencys. If you plan to implement a dimmer, try to avoid PWM for this reason

[jonovid]

the other microscope video by EEVblog #992 (Part 2) - How To Clean & Service A Microscope 


the other Microscope video by David and Dave -audio input . got to love the LED light bodge Job.  it's a hobbyist classic
is to avoid any one off pcb fabrication and just hot glue it. Hot Snot it!   
also IMO
it just goes to show why, a good consumer 3D printing service is useful for that one off 3D printing job.  no everyone needs a 3D printer at home.
 

[bktemp]

Does the LED current source actually work with the LED attached?
A white LED needs 3-3.3V. Using 5V from USB gives 1.5-2V for the current source. 1.25V are lost on the current sense resistor. So we have 0.25-0.75V available for the LM317 to work with.
I doubt it is able to reach the full current with this low voltage available. The datasheet recommends at least 3V input-output differential (most LM317 work fine down to at least 2V at lower currents).
In the video David only measured the current by shorting the output of the current source directly to GND (so with almost no voltage drop instead of the 3V for the LED).
A simple 5.6 ohms 1W resistor would probably be a better and even cheaper solution.

[jonovid]

too late to move EEVblog #992 (Part 2) post, to its own thread.

[mmagin]

A bit of bubblewrap around some equipment and the rest of the box filled with (or not quite) with packing peanuts is my least favorite experience with far too many ebay sellers.  It bums me out even if I'm getting something that's not functional to have it needlessly beat up in shipment.  A couple extra layers of bubble wrap taped around securely and some packing that doesn't migrate around as much as peanuts (like crumpled newsprint paper) isn't any more expensive but works so much better.

[mrpackethead]

Any chance you want to do a cleaning of a Mantis.. I'm sure mine is pretty dirty, but i'm not sure how to attack it.

Also, related.. can you take photos of what you can see in a mantis..  I've tryed, but just cant it to work.

[SeanB]

With cleaning glass it really is a case of "it depends" as to what you use. In general most optical glass will have a coating of one or more antireflective or filtering material on the surface, and these can be affected by many things, water included. Bare glass not so much, but those coatings are proprietary per lens manufacturer, and vary a lot. So, best is to start with the things that are least likely to damage, like a clean lint free absolutely dry cloth used to pat the surface ( not a wipe, as this will leave scratches) clean of dust.  then you can use distilled water, using another clean microfibre cloth as a wash. Alcohol ( 20% ethanol in water, also called 40 proof cane spirit, not Vodka or anything else, only the clear tasteless stuff ( Mainstay, Mainstay as a brand here) from a nip bottle still sealed) is the next, as it leaves no film, as the base ethanol is 99.6% ultra pure, and the water for blending is first DI and then RO filtered. This may affect some AR films, so the low alcohol content is good.

Do not use any stronger solvents, as many AR films are organic acids that dissolve or react with them. With many lenses you will also need to do the most horrid thing you can think of, and dismantle the lens out of the housing, so that you can take the lens and use a very destructive cleaning on it to kill fungus that has colonised the AR coating and is using it as food. With that your only choice is to strip the layer off, or kill the fungus with either bleach or peroxide, which can further damage the film. When doing the fungal kill, take the housing ( remove all the glass and other parts) and put the aluminium parts in a glass high temperature container, and boil on a stove to kill the spores in the pores of the metal. Plastic you have to chance the peroxide, and with compound lenses you also have a risk of the optical epoxy that glues them together being attacked, so you will need to contact the manufacturer as to which one they used, and if you can buy a lens off them if available. Otherwise you will have to dissolve the epoxy, split the lenses and clean them ( AR coating is pretty much going to be toast here) in strong solvents like MEK, and use some optically clear UV cure resin to put the kit back together. Will never be the same, but better to have at least clear glass as opposed to a growing fungal film that eats the other lenses.

[German_EE]

OK, David's first video:

8/10 for presentation skills, you're obviously treading in the steps of the master 

5/10 for editing technique. The constant editing especially at the beginning was really annoying and at one point we saw four edits in six seconds. That's way too much.

Having said that, David is a lot better in front of a camera than I will ever be and this is a great start. Well done!

[richnormand]

I have a similar setup Reichert microscope.
What I find very useful to examine wafers and ICs is two high quality polarisers.

One is inserted at the light source (you seem to have a place to insert filters as shown in the video).
The other one is in the light return path from the sample. Either one should be rotatable.
Close to 90 degrees it will greatly enhance any defect and change in surface contact.

[smashIt]

Just a few tipps from a fellow olympus-owner:

If you need any documentation for your microscope take a look at http://alanwood.net/photography/olympus/microscopes.html
He has collected a ton of manuals and even service-manuals.

Regarding a plan objective (and the "Plan" realy makes a difference at low magnifications)
There is one up on ebay at the moment:
http://www.ebay.at/itm/201931201422?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Alternatively you can convert your microscope to the current UIS2-system.
All you need is a little piece of aluminium (see attached picture).
With this you can mount UIS components onto your microscope.
But you have to stay in one system (all UIS or all 160mm).

[etelmo]

Did you regrease your microscope after you cleaned it?

Given the magnification some lash and play is to be expected (and vibration will be easily noticed) however it would be exacerbated if all of the old grease was removed in cleaning and none was reapplied. I don't know what product would be specified by Olympus however a tiny amount of NyoGel 774H (heavy) or 774VH (very heavy) on the rack would probably help, it's a pretty generic damping grease.

If someone knows the correct grease (does it need an extra heavy grade? or perhaps this type is designed for none at all?) it would be interesting to find out.