Watch out for fake 74LS logic ICs.

[Bud]

Jeesus Murphy, so much trouble just because you boys do not want to buy from an industrial distributor?...

[Fflint]

So here are the pictures of the die that is supposed to be SN74LS133N 82E24DKE4, also on the bottom of the chip plastic E9 is molded. It is supposed to be a 13 input Nand gate. It sure looks like one to me, but I have no idea about a brand. Do TI dies have a logo usually?

I don't think it's a TI part, because TI used aluminum wedge bonding, not gold ball bonding.

That sure is interesting.

One of your images is mirrored
Assuming this is a 133 type chip, pinout suggests that the first (darkfield) image is correct. So the marking reads:

Code: [Select]

S133
 Bwhich makes sense, I guess. Perhaps the S even stands for Schottky.

The circuit is definitely bipolar and looks like a likely TTL NAND gate.

Most TI dice have a logo: letters TI in rectangle. See here.
Not sure who made this. Are base areas of transistors visible under brightfield? I can't really see them, but it could be just the quality of your image. If they are invisible, that strongly hints towards modern Chinese production. Which is perhaps good news in a way, it means somebody is still making these things.

Is there some site with 74LS die photos(or any 74 series) ? I couldn't find one. 

Well, there is https://project5474.org but not a lot in there.

Indeed yes, the microscope mirrors it and I have to u mirror it in software. I took lots of photos (not knowing which one will come out OK). So I mistakenly added non-un-mirrored if it makes sense.

Areas of transistors are clearly visible under reflective brightfield (if this is the correct term). I speak more about the setup below. It looks a lot better by eye.

Here are two fragments of the same area zoomed in from full resolution with a little contrast and sharpness added before mirroring. You can just make out boxy outlines around transistors. Click on an image for a bigger version.



This is generally a biology transmission microscope, with a led light from another microscope(much less magnifying) used for dark field. The first image in the previous post  is dark field (light source positioned on top at an angle). The pinkish tinted image is an attempt at reflective brightfield by shining a led light into one eyepiece while looking through the other (it is an old Zeiss Letzlar microscope made in 1950s).

Jeesus Murphy, so much trouble just because you boys do not want to buy from an industrial distributor?...

Aren't you curious how they do it and remain profitable? (previous post about them having a huge bowl of solder and paying peanuts for labour notwithstanding). I recently heard average labour rate in China now is double that of Mexico(no idea if true). And we have no(up to my knowledge) fake chips from Mexico!

If this turns out to be the real deal from TI, then it is even more puzzling. It lacks the dice logo though so who knows. I would rather not sacrifice a real TI chip to find out(perhaps I will of there'll be no other way, but the 74ls to chips I have aren't 133s so I'm not sure how useful it's going to be. )  If anyone has a genuine TI made 74LS133 and wants to sacrifice it for decapping (and wants to post it to me in Poland) msg me for the address :-)

Going that far (sanding relabelling, fixing leads) just to sell it as new when they could easily sell same 74ls133 as used to me for the same price (I would probably prefer old tested due to less chance of scam on used parts).

I thought this reads 2133, but I agree it looks like S133B - bipolar so definitely not HC ergo LS, but made by whom? There is always a possibility it is a western made 74ls133 made by another manufacturer pulled from old hw and some person thinking TI chips sell better so they went to change the markings. Probably the simplest answer is the correct one.

BTW, for anyone curious. You don't need nitric acid for decapping. Concentrated sulphuric acid will do, but it has to be hot (over 120C roughly). It has to be done in a fume hood or outside because it gives off noxious gasses. Also, don't heat the acid and plop the chip in. Put it in cold and slowly heat in a beaker on a hot plate. After 20min let water run through the black goop that results until it clears. Then throw it out onto a filter paper and look for your prize. Finally clean with isopropyl alcohol and a small brush. Physical cleaning is required as no amount of stirring will clean specs of black goop from the die. Overall a much more approachable process than dealing with hot nitric acid (potentially fuming). If I couldn't do it with sulphuric I would give up. I don't know why online sources prefer nitric.

Edit I was thinking about this letter B being there. I looked for anyone with 74ls133B chip, but only SGS Thomson made one (74ls133b1) and that b1 was for plastic package. There is no sense putting package code on the die... So perhaps it is a clue someone can decipher.

[Fflint]

I thought this reads 2133, but I agree it looks like S133B - bipolar so definitely not HC ergo LS, but made by whom? There is always a possibility it is a western made 74ls133 made by another manufacturer pulled from old hw and some person thinking TI chips sell better so they went to change the markings. Probably the simplest answer is the correct one.

It seems like you are not aware that the original '133 part was the 74S133.  The 74LS133 came much later.  Have you measured the supply current of your parts?

I haven't yet. I will in a couple of days. Yes, I knew there was 74S133 (in TI' s "digital logic, pocket data book" for example), but I didn't connect the dots with this.

In the meantime I looked for any datasheets from manufacturers that made this chip and I found TI, Motorola, ST (SGS Thomson I guess is the same thing, is it?) and that's it. Interestingly on TI pages if you go into products they don't even list 74ls133. There are lots of other 74ls products, but no 133. However, the S version is present in their book mentioned above.

I wish someone wrote an article about the detailed history of simple logic before CMOS with die photos.

Also I wonder if any an help with reverse engineering the schematic. Could someone comment on the below:

These are clearly bipolar transistors. Right?


And this I assume is a diode.


But then what are these?


And finally this. If anything? Maybe these are just connections.


When I have more time I might send this die through an ultrasonic cleaner. I read online it helps. Also in future I might be interested in perhaps grinding the top surface down of this or similar chips slowly to reveal hidden layers. Anyone has an idea how thick those layers are on early pre-CMOS logic and how many there may be? I have a surface grinder precise enough to take off a sharpie mark from really flat steel without touching said steel(a micron or so) so with a diamond wheel it might work. However work holding this die and aligning it may be difficult.(encasing it in something would be obvious, but diamond wheels don't like epoxy..) Also I have no arbor for the diamond wheel yet (it's on a long list of upgrade projects). That's why I said "in future".

[magic]

I don't think it's a TI part, because TI used aluminum wedge bonding, not gold ball bonding.

That sure is interesting.

It's not, though, because Noopy showed an old genuine TI 7400 with ball bonds.

Areas of transistors are clearly visible under reflective brightfield (if this is the correct term).

Nope, the bases aren't visible. Look at Noopy's 7400, the bases are red. You can clearly see one patch of base area with all the input emitters on it and a link to the pullup resistor, which is also made of a long and narrow "base". This is not visible on your chip and I don't think we have ever seen any classic bipolar chip from TI/ST/Motorola/Fairchild/Philips/etc which was like that. Smells like China to me.

BTW, there is a schematic on Noopy's 7400 page. Hopefully you can figure out the difference between 2 and 13 inputs yourself

The pinkish tinted image is an attempt at reflective brightfield by shining a led light into one eyepiece while looking through the other (it is an old Zeiss Letzlar microscope made in 1950s).

Yeah, I was wondering if it's like that. It's not too bad, but maybe not optimal yet because there is some unevenness. Are you placing the LED exactly in the exit pupil of the microscope?
If not sure what this means, set it up for transmitted observation, pull out one ocular to confirm that exit pupil of the objective is filled 100% with light (open the aperture diaphragm as necessary), reinstall the ocular, hang a paper above it and look where you can see a sharp image of the objective's exit pupil. This image determines the optimal location and diameter of your LED. Fairly obvious if you think about it.

When (ab)using biological objectives for reflected illumination, also make sure they have matte black backs, no shiny metal parts surrounding the pupil. Fairly easy to fix with matte electric tape if necessary.

BTW, for anyone curious. You don't need nitric acid for decapping. Concentrated sulphuric acid will do, but it has to be hot (over 120C roughly). It has to be done in a fume hood or outside because it gives off noxious gasses. Also, don't heat the acid and plop the chip in. Put it in cold and slowly heat in a beaker on a hot plate. After 20min let water run through the black goop that results until it clears. Then throw it out onto a filter paper and look for your prize. Finally clean with isopropyl alcohol and a small brush. Physical cleaning is required as no amount of stirring will clean specs of black goop from the die. Overall a much more approachable process than dealing with hot nitric acid (potentially fuming). If I couldn't do it with sulphuric I would give up. I don't know why online sources prefer nitric.

Because common 65% HNO₃ at 120°C produces same result in 5 minutes and without black goo. Bonding pads are completely etched away, though.
Because FNA at 70°C or so (never done it) produces again the same result, but without bonding pad corrosion at all. The part remains functional.

Online sources also say to clean with acetone if preservation of bonding pads is important. In my experience, H₂SO₄ damages them slightly by itself, even without water cleaning. I have seen claims that corrosion can be avoided with short bath in acid pre-heated to 200°C or more.

Also I wonder if any an help with reverse engineering the schematic. Could someone comment on the below:

These are clearly bipolar transistors. Right?

Yeah, these look like NPNs, unconnected of course. This silicon is likely capable of being several other ICs, depending on the final metal layer on top.

Also in future I might be interested in perhaps grinding the top surface down of this or similar chips slowly to reveal hidden layers. Anyone has an idea how thick those layers are on early pre-CMOS logic and how many there may be?

Nothing interesting in there. It's silicon with different dopants diffused into it (invisible without staining with nasty chemicals), some silicon oxide or nitride (transparent), some metal for connections, some more oxide or nitride. The oxide and metal layers can be cleanly removed chemically if desired.

The oxide layer itself contains most useful information, because its thickness depends on the number of processing steps received by given area and this determine the color of reflected light by thin film interference.

You may also want to take a look at this thread
https://www.eevblog.com/forum/projects/decapping-and-chip-documentation-howto/

[amyk]

That's definitely bipolar, but here's another simple tip to distinguish them without decapping: both TTL and CMOS families will have diodes between Vcc and inputs, but they point in opposite directions; the input diodes on TTL point towards the inputs, while those of CMOS are protection diodes that point toward Vcc.

[Fflint]

I don't think it's a TI part, because TI used aluminum wedge bonding, not gold ball bonding.

That sure is interesting.

It's not, though, because Noopy showed an old genuine TI 7400 with ball bonds.

It took me a while to figure out what is it you were talking about... For anyone else reading this. There is a user on this forum with a nick of Noopy, he has a personal blog/site where he has lots of cool die photos and stuff. Here is the page magic is referring to (I think)

https://www.richis-lab.de/logic14.htm

So those ball bonds remove the only definitive proof of fakery we found so far (gold wires balled instead of aluminium wire bond)

Areas of transistors are clearly visible under reflective brightfield (if this is the correct term).

Nope, the bases aren't visible. Look at Noopy's 7400, the bases are red. You can clearly see one patch of base area with all the input emitters on it and a link to the pullup resistor, which is also made of a long and narrow "base". This is not visible on your chip and I don't think we have ever seen any classic bipolar chip from TI/ST/Motorola/Fairchild/Philips/etc which was like that. Smells like China to me.

I wouldn't put a lot of importance at colors on my pink-violet tinted photos. Because of the way I have to take them there is a lot of Illumination light that bleeds back into the "receiving" eye piece. It takes a lot of fiddling to get them even as "good" as shown. I'll refer more to it lower(perhaps I can use the advice given). Also there are other color differences in certain areas between that noopy's photo and mine that suggest to me either his is some special way of photography with false color, or my colors are bad due to that over bleed of the Illumination light that comes off as pink-violet. Why is it pink-violet? I'm guessing it is a reflection from coatings on all the optical components. (those you can see with the same pink-violet tint when looking at certain lenses on binoculars etc - I always thought this is there to improve chromatic abberations, but I may be wrong).

BTW, there is a schematic on Noopy's 7400 page. Hopefully you can figure out the difference between 2 and 13 inputs yourself

The pinkish tinted image is an attempt at reflective brightfield by shining a led light into one eyepiece while looking through the other (it is an old Zeiss Letzlar microscope made in 1950s).

Yeah, I was wondering if it's like that. It's not too bad, but maybe not optimal yet because there is some unevenness. Are you placing the LED exactly in the exit pupil of the microscope?
If not sure what this means, set it up for transmitted observation, pull out one ocular to confirm that exit pupil of the objective is filled 100% with light (open the aperture diaphragm as necessary), reinstall the ocular, hang a paper above it and look where you can see a sharp image of the objective's exit pupil. This image determines the optimal location and diameter of your LED. Fairly obvious if you think about it.

When (ab)using biological objectives for reflected illumination, also make sure they have matte black backs, no shiny metal parts surrounding the pupil. Fairly easy to fix with matte electric tape if necessary.

Thanks, I didn't know getting rid of the reflection of the objective front surface is important. Also for how to find the optimal location of the led light.

Unfortunately at the moment I have no way to precisely position the light other than holding one of my two led torches to the other eyepiece (touching it or slightly backing off). I can also remove the eyepiece and point the light down it's tube.

My microscope's aperture diaphragm is not in the light path when using the episcopic (through eyepiece and the ocular) Illumination. The microscope is laborlux 3 (1954) with "Leitz brightfield condenser 600"under the table. The diaphragm is there to be used with a built in light source that is in the base. By saying it is setup for transmission microscopy (is this the right term?) I meant the light is on the bottom, then there is the condenser with diaphragm, then the table, objective revolver, then a head unit with two eyepieces. (a head unit with a tube for episcopic Illumination exists for this model, but it is rather expensive).

To achieve Illumination through the objective, which seems the only way to resolve those tiny height differences in the oxide layer. Unfortunately the extraneous result is this pink-violet hue. If anyone has any idea how to get rid of it, please let me know. Perhaps I should try laser Illumination with a webcam, or try making polarizing filters.

BTW, for anyone curious. You don't need nitric acid for decapping. Concentrated sulphuric acid will do, but it has to be hot (over 120C roughly). It has to be done in a fume hood or outside because it gives off noxious gasses. Also, don't heat the acid and plop the chip in. Put it in cold and slowly heat in a beaker on a hot plate. After 20min let water run through the black goop that results until it clears. Then throw it out onto a filter paper and look for your prize. Finally clean with isopropyl alcohol and a small brush. Physical cleaning is required as no amount of stirring will clean specs of black goop from the die. Overall a much more approachable process than dealing with hot nitric acid (potentially fuming). If I couldn't do it with sulphuric I would give up. I don't know why online sources prefer nitric.

Because common 65% HNO₃ at 120°C produces same result in 5 minutes and without black goo. Bonding pads are completely etched away, though.
Because FNA at 70°C or so (never done it) produces again the same result, but without bonding pad corrosion at all. The part remains functional.

Online sources also say to clean with acetone if preservation of bonding pads is important. In my experience, H₂SO₄ damages them slightly by itself, even without water cleaning. I have seen claims that corrosion can be avoided with short bath in acid pre-heated to 200°C or more.

I saw no difference between acetone, or isopropyl alcohol, and even good old water when cleaning. The only benefit of acetone is that it dries faster perhaps.

Also I wonder if any an help with reverse engineering the schematic. Could someone comment on the below:

These are clearly bipolar transistors. Right?

Yeah, these look like NPNs, unconnected of course. This silicon is likely capable of being several other ICs, depending on the final metal layer on top.

Could you explain what you mean by the "final metal layer on top"? I can clearly see aluminium metallic traces. The colors are true in the dark field photo I attached (the dark one). There you can clearly see the grey, slightly reflective sheen of aluminium. Perhaps including a zommed in bit will help.


I used concentrated, 96% sulphuric acid, I didn't add any water to it. It shouldn't touch metals by much. Where gold balls are missing they were ripped during physical cleaning.

Also in future I might be interested in perhaps grinding the top surface down of this or similar chips slowly to reveal hidden layers. Anyone has an idea how thick those layers are on early pre-CMOS logic and how many there may be?

Nothing interesting in there. It's silicon with different dopants diffused into it (invisible without staining with nasty chemicals), some silicon oxide or nitride (transparent), some metal for connections, some more oxide or nitride. The oxide and metal layers can be cleanly removed chemically if desired.

The oxide layer itself contains most useful information, because its thickness depends on the number of processing steps received by given area and this determine the color of reflected light by thin film interference.

You may also want to take a look at this thread
https://www.eevblog.com/forum/projects/decapping-and-chip-documentation-howto/

Thanks.

Edit: Having spent some time looking at all the logic ICs photographed on the website linked above (Richie's Lab) I too am leaning towards thinking this is some old Chinese production IC. All TI chips I saw there have ti logo on the die. Other western manufacturers had more extraneous stuff on their dies. More numbers, letters, mask alignment features. Smaller feature sizes.

This one seems a bit spartan in comparison. So I wouldn't be surprised if it was some Chinese make.

[magic]

Also there are other color differences in certain areas between that noopy's photo and mine that suggest to me either his is some special way of photography with false color, or my colors are bad due to that over bleed of the Illumination light that comes off as pink-violet. Why is it pink-violet?

Different chips made in different fabs end up with different color.
It's determined by thickness of the transparent oxide layer (and angle of reflection, but that's fixed at 90° in this case).
Same principle as oil spills on water appearing in random colors.

I'm guessing it is a reflection from coatings on all the optical components. (those you can see with the same pink-violet tint when looking at certain lenses on binoculars etc - I always thought this is there to improve chromatic abberations, but I may be wrong).

AR coatings are there to reduce reflections off lens surfaces and increase transmission of the objective.
Pretty much all optics have them, objectives purpose made for episcopic illumination even more so. They are not a problem.

Metal traces on your image are reasonably neutral grey, so it isn't a serious problem for you either.
Objectives may have slight color cast and "white" LEDs too. Camera white balance can fix it most of the time.

If anyone has any idea how to get rid of it, please let me know.

Try again with an IC which isn't violet

Thanks, I didn't know getting rid of the reflection of the objective front surface is important.

Ideally everything inside the scope should be black to absorb unintended stray light rather than reflect it towards the image.
The edges of the objective are particularly at risk of being illuminated if you don't align your setup perfectly.
This was not a problem in transmitted light applications so some old objectives have bare metal there.

Unfortunately at the moment I have no way to precisely position the light other than holding one of my two led torches to the other eyepiece (touching it or slightly backing off). I can also remove the eyepiece and point the light down it's tube.

I suppose removing the ocular would only increase light pollution inside the scope. And dust ingress.
My suggestion is to use the ocular to project an image of the LED precisely into the pupil of the objective, and only there.
I think a round THT LED (3mm or 5mm) would be good fit, depending on ocular magnification. Make a mounting attachment for it out of cardboard/plastic/etc.
You can pull the other ocular to inspect where the light went, make adjustments as necessary.

My microscope's aperture diaphragm is not in the light path when using the episcopic (through eyepiece and the ocular) Illumination.

Of course it isn't and the built-in illumination is useless for IC imaging.
I only mentioned it as a way of locating the real image of the pupil through the ocular.

Could you explain what you mean by the "final metal layer on top"? I can clearly see aluminium metallic traces.

Exactly what I meant.
This layer of aluminium traces is fabricated at the end of the process "on top" of the silicon to provide connections between circuit elements.
Hence the same silicon may sometimes become a few different circuits, depending on metal configuration.

[rsjsouza]

The "S" letter is very similar to the Signetics logo. They had a 74LS133 on their product line back in 1978.

Page 204:
http://w.bitsavers.org/components/signetics/_dataBooks/1978_Signetics_TTL_Data_Manual.pdf

[Fflint]

The "S" letter is very similar to the Signetics logo. They had a 74LS133 on their product line back in 1978.

Page 204:
http://w.bitsavers.org/components/signetics/_dataBooks/1978_Signetics_TTL_Data_Manual.pdf

Thanks :-)

I wish there was a site which collected die photos of same chip made by different manufacturers for simple logic. No doubt as those chips become more collectible over time people will be even more interested in identifying them.
edit that's what  https://project5474.org is :-)

Metal traces on your image are reasonably neutral grey, so it isn't a serious problem for you either.
Objectives may have slight color cast and "white" LEDs too. Camera white balance can fix it most of the time.

You can see how much extra color there is by looking at an area around the chip. It is a white double sided sticky tape. It should be white, but it appears dark blue/violet. Most likely this color is reflected from the chip.

My suggestion is to use the ocular to project an image of the LED precisely into the pupil of the objective, and only there.
I think a round THT LED (3mm or 5mm) would be good fit, depending on ocular magnification. Make a mounting attachment for it out of cardboard/plastic/etc.
You can pull the other ocular to inspect where the light went, make adjustments as necessary.

My microscope's aperture diaphragm is not in the light path when using the episcopic (through eyepiece and the ocular) Illumination.

Of course it isn't and the built-in illumination is useless for IC imaging.
I only mentioned it as a way of locating the real image of the pupil through the ocular.

I see. I haven't got any white THT leds, just some smd ones. I might make something to hold a little pcb. Although now I'm exploring the possibility of adding a diy "vertical light" to my microscope. The binocular head is removable. It should be possible to fabricate an attachment that will move the head let's say 20mm higher and use the space to integrate a episcopic light from a diy 3d printable microscope project like PUMA (Maybe polarizing filters too).

Could you explain what you mean by the "final metal layer on top"? I can clearly see aluminium metallic traces.

Exactly what I meant.
This layer of aluminium traces is fabricated at the end of the process "on top" of the silicon to provide connections between circuit elements.
Hence the same silicon may sometimes become a few different circuits, depending on metal configuration.

OK, I see.

[Fflint]

Here is a slightly better photo I took with light reflected off the face of the objective showing silicon features of a few transistors. Here you can clearly see green tint and even larger orange tint around the base below.

I think this is the best it can get without extra vertical Illumination. I took that photo by using a powerfull led lamp in place of the original light, covered by a piece of thin paper, then using the original diaphragm and condenser. The ic die obscured direct light, but surrounding light reflected off the objective face illuminated it from the top. Unfortunately I have no dslr so I can't use a method used by Nooki for example. This one has a slight greenish tint unfortunately.



Having seen all other die photos I still suspect Chinese production. I'll update this thread once I have measurements of current consumption and input voltage sweep.

[magic]

You can see how much extra color there is by looking at an area around the chip. It is a white double sided sticky tape. It should be white, but it appears dark blue/violet.

Okay, I see it. Not gonna lie, I sometimes get similar color cast if I use a crappy LED. I also have one ruskie objective which is somehow more blue than others. But it can be tweaked away with WB controls on the camera. I'm yet to see something so bad that even red or green turns blue.

I see. I haven't got any white THT leds, just some smd ones. I might make something to hold a little pcb. Although now I'm exploring the possibility of adding a diy "vertical light" to my microscope. The binocular head is removable. It should be possible to fabricate an attachment that will move the head let's say 20mm higher and use the space to integrate a episcopic light from a diy 3d printable microscope project like PUMA (Maybe polarizing filters too).

SMD is not a problem, though many may prove too small. Maybe illuminate a small piece of ground glass with LEDs from behind.

DIY attachment is an option too. In fact, I have bought an incomplete biological scope without any head at all and frankensteined this tower on top of it. This contains a DIY illuminator, appropriate extension tube and a camera (the Raynox "tube lens" got removed).

Remarks:
Smartphone mirror foil sucks, I'm looking for something better. There is noticeable loss of resolution when this mirror is inserted (while other lighting is used).
I do use an extra lens to project an image of round LED into the objective, bouncing off the angled mirror.
Some black paper in strategic places helps. I hope to post more about it in the future.
Not sure how this compares to using the beam splitter of a bino head. I don't have a bino head.

[magic]

Here is a slightly better photo I took with light reflected off the face of the objective showing silicon features of a few transistors. Here you can clearly see green tint and even larger orange tint around the base below.

Yes, that's nicer. You are right, base regions are distinct now. Maybe it can be a recycled Western chip after all.

I think this is the best it can get without extra vertical Illumination. I took that photo by using a powerfull led lamp in place of the original light, covered by a piece of thin paper, then using the original diaphragm and condenser. The ic die obscured direct light, but surrounding light reflected off the objective face illuminated it from the top. Unfortunately I have no dslr so I can't use a method used by Nooki for example. This one has a slight greenish tint unfortunately.

That sounds like about the same technique Noopy developed, but used with higher resolution optics. So what the heck.
I tried it too with mixed results. Oftentimes there is not enough light passing around the die and still hitting the objective. I guess SLR lenses have an advantage here.

Another option is angled transparent mirror between the specimen and the objective, but it causes quite bad blur even with 10x0.25. IIRC it was passable with 5x0.12.

[Fflint]

You can see how much extra color there is by looking at an area around the chip. It is a white double sided sticky tape. It should be white, but it appears dark blue/violet.

Okay, I see it. Not gonna lie, I sometimes get similar color cast if I use a crappy LED. I also have one ruskie objective which is somehow more blue than others. But it can be tweaked away with WB controls on the camera. I'm yet to see something so bad that even red or green turns blue.

The really bad colors resulted from using two cheap battery powered torches. One (the slightly more expensive one) even caused photos with visible vertical bands of light-dark (probably due to the low led drive frequency). There was an immediate improvement when I tried to use another led light that was much warmer, but heavy misting was present regardless of light position(through eyepiece). The original light that came with the microscope is incandescent so I wouldn't be surprised if the optics was optimised for this band somehow.

I see. I haven't got any white THT leds, just some smd ones. I might make something to hold a little pcb. Although now I'm exploring the possibility of adding a diy "vertical light" to my microscope. The binocular head is removable. It should be possible to fabricate an attachment that will move the head let's say 20mm higher and use the space to integrate a episcopic light from a diy 3d printable microscope project like PUMA (Maybe polarizing filters too).

SMD is not a problem, though many may prove too small. Maybe illuminate a small piece of ground glass with LEDs from behind.

DIY attachment is an option too. In fact, I have bought an incomplete biological scope without any head at all and frankensteined this tower on top of it. This contains a DIY illuminator, appropriate extension tube and a camera (the Raynox "tube lens" got removed).

Remarks:
Smartphone mirror foil sucks, I'm looking for something better. There is noticeable loss of resolution when this mirror is inserted (while other lighting is used).
I do use an extra lens to project an image of round LED into the objective, bouncing off the angled mirror.
Some black paper in strategic places helps. I hope to post more about it in the future.
Not sure how this compares to using the beam splitter of a bino head. I don't have a bino head.

Good idea about the ground glass, I've been using paper, but I have a sander to I can sand/grind a piece of glass  easily.

Cool idea with your custom head/photo adapter in the thread you linked.  One thing worth mentioning is that back in the 1900 when people were inventing microscopy they routinely used cover slips angled at 45 degres as beam splitters. If I can't get a tiny rod mirror, I might try that.

The main problem with illuminating via one eyepiece of a bino head I have is misting (and glare, but that can be fixed a bit by positioning/intensity). I think it must be due to multiple prisms in the light path and lots of unwanted reflections. Also, some if the light (half?) is probably passing right from one eyepiece to the next.

I have lots of broken Gu10 7W,600lm,4k K, LED lights (I collect them until I have 20 or so and then I repair them en-mass) so I pulled out the little aluminium board from one. I replaced the burned out led and I powered it with a bench psu at 60V (it was running at about 3W, not enough to heat a lot, but still lots of light. This is the light I used in the base. Until I fabricate the proper vertical light I might play with this a bit.

Specifically, the chip's die is roughly a mm square, the white sticky tape holding it is maybe 0.3 mm bigger. So there is lots of light that goes around it that enters the objective. I have to try putting a piece of black backing behind it sized same as the objective lens so no strong direct light enters it. To get nice colors the light has to be very intense, but the slightest amount hitting the objective directly bounces around the scope on the inside and produces misting (despite Zeiss blacking it pretty well internally).

I think this is the best it can get without extra vertical Illumination. I took that photo by using a powerfull led lamp in place of the original light, covered by a piece of thin paper, then using the original diaphragm and condenser. The ic die obscured direct light, but surrounding light reflected off the objective face illuminated it from the top. Unfortunately I have no dslr so I can't use a method used by Nooki for example. This one has a slight greenish tint unfortunately.

That sounds like about the same technique Noopy developed, but used with higher resolution optics. So what the heck.
I tried it too with mixed results. Oftentimes there is not enough light passing around the die and still hitting the objective. I guess SLR lenses have an advantage here.

Another option is angled transparent mirror between the specimen and the objective, but it causes quite bad blur even with 10x0.25. IIRC it was passable with 5x0.12.

I read about Noopy's technique. If I had a dslr and some objectives I would do the same.

When searching online I found this open source 3d printed microscope project called PUMA (Anyone looking for it just Google it). They have lots of cool attachments and the arrangements /parts they use are all great as a starting point for commercial modular scope mods. I think their vertical light is especially interesting. They use a beam splitter in the tube between objective/head. The light source is a single white led plus a condenser made with two cheap lenses. I think I might try a cover glass as a "poor man's beam splitter".

[wraper]

This thread proves yet once again that buy dodgy parts from China if you want to tinker with those parts instead of making things.

[magic]

No joke, they truly are the cheapest shit out there for decapsulation and chip imaging R&D

One (the slightly more expensive one) even caused photos with visible vertical bands of light-dark (probably due to the low led drive frequency).

Yeah, rolling shutter effect. For me it's 5V and 100Ω. These dice are very reflective, so it doesn't take a lot of light at all with illumination through the lens.

The main problem with illuminating via one eyepiece of a bino head I have is misting (and glare, but that can be fixed a bit by positioning/intensity). I think it must be due to multiple prisms in the light path and lots of unwanted reflections. Also, some if the light (half?) is probably passing right from one eyepiece to the next.

Start with basics: a well defined uniform circle of light, projected by the ocular exactly into the glass of the objective.

If there is still glare then yes, it's insufficient absorption of light in the beamsplitter, nothing that can be fixed without surgery, if possible at all. Taking the simple example of angled mirror foil (or cover glass as you suggest), some light fails to reflect down to the objective and passes through to the other side. You need to absorb that light well, because anything that returns to the mirror can be reflected upwards and mixed with the image. Similar effect likely occurs inside those bino heads somewhere. This is not a normal light path for them.

My solution so far: use a piece of black paper, angled at 45° so that most of the reflection goes sideways. Maximize the distance from the mirror to the absorber to maximize light loss by scattering in random directions.

When searching online I found this open source 3d printed microscope project called PUMA (Anyone looking for it just Google it). They have lots of cool attachments and the arrangements /parts they use are all great as a starting point for commercial modular scope mods. I think their vertical light is especially interesting. They use a beam splitter in the tube between objective/head. The light source is a single white led plus a condenser made with two cheap lenses. I think I might try a cover glass as a "poor man's beam splitter".

Thanks, will look it up. Their illuminator seems similar to mine, but I found a way to do it with one lens. But I don't have a variable aperture diaphragm - it's fixed by LED diameter. Not a big deal, because diaphragm setting must match exit pupil diameter rather than NA of the objective, and that varies little with magnification.

edit
Inserting anything below the head increases primary image distance, and if your scope is not infinity corrected, spherical aberration correction in the objectives will be thrown out of whack. It shouldn't be a problem with 10x0.25, but not sure about more powerful optics like 40x0.65 or oil if you fancy going that far. It is my understanding that original attachments of that sorts contained extra optics which compensate for this. Magnification increases too. Again, not an issue in infinity scopes.

[Fflint]

Inserting anything below the head increases primary image distance, and if your scope is not infinity corrected, spherical aberration correction in the objectives will be thrown out of whack. It shouldn't be a problem with 10x0.25, but not sure about more powerful optics like 40x0.65 or oil if you fancy going that far. It is my understanding that original attachments of that sorts contained extra optics which compensate for this. Magnification increases too. Again, not an issue in infinity scopes.

That's not good :-( I was hoping to use it primarily with my 40x objective (and perhaps with the 100x oil one). For other objectives like 10 (and even 3) I could figure out alternative arrangements as the distance between the sample and the objective is many mm.

I doubt my laborlux 3 is infinity corrected, because it was made in 1954. Also the head has 1.5x written on it and it contains a lens in addition to prisms and eyepieces. This is useful info. If I choose to go down this route I have to weight the potential increase in image quality due to better lighting with deterioration due to new spherical abberations. I imagine I can limit them by attempting to insert as small amount of extra height as possible. With a rod mirror I could probably manage to do it all in 5mm. (total length is 150mm). A rod mirror is just a very thin rod with its end cut at 45deg and polished. It will obscure some of the light, but not a lot.

This thread proves yet once again that buy dodgy parts from China if you want to tinker with those parts instead of making things.

I can't say I disagree...

BTW, Is anyone aware of any software (preferably open source / free, but knowing about commercial options would be nice too) that can do automated image stitching and focus stacking at the same time? The "at the same time" is the key requirement. There is plenty of tools that can do one or another with limited alignment capabilities. However, yesterday I took a number (about 50) photos of that die with 3~4 of each area intending to focus-stack and stitch them, but there are few issues. Just focus stacking photos covering same area results in photos that are slightly out of focus in corners, those corners often overlap other areas that are on very good focus. If there was some software that can do both a panorama stich and a focus stack at the same time, it could make a nice very high resolution picture from it. There is a lot of detail on those photos, but they are all generally half in half out of focus due to the tilt of the chip. Of course it is possible to generate a focus map for each picture. Make it into an alpha channel and then stitch it manually all in gimp, but it sounds like hours of work...

Two examples of those photos :

[magic]

My goal is to avoid stacking by having everything aligned. Flat die, flat slide glass, flat stage.

I'm almost there, but the stage is misaligned somewhat. I plan to insert corrective shims under it, but I want to do it together with other mods/repairs and I'm still gathering parts and procrastinating around it a little

That's not good :-( I was hoping to use it primarily with my 40x objective (and perhaps with the 100x oil one). For other objectives like 10 (and even 3) I could figure out alternative arrangements as the distance between the sample and the objective is many mm.

There is a simple test: raise the ocular in the head tube by the same amount, refocus, see if it made a difference.
150mm sounds like a 160mm mechanical tube length system, so finite. 160 should be printed on the objectives.
As far as I'm concerned, 10x0.25 is good enough for most bipolar tech.

Not sure how a rod mirror could work as a splitter? The image forming light needs to pass undisturbed...

[hanakp]

@Fflint: I have Tesla MH74ALS02 from 1980s, PM me if you're interested.

https://www.teslakatalog.cz/MH74ALS02.html

Many such Tesla-made ICs are also available on Ebay, I doubt anyone would go though the trouble of faking those. I alone own literally thousands of MH74xxx TTLs, including industrial-temp MH84xxx and even some military-spec MH54xxx. But they're so cheap now it's not worth the hassle putting them up for sale.

Oh, and in case the non-european crowd wonders, Tesla was a czechoslovak state-run electronics company:

https://en.wikipedia.org/wiki/Tesla_(Czechoslovak_company)

[magic]

PUMA microscope epi illuminator and related parts:

https://youtu.be/cAEB10K8PqI
https://youtu.be/uB48D5KueU8

I watched most of both videos and agree with most everything there. Similar design to mine.

Interesting ideas for light traps.
Polarization is something I haven't considered but looks potentially interesting.
I wouldn't bother reproducing their Nelsonian illumination optics, Köhler is the way to go.

I have an idea how to do almost perfect Köhler setup with one lens and two stops (variable or fixed) for aperture and field control.
Preliminary prototype shows that the field stop does provide some degree of improvement.

The illumination path is as follows:
beam splitter ← converging lens with a few cm focal length ← field stop in the image conjugate plane ← aperture stop in the pupil conjugate plane ← uniform diffuse light right behind the aperture stop

The conjugate planes are guaranteed to come in this order as they are on different sides of the focal plane of the converging lens.
Two final elements are currently replaced with an LED of appropriate diameter.

[Fflint]

My goal is to avoid stacking by having everything aligned. Flat die, flat slide glass, flat stage.

I'm almost there, but the stage is misaligned somewhat. I plan to insert corrective shims under it, but I want to do it together with other mods/repairs and I'm still gathering parts and procrastinating around it a little

I see. With a 10x objective my current chip seems flat enough, but when I go to 40x (let alone 100) it is then I discovered it is not flat at all. I would be interested in finding out if there are any ways anyone might have to flatten it. The plastic is completely gone. The die lies on its clean back on a microscope slide already. Also I found I can get better image quality by adding a drop of microscope oil on the due and a cover slip on top, but of course the cover slip is not flat either so this adds to the alignment problem.

That's not good :-( I was hoping to use it primarily with my 40x objective (and perhaps with the 100x oil one). For other objectives like 10 (and even 3) I could figure out alternative arrangements as the distance between the sample and the objective is many mm.

There is a simple test: raise the ocular in the head tube by the same amount, refocus, see if it made a difference.
150mm sounds like a 160mm mechanical tube length system, so finite. 160 should be printed on the objectives.
As far as I'm concerned, 10x0.25 is good enough for most bipolar tech.

Not sure how a rod mirror could work as a splitter? The image forming light needs to pass undisturbed...

I'll try the head raising test.

Unfortunately for taking photos I'm using a mobile phone, also due to the way I'm illuminating the chip if there is light passing around it it will make the mobile phone photo unusable. So the die ideally should fill the entire field of view. That is "almost there" with 10x, so I have to use 40x. Also there is so much more detail visible with 40x...

As for the rod mirror, it can't function as a splitter, it is a simple mirror, but it is positioned far from any focus point. My understanding is obscuring, let's say 5% of the light path with it will not cause the image to be much worse.

PUMA microscope epi illuminator and related parts:

https://youtu.be/cAEB10K8PqI
https://youtu.be/uB48D5KueU8

I watched most of both videos and agree with most everything there. Similar design to mine.

Interesting ideas for light traps.
Polarization is something I haven't considered but looks potentially interesting.
I wouldn't bother reproducing their Nelsonian illumination optics, Köhler is the way to go.

I have an idea how to do almost perfect Köhler setup with one lens and two stops (variable or fixed) for aperture and field control.
Preliminary prototype shows that the field stop does provide some degree of improvement.

The illumination path is as follows:
beam splitter ← converging lens with a few cm focal length ← field stop in the image conjugate plane ← aperture stop in the pupil conjugate plane ← uniform diffuse light right behind the aperture stop

The conjugate planes are guaranteed to come in this order as they are on different sides of the focal plane of the converging lens.
Two final elements are currently replaced with an LED of appropriate diameter.

This went over my head a bit, nevermind. Regarding polarisation I ordered two cheap photography polarisation filters to do an experiment with my "through eyepiece" Illumination.

[magic]

I see. With a 10x objective my current chip seems flat enough, but when I go to 40x (let alone 100) it is then I discovered it is not flat at all. I would be interested in finding out if there are any ways anyone might have to flatten it. The plastic is completely gone. The die lies on its clean back on a microscope slide already. Also I found I can get better image quality by adding a drop of microscope oil on the due and a cover slip on top, but of course the cover slip is not flat either so this adds to the alignment problem.

Cover slip is necessary for biological objectives more than ~0.4NA and it's a PITA because any junk on the die, bonding balls etc, causes it to tilt.
As for the rest, I think my slide glasses are flat enough because I can rotate them 180° and nothing changes. This hints that my problem is in the stage itself.

Common 10x0.25 achromats deliver plenty of detail already but it may take high mag oculars to see. I find 20x0.4 to be acceptable without cover glass too, but they are rarer.
I use direct projection onto a USB camera image sensor. I see a small crop of the FOV (don't need to care about plan ) and pretty much the maximum resolution of the optics.

This went over my head a bit, nevermind. Regarding polarisation I ordered two cheap photography polarisation filters to do an experiment with my "through eyepiece" Illumination.

It's fairly simple stuff. You could probably understand more just by looking up the definitions of all that jargon.

Unfortunately, I haven't had much time to play with it (and still don't) so this project is stuck for now. I hope to build something more permanent than a paper prototype and show how to reproduce it.

I think bino-epi is a good starting point, your images aren't too bad. If polarizing filters could improve contrast further, that's a win.

[Fflint]

I played with the polarisation filters a bit. To be honest I expected a bigger difference, but there is some difference. It seems height differencec in the die are slightly enhanced and it becomes apparent flat surfaces aren't ideał (a diffraction like pattern - rainbow of colors) can be seen in few places. Mind that the best photos I took were with the die submerged in microscope oil (these are not the photos with the entire die visible, I'm talking about individual transistors etc). While today I kept the die dry, because I had only few minutes of time to spend on it. This is the same reason why I haven't taken any photos. It is possible to hold a light, and two filters in one's hand while observing, but adding a mobile phone to this is impossible (not enough hands..). So I would need to fabricate some adapters/holders.

I have more interesting chips coming (no - not from China... These will be Polish CEMI made logic and hopefully Czech Tesla chips). So I'll fabricate those adapters, but it'll take some time and it will go into it's own thread (or into an already existing thread we have here).

I've been thinking what is the best photography device accessible to me (no DSLR) and I decided in favor of medium resolution and very good dynamic range. I have a Cctv Camera board that uses a Sony Imx335 sensor. It is a 5M pix sensor with very good dynamic range. I plan to make an adapter for it to mount it on the eyepiece. Pehaps with no lens, using direct imaging like magic. Maybe 10x objective will be good enough for me too then. A led light woukd be mounted to another eyepiece.

I might be able to remove "misting" in software. By taking a picture with no specimen and then subtracting it from the picture of interest.

I have lots of ideas, I wish I had more time in a day to try them all :-)

[philpem]

Indeed yes, the microscope mirrors it and I have to u mirror it in software. I took lots of photos (not knowing which one will come out OK). So I mistakenly added non-un-mirrored if it makes sense.

Areas of transistors are clearly visible under reflective brightfield (if this is the correct term). I speak more about the setup below. It looks a lot better by eye.

Here are two fragments of the same area zoomed in from full resolution with a little contrast and sharpness added before mirroring. You can just make out boxy outlines around transistors. Click on an image for a bigger version.

That "S" looks a lot like the Signetics logo. I can't find any evidence of them producing the 74LS133, but it could be a 74S133:

https://en.wikipedia.org/wiki/Signetics

https://www.datasheetarchive.com/pdf/download.php?id=0720bef359401ea49557d1f057a04258271e46&type=M&term=74S133


The giveaway would seem to be the power consumption. The 74S133 pulls between 5 and 10mA - http://www.bitsavers.org/components/signetics/_dataBooks/1978_Signetics_TTL_Data_Manual.pdf (page 206)
Whereas the Motorola 74LS133 pulls between 0.5 and 1.1mA: https://pdf1.alldatasheet.com/datasheet-pdf/view/5654/MOTOROLA/74LS133.html

[rsjsouza]

That "S" looks a lot like the Signetics logo. I can't find any evidence of them producing the 74LS133, but it could be a 74S133:

Check reply #57 above from me.