TNP #3 - Zeiss Microscope LED Upgrade & Differential Interference Contrast (DIC) Microscopy Preview
Vložit
- čas přidán 29. 01. 2021
- In this episode Shahriar works on a Carl-Zeiss Axioskop with Differential Interference Contrast Microscopy. The instrument has bene meticulously restored to a working condition. The scope is retrofitted with an LED illumination source. Although not idea, it does offer good brightness and uniform illumination. An ideal source would have to be a point source.
The microscope is used to examine an integrated circuit revealing the intricate BEOL of the chip. Furthermore, DIC imaging shows the non-uniformity of the internal integrated circuit metal surfaces.
www.TheSignalPath.com
/ thesignalpath
www.Patreon.com/TheSignalPath - Věda a technologie
I wanna know where to find dumpsters like yours!
Would you comment on how SpaceX's starlink would fair in countries where it was 'banned'? Is the xciever easily detected/located?
Yep, this needs some explanation. I found this 5 grand scope in a dumpster ain't cutting it.
@@kevinbreslin5718Are you lost?
It's not the best idea to have a big area LED in this setup. Our aim (I'm a Zeiss developer) is to have a point like source as collimation is designed so that a point source optimally fills the back focal plane. Modern LED sources use fiber guides. However in the olden days indescandend bulbs needed to be aligned so that the filament is never reflected onto itself which would lower the lamp life. Also when aligning you should remove the objective and aim for uniform illumination with the adjustment screws on the lamp house. Usually you will see the filament and it's reflection when you hold a piece of paper in the back focal plane. With halogen lamp make sure to have filament and reflection on top of each other. This no longer applies for you but make sure even illumination when the objective is removed. After that you can proceed with Koehler illumination setup to get the best possible Abbe limited resolution.
Damn! It's always good to hear from a real expert, even if you are pouring oil on someone else's design. Any chance you could point Shahriar to the right parts?
@@michaeltonge1971 yeah sorry about the slight rant. It's probably not too bad at low magnification but if you are working with 63x immersion lenses alot of light is wasted and it gets pretty dark. Not sure but for the old pyramid style scopes might be LED retrofits available. But not from Zeiss.
@@LutzSchafer I don't think it's a rant. I think it's absolutely fantastic to have subject matter experts to provide information and I am sure Shahriar appreciates the comments too. This is what makes this channel and others such as Paul Carlson's or Dave Jones's for examples so educational and SO worth while. The info you provided may be helpful to others and passes on wisdom to all of us who have interest in such things!
I am aware of the limitations of the LED mod and the procedure of uniform illumination without the objective lens. Zeiss also has a few good videos on how to align the light, etc. which I studied before the mod. In fact, I emailed them to let them know I was going to make a video but never heard back unfortunately.
I had to work with what I had and the result is acceptable but not ideal. I aligned the LED with the existing optics as best I could, including the position of the LED with respect to the focusing lenses in the light source.
I forgot to add, thank you for the useful information and addition to the video! Having an expert comment is always a blessing.
Amazing! I love how you don't waste time on unnecessary things. With your level of skill, you could have designed your own PSU and buck converter to run the LED and fan, but you went with easy available solutions.
If thats 100W LED from ebay, I would have my doubts about longevity, efficiency and color rendering index
Such exquisite optics! Zeiss builds some wonderful optical systems. One of my expensive hobbies is amateur astronomy and astrophotography. I own a number of Astro-Physics telescope, which are just wonderful, including a 155mm (6") aperture apochromatic triplet (with hand figured, oil spaced lenses). So wonderfully color corrected, high contrast and a joy for planetary observations. Some years ago, during a Mars opposition, I had a rare opportunity to observe Mars through my scope's very limited production big-brother.. the 8 inch aperture version of my scope, another hand-figured triplet refractor. And to top it off, the owner had a Zeiss binocular view and Zeiss orthoscopic lenses. This was just about a religious experience! The Zeiss optics, both the binoviewer and eyepieces had very little scattering which presented such a wonderful high-contrast image.. Just a joy to behold!
The "problem" with optics like these is the same as test equipment -- there's this asymptotic curve when you plot cost against performance. Eeeking out that next few percent of performance costs many $$$ as your well up the steep part of the curve. One of my scopes, my pride-and-joy, cost more than the first car I purchased new.. it's hard to explain to normal people why you'd do such a thing!
Nice clean mod. Those objectives look really sharp.
why don't you make videos any more?
Relieved to see you alive and well! I was starting to assume the worst. Really hope to see you back on the platform, your stuff was the best
WE MISS YOU!!! Hope you are well!
Dude, more vids from you please. I love your conciseness and your cool narration voice and of course your very high s/n.
Carl Zeiss optics, nothing else is acceptable than the best.
Impressive! Congratulation with "the new" Axioplan! Thanks.
I love how the optics on this 'old' microscope are still magnificent. Great job bringing this back to life!
It's hard to believe the resolution of such a shallow depth of field with natural light illumination. Adding to the great history of optics!
Would love to see analysis of some active differential probe amplifier ASICs. Or high speed ADCs/DACs.
Great find and very interesting capabilities.
Awesome microscope images. And some one throw it away, but you found it and now you can make us drool at the chip images. More of them with your nice explanations. BTW, your explanations are what really make the difference. Thanks for the knowledge you gives us.
Wow. Nice score. The image is really sharp and no chromatic aberrations or vignetting. That is really good.
Amazing that you can see such details!
I like the "noise path" videos. Also, I am sure there is a noise path designed into all complex systems :)
How about decapping Starlink ASICs and looking under this beast!
I bet it would be so dense and on so many layers
How about just normal commonly used ASIC`s from reputable brands instead of some exclusive stuff only a minority uses......
@@jb5631 Well, that would be nice too. Fun is in seeing the internal organization ;)
Looks Great,
Wonderful!
wow! great stuff...cheers.
Great work! Especially imaging the surface roughness by manufacturing tolerances. You could use those images now and simulate it with FEM (e.g. HFSS, CST) to determine in detail the influence of the manufacturing tolerances on the results.
Oh yes that's mean great stuff are coming , I can't wait for reverse Engineering RF/high speed circuits at the die level .
Nice, I won a Nikon Optiphot 66 in a raffle at work a couple months ago, then spent money to get the parts off ebay to add DIC to it, it does both Diascopic and Episcopic light paths. Fun stuff, I put my Canon 5D Mark III and it's ok for this.
this looks like the predecessor to the confocal microscopy, they do some pretty amazing stuff with those, servos to move the optics around guided by software and you get a 3d representation of the stuff you're looking at...
regarding the halogen lamp - I think they use them still because of the emitted light spectrum, LEDs can have dips in the spectrum which may hide some detail in the sample
That is a very nice microscope, can you do a video about how components on an IC die look (so I know where I am looking at through my microscope) and the physics behind the shapes.
Great microscope find, and geat conversion!
Would a higher CRI led do anything or improve polarisation?
Saved from trash? Whenever I am at the local recycling site, I look for a Carl Zeiss phase contrast microscope, but all I can find is some crt TVs, broken laptops and a keyboard with what looks like cat vomit on it. I’ll just keep looking - might get lucky some day.
Thanks. Ha my school days when I searched for a microscope and swapped a model locomotive for a brass beauty!.
Wow! I had to build my equipment with parts I bought off of eBay. I can't imagine what the equipment you have cost when new. Anyway, I'd like to suggest you try out focus stacking. The technique allows you to build an image where pretty much every layer appears in focus. There are a few free focus stacking programs you can use to see if you like the technique. Some of these programs can also display the image in such a way as to produce a more 3D look. These tools are truly amazing.
Amazing picture quality, really pumped for more of this! Btw, who put "hugo" on the top metal layer in Katakana? Is this an easter egg from the manufacturer?
What are the squares around? Are those just to maintain uniform metal layer fill percentage for uniform etch, like on PCBs? Sorry if question is lame, never designed a chip, but did quite a number of PCBs.
They could be, but as some of them are not in a uniform pattern, like in concentric squares, I'm guessing not all of them are. They must be there to influence something at a lower metal layer, either capacitively, or inductively. When he focuses on the 3rd layer from the top, it looks like there starts to be tiny squares on a grid, possibly to meet minimum metal fill requirements. In CMOS, the minimum metal fill isn't only for uniform etch, but also to maintain a uniform topography - large areas with no metal will be slightly lower than areas that have metal, and that non-planar surface can cause issues at higher metal layers. At the highest metal layers, you can get away without having fills, as there aren't that many more metal layers above it, so your topography isn't going to get too uneven, but usually at the lower metal layers, you have to follow the fill rules.
Pashmaam!!!
Ohhh, I'm hoping that video means an upcoming dishy IC reverse engineering! :)
Is that round white cover on the front base foot a port for additional light source?
On mine (cheap Chinesium) it just a white disc painted black on the backside for image contrast?
That is for bottom illumination.
Could it be that one reason to use a halogen lamp is the very broad spectrum of light wavelengths emitted?
In order to kind of replicate that, using a high CRI LED would make sense. :)
great Shahriar , wonder which FAB you use for manufacturing your chip ?! how much did cost ?!
I liked digital powermeter.Where buy
this?
😲😲😲
Did you check the spectrum of the LED? What were the results? BTW: Can I recommend the Nikon D7100 as a replacement for the d700. I still own both. The D7100 has a smaller sensor but a higher pixel count and a much wider dynamic range. It's also great for astrophotography.
Do you know what frame rate it offers in live-view through the HDMI port?
@@Thesignalpath You could also try the sony A7RIII or A7II. Full Frame 60Fps HDMI live view. Any other mirrorless camera also will have high frame rate HDMI monitoring.
Probably better to use dedicated HDMI large-sensor camera for this microscope setup. I have D800 with fancy adapter on my microscope, but while HDMI output is 30 or 60 fps, focusing heavy camera on top and messing with batteries and controls is just inconvenient.
@@xDevscom_EE Which is why mirrorless cameras excell in this role. They are light, high fps live view, no messy focus, crazy high ISOs with very low noise with the sony ones, and many more benifits.
@@anonymousarmadillo6589 yeah why not use a Arri Alexa Mini LF??! I think a 2k+$ camera just for viewing through a microscope is a bit overkill
interesting
Finally THz and not that GHz rubbish.
Haha...
by the way common misconception it is not the oil it is the salt on your fingers, important difference
How education enriches life...
Yes it does. A great human jou.
Are you sure you have reflected light differential interference contrast setup? That is quite expensive stuff - not to be thrown away that easy.
Yes, absolutely.
@@Thesignalpath Do you sense the difference between polarized reflected light microscopy and DIC microscopy ?
Chip ID ICCTRDUP? CMC microsystems, TR = Toronto? All my chips in grad school were ICKAAxxx - university of Alberta, built in the Dalsa 300V CMOS process through CMC! That is a pretty nice microscope - even nicer than the one on the probe station I used in grad school. What really got me is the microscope we had for taking die photos for thesis or papers was really crappy! I tried to attach a camera to the probe station microscope, but we didn't have the right adapter, and my supervisor didn't want to spend money to get one made in the machine shop, so all my die photos in my thesis are terrible!
Yes, it is from University of Toronto. That is where I did my Ph.D.
Some LED makers offer very high CRI so that might be a future upgrade. . .
That is a good idea, I will look into it. :)
@@Thesignalpath Yuji and Soraa have interesting technology, although I don't know if Soraa have anything for this application. If you can underdrive the LED, you will get much longer and better light.
Look for yuji LED'S. They have a 500watt version that is tiny and fairly point sourced
Likely the illumination will not be as good with the LED due to the large source area. Also halogens are still mainstream on microscopes like this because of their perfect CRI. Looks great though
My thoughts exactly. "If it ain't broken don't fix it".I don't really understand the rationale for switching to LED in this case, did Shariar mention it? Improved brightness perhaps? This is not even guaranteed.
The Noise Path? It's a new brand name or some sort of new channel?
It is a subset of The Signal Path. I post short/update videos under this name.
Shahriar finds Ziess microscopes in his trash, Dave Jones finds industrial computers and flat screen TV's in the dumpster. In my trash I find, you know, trash. What am I doing wrong?
How much would a microscope like this cost?
Well, depends on how good of a shape it is in. But I guess less than 10K from eBay. (At least now that it has been fully serviced.)
With basic Acroplans, you can probably find one for 3000$ if you look often on ebay.
Nice scope. Even though you had removed the sensitive parts, I had to cringe a little bit when you said you took a grinder to a part of the microscope. You really should look into getting a mill. An old knee mill or used table top mill should not be too hard to find or cost too much.
The lenses were removed when I grinded the unwanted parts away. These parts are not precision parts, so they didn't have to be perfect. Indeed for more critical parts one should use a mill.
@@Thesignalpath Can you get a 3,000 lb Bridgeport in your lab? The oil and metal chips are kind of messy. Seriously, Shahriar, thanks for another great video. I have a Zeiss contrast microscope I thought could be used only for transparent specimens.
Wouldn't this video be called The Optical Path?
How can't I find something like this in trash can?😁
Updated your channel name?
It is a subset of The Signal Path. I post short/update videos under this name.
Not sure how you used the DIC with the epi illumination. You need the Optovar setting with the polarizer. What would be interesting for you is the Zernecke phase contrast. But for that you need PH (1...3) marked objectives. They have a semitransparent phase ring (quarter wavelength). And also a corresponding dark field aperture in the filterchanger I think. Nonetheless there is a "digital" equivalent image processing (TIE) that I developed for Zeiss. With that you need two images at different focus. Positions. The transport of intensity equation (TIE) algorithm transforms every camera pixel into its corresponding phase shift of the partial coherent light reflected from the sample. Maybe something of interest to you?
I would have loved to get a hold of a more modern unit from Carl Zeiss. I wrote them, but never heard back. :(
@@Thesignalpath That's a bit surprising. They are usually quick with quotations on potential customers. I don't know which area in the US you are to refer you. Here in Canada I kindof know the team although I actually have not much to do with them as I work remotly for advanced development Jena and Munich.
Regarding your Axioscope being from the 1990s it's a good universal instrument. Although you don't have any automation, you could create DIY solutions. Zeiss has a whole semiconductor family of instruments that are at TI, Infinion etc. That you possibly can't afford. Also there are systems dedicated to metallurgy. As I could see in your video, the objectives are not designed for reflected light metallurgy applications but rather for fluorescence widefield and transmission brightfield. But of course you can use them in reflected non fluorescent mode. To achieve the best theoretically possible resolution (1/(2lamda)) you would need to use coverslip and where necessary immersion oil. The metallurgy objectives are designed without that. Also mind you that Zeiss lenses start at $5k. The ones you are having already have a significant value especially the Fluar's.
Where is this mythical trash can? Please tell us
ヒューゴって？
That's a bit odd; it's labelled West Germany, and it's Karl Zeis - i must've missed a memo, because all i know is that their HQ is in the town Jena, which is near the middle of what was East Germany, eastern Thuringia. Now i wonder, did the Karl Zeis company maintain subsidiaries in both Germanies? (LOL, i've used a plural of a country * nerd giggles *)
after WW2 a part of the Carl Zeiss scientists got moved by the allied forces and a new branch of Carl Zeiss got started in "Oberkochen". During the time, where Germany was split there were two independant Carl Zeiss. Both led the market in their respective areas (western/eastern part of the world). In the 90s they got united. Today the HQ is in Oberkochen, but there is still a large location in Jena
@@matze1508 WOW, thank you the brief history lesson.
Why change from halogen?
Because the lights were hard to get a hold off and I wanted something with colder light colors.
@@Thesignalpath Perhaps adapting it to use an automotive headlight bulb would be the way to go? Apart from very common halogen bulbs, HID would be worth a consideration.
I don't find this kind of trash at my house.
Can anyone here point me in the right direction? I'm trying to buy used electronics testing equipment. Like, old surplus hardware at an auction or something. Don't know if I'm making sense buttt, help lol
I think the word you're looking for is eBay