Transistors - die pictures

[T3sl4co1l]

I think I've seen Vgs(max) by characteristic (zener voltage) rather than rating before, but I can't recall a specific example so it's probably been a while -- uncommon.

They should do fine, being small zeners, but beware the voltage might not be well controlled, the current rating quite small, and the breakdown voltage is usually pretty high besides, and you might not want to idle the gate at 20V or whatever, when all you need is 5 or 10V.

Tim

[David Hess]

I wondered if it was Control Data Corp? but they didn't use house-labelled that I could find. I got them from a surplus store around 1976, they had bushels and couldn't give them away almost. Some on mainframe looking huge PC boards.
Minimum Cost Semiconductor Silicon Survey (April 1969) of Fairchild TO-105, TO-106 parts. 2N5136 $0.11 "Not recommended", " need not supply a data sheet for publication... none of the readers of 73 would be interested." 

That is what I was thinking; maybe Fairchild house numbered them for Control Data Corp.

I just checked and the small white CDC ones that I can find immediately are marked CDC CS2923 and CDC CS2924, and also I have these part numbers as 2N in the older "top hat" style TO-92, and probably in real TO-92.  The all black ones seem to all be Fairchild, and some of those are also marked Singapore.  All appear to have date codes through the 1970s.

Somewhere I am sure I have some of the larger packaged white CDC ones.  I might have pulled all of the white CDC ones off of forgotten boards, but many appear new with the original lead lengths.

Here we see a transistor built by Continental Device Corporation similar looking to ours with a "CDC" on the side of the white ceramic:

https://archive.org/details/bitsavers_ElectronicignV15N1419670705_116901282/page/72/mode/2up?q=%22Continental+Device+Corporation%22

That resolves who CDC is, but I notice something else in that advertisement which explains why CDC is not listed in my transistor D.A.T.A BOOK.  That advertisement also shows RTL ICs with the Fairchild part numbers and packaging.  Is there any reason not to think that CDC was simply reselling Fairchild products with a CDC house numbers and markings?

I did not know that the Electronic Design trade magazine went back that far.

[exe]

I don't think anyone ever actually put any rating datapoint on them; max shunted current/energy, reverse leakage, etc.

No, and I'd say most transistor parameters are loosely specified. Like, leakages, or threshold voltage.

My expectation is that  VGS(MAX) should be safe to apply for extended time, and built-in zenner should not leak more than maximum leakage current over the whole temperature range. However, I'd also derate gate voltage for two reasons: 1) I don't know if this is harmful or not (some DS have disclaimer that working at maximum specs reduce reliability, but others don't have it) 2) transients etc can make it exceed the specs.

On the other hand, zener should start conducting below dangerous for mosfet voltages, right?) So, if gate current is limited, then it should be safe at all times). But I wouldn't bet on this.

[Noopy]

Regarding the protected MOSFETs:
It´s always the same: If it is not specified in the datasheet you can´t be sure. 
Nevertheless I would expect that the abolute maximum rating of Vgs should be possible without relevant leakage current.
For the PMPB19 it´s not a bigger problem. The absolute maximum rating is +/-10V. That is pretty low. I assume the Gate can withstand 15V or even 20V. Between 10V and 20V there is enough room for zener breakdown voltage tolerance and leakage current.
More interesting is something like the STY60NM60 (https://www.st.com/resource/en/datasheet/sty60nm60.pdf): Absolute maximum rating Vgs=30V! 
But here the datasheet states that the zener doesn´t break down below 30V and up to this voltage up to 1mA leakage is allowed.

That resolves who CDC is, but I notice something else in that advertisement which explains why CDC is not listed in my transistor D.A.T.A BOOK.  That advertisement also shows RTL ICs with the Fairchild part numbers and packaging.  Is there any reason not to think that CDC was simply reselling Fairchild products with a CDC house numbers and markings?

Well that´s of course possible. Perhaps these transistors are Fairchild parts with a CDC branding.

[floobydust]

Transistor reliability was poor back then, Fairchilds known to fail if you tapped them on the side... it must have been interconnects/bonding wires or something.
As well as the die design and fab not performing or competing well. Motorola was giving them a very good run back then. Competition was fierce.
So they would likely have a sales/surplus channel to sell off the poor performers at discount pricing. Perhaps CDC was that.

[SeanB]

Found a board with those small white ceramic transistors, with IC's on it (TTL 54 series) that are dated 1977. White ceramic are marked E111, dated 7726, next to a LM218 from National, dated 7809, and an odd transistor from Vactek, partial part number VTL2??? wonder what that 4 leg device is, and looked it up and it is an optocoupler, LED driving a CDS cell, looking like an IR silicon LED.

https://www.datasheetarchive.com/pdf/download/distributors/Datasheets-8/DSA-152236.pdf?h=838accbe3c51ebcbb1e90291f22d175b%3A6eb27fe37141c7c2ba00cc294ac7594ee0%3Ae0393f1b8a637e5dd423f997a5749c5e

Interesting little unknown part. Should see about popping it into an envelope if anybody wants to decap it. The R&S boat anchor it comes from is both a dead item, and was heavy. But really interestng design wise though, and dating from around 1979, though it looks like there were some infant mortality on a lot of IC's, with them having 1984 to 1986 date codes, and not mil spec or industrial range on random ones that likely failed. Looks like the LM1558 IC's all failed in warranty, all changed out with 1980 date codes. Must have had a leaky batch, or they all got purple plague.

[Noopy]

This VTL2 part sounds interesting. I can take a closer look at it if you want me to... 

[SeanB]

Well PM me a postal address and I will drop it into the black hole that is called SA Post office. Might be there possibly Christmas 2025.

[Noopy]

Done!
I will wait patiently. 

[Noopy]

The Signetics SD304 is a dual-gate n-channel MOSFET. It is a DMOS transistor. The gate electrodes are protected against overvoltage by Zener diodes. In addition to the SD304, the datasheet also lists the SD303, the SD301 and the SD300. These are probably different bins. The SD304 can be used at frequencies up to 1GHz. It blocks up to 25V and conducts up to 50mA.




Several scratches and remnants of bondwires can be seen on the bottom of the SD304 package.






The edge length of the die is 0,50 mm. Square structures in the corners make it easier to place the masks. The auxiliary structure in the top left-hand corner makes it possible to check the alignment of the masks against each other after production. On the right edge, numbers appear to be shown in different levels. However, the characters are so small that the process could not work them out cleanly. There is a string of characters in the upper area. This could be an internal project name.




The MOSFET structures are clearly visible. In the center is the drain potential, which is surrounded by two gate electrodes. The outermost ring transmits the source potential. The two protective diodes are integrated at the right-hand edge.

The structures are similar to those in the SM200 (https://www.richis-lab.de/FET45.htm), but here the additional strips that increase the active area are missing. This has a noticeable effect on the transconductance. While 12-24mS are specified for the SM200, the SD304 is specified with just 10mS.


https://www.richis-lab.de/FET48.htm

 

[Noopy]

Like the SD304, the Signetics SD305 is a DMOS transistor with two gate electrodes. It is specified for applications in the range of 200MHz. It is therefore somewhat slower than the SD304. The maximum reverse voltage is 20V. The drain current may increase up to 150mA. In addition to the SD305, the datasheet also lists the SD306. The specifications of the components show that they are not just different bins. The SD306 has a lower current carrying capacity, lower parasitic capacitance, a lower transconductance and a higher channel resistance. Everything indicates that the SD306 contains a significantly smaller transistor.




The source potential of the MOSFET is connected to the package via two bondwires.






The cross-shaped structure of the MOSFET offers a significantly larger active area than in the SD304. This explains the higher current carrying capacity and the higher transconductance, but also the higher parasitic capacitances. The auxiliary structures in the corners and at the upper edge are the same as those found in the SD304.




The slightly thicker metal strip on the right-hand source bondpad does not appear to have any electrical function. It could be an auxiliary structure that facilitates the correct alignment of the die. The Zener diodes, which protect the gates against overvoltage, are integrated near the bondpads.


https://www.richis-lab.de/FET49.htm

 

[Noopy]

The above advertisement is from the magazine Electronics (May 30, 1974). Unitrode advertises a fast power switch for use in switching regulators. There are a total of eight variants, four for positive currents and four for negative currents. The freewheeling diode of the switching regulator is already integrated in this component.






The Unitrode Semiconductor Data Book from 1980 describes even more variants. Whereas in 1974 only reverse voltages of 60V and 80V were available, in 1980 there were already variants with a reverse voltage of 100V. The permissible currents could be selected between 5A, 15A and 20A. The 5A and 15A variants are in TO-66 packages with four connection pins. The packages of these types are insulated. The 20A variant, on the other hand, uses a TO-3 package with three pins. There, the package is connected to the reference potential. A closer look at the specifications shows that the 15A variants may be permanently loaded with 15A, while the 20A variants are the peak current. No more than 15A may flow continuously here either.

The components contain an NPN power transistor and a driver transistor. The variants for negative load currents use NPN driver transistors and are therefore Darlington transistors. The variants for positive load currents use PNP driver transistors and are therefore Sziklai transistors. All variants also contain a freewheeling diode. The variants shown here use pn diodes. There are also variants with Schottky diodes (PIC730, PIC740). By matching the transistor/diode combination, the current commutation between the two elements is cleaner and cause less noise.




The component shown here is a PIC647. The marking is barely recognizable. The date code refers to the year 1984.






The PIC600 variants are hybrid components. The TO-3 package contains a large heatspreader on which a disk of beryllium oxide is placed. All active elements are located on the ceramic disk.




In order to be able to transmit the high currents, several bondwires lead from the connection pins to the ceramic carrier. The traces on the back of the pins show that the wires have been electrically welded there. Bonding processes are otherwise usually carried out by friction welding.




The PIC647 is one of the Sziklai transistors. The NPN power transistor is located in the upper area of the ceramic carrier. The significantly smaller PNP driver transistor is located below the power transistor. The base-emitter resistors were applied to the ceramic carrier as thick-film resistors. The freewheeling diode is located on the left-hand side. A metal block on the heatspreader enables a bond connection to the package.

The input pin is connected to the collector potential of the power transistor via two bondwires. Four bondwires had to be attached to the output pin, as two end directly on the emitter of the power transistor and the freewheeling diode also had to be contacted.






The edge length of the power transistor is 2,7 mm. The structures of the base and emitter are not unusual. However, it is noticeable that the transistor structures do not extend to the lower edge. There is obviously a second, smaller transistor in the lower area, which is inactive here. Apparently, the design offers the possibility of building a Darlington transistor. If a differently structured metal layer is applied, the small transistor can be used as a driver for the large transistor.




Outside the active area, the base area ends surprisingly quickly.




Where the base surface is contacted, the metal surface has become discolored. There was probably a strong build-up of heat there. A discoloration of the silicon can be clearly seen in the right-hand area. Perhaps a breakthrough occurred at this point.






The edge length of the driver transistor is 0,77mm. The picture above clearly shows that it is a transistor with a so-called perforated emitter. The base area under the emitter surface is contacted via small, square windows in the emitter surface. It is not a large power transistor, but the base current can still be up to 400mA.






The edge length of the freewheeling diode is 2,8 mm. The upper right area appears to be damaged.




The resistors are ordinary thick-film resistors.


https://www.richis-lab.de/BipolarA62.htm

 

[iMo]

That transistor would cost you an arm and a leg today..

[floobydust]

HP gave it a part number, and there is some aerospace mention but no idea what it might have ended up in.
HP 1826-1080, PIC647, NSN 5962-01-249-2101.

HP 1826-1079, PIC602 still in production by Microsemi (Microchip), special order price via quotation only at Mouser, Digikey etc.  must be a military application being supported.

[Noopy]

Mouser states 462€ for the PIC647:

https://www.mouser.de/ProductDetail/Microchip-Technology/PIC647?qs=bP%2FQ3uAU8wuSB7VeNzxz6g%3D%3D

 

[Zoli]

...
The edge length of the power transistor is 2,7 mm. The structures of the base and emitter are not unusual. However, it is noticeable that the transistor structures do not extend to the lower edge. There is obviously a second, smaller transistor in the lower area, which is inactive here. Apparently, the design offers the possibility of building a Darlington transistor. If a differently structured metal layer is applied, the small transistor can be used as a driver for the large transistor.
...

Most likely the Darlington is used to build the NPN version.

[Noopy]

...
The edge length of the power transistor is 2,7 mm. The structures of the base and emitter are not unusual. However, it is noticeable that the transistor structures do not extend to the lower edge. There is obviously a second, smaller transistor in the lower area, which is inactive here. Apparently, the design offers the possibility of building a Darlington transistor. If a differently structured metal layer is applied, the small transistor can be used as a driver for the large transistor.
...

Most likely the Darlington is used to build the NPN version.

That would be a logical explanation. Nevertheless it´s surprising that they wasted so much inactives silicon area. Well this part is so expensive the unused silicon probably wasn´t a problem. 

[dzseki]

Then what is the advantage of having everythimng in one package for 600+€, what anyone could build for 20€?

[Noopy]

You have to go back in time:
- switching regulator were quite new and special for most people
- it wasn´t as easy as today to buy a whole bunch of perfect switching power transistors and diodes
- everything was bulky so one part with "high integration" was a huge benefit regarding EMC
- everything (regarding semiconductors) was more expensive back in these days
- today you are paying even more because these products are out of time and (normal) production

[iMo]

It so costly because it is hand made. It is like when a watchmaker has to assemble a mechanical watch.
A latest AMD or Intel CPU may cost you the same, but that is because there is no human hand touching it during its production..

[Noopy]

There was a time when hands were relatively cheap and every transistor was made by hand. 

[Noopy]

The UMT1007 is a fast switching power transistor built by Unitrode. The datasheet specifies a maximum collector voltage of 400V. The collector current may be 5A continuously and 10A for a short time. At a current of 3A, the switch-on and switch-off time is specified as 0,4µs. At a package temperature of 25°C, 100W can be dissipated.






There is a large heatspreader in the package. Two metal strips contact the die. These were pushed onto the pins.






The edge length of the die is 3,7 mm. A kind of protective lacquer somewhat disturbs the view of the structures. As it is usual with power transistors, the emitter surface (inside) is divided into strips between which the base surface is contacted. This ensures that the base resistance in the active area is as low as possible.

The two large darker areas, one on the emitter surface and one on the base surface, are interesting. In another component, the optical appearance would suggest that the metal layer contacts the silicon surface in this area. However, this is certainly not the case here, as the metal layer must distribute the current over the entire surface and therefore must be in contact with the silicon everywhere. Another explanation would be a recess in a passivation layer. However, this also seems illogical here, as this area would then be more limited around the contacts. The most likely explanation is that the metal layer has somehow been reinforced in these areas. As the highest current densities occur there, a slightly lower-resistance metal layer would certainly be helpful.




The die was etched down a little in the edge area. This results in the familiar MESA structure. The etched trench ensures that the outer edges of the collector-base junction are exposed to as few impurities as possible. This reduces the leakage current and increases the dielectric strength accordingly.






The surface of the silicon appears irregular. However, this may also be due to the protective lacquer. The edges of the metallization are very unclean compared to more modern transistors.












The base-emitter junction only breaks down at -26V. This indicates relatively low doping. In the pictures above, the current increases as follows: 20mA, 50mA, 100mA, 200mA, 500mA. At low currents, the distribution of the breakdown areas is very irregular, which indicates a certain inhomogeneity in the structures.


https://www.richis-lab.de/BipolarA63.htm

 

[3roomlab]

by chance i saw a japanese youtube (denki otaku) channel talking about some aliex electronics purchases
https://www.youtube.com/@DENKI-OTAKU/videos
and there he bought some full wafers
i didnt think they could be exported ?

the link on his channel is broken but i found the shops (plural)
they are labelled as birthday gifts 

https://www.aliexpress.com/item/1005006508593662.html?



decorative tiles for the next man cave ?

[Noopy]

Nice!
If i had more time, i would buy a few... 

[Noopy]

The BDX66 is a PNP Darlington transistor with an integrated free-wheeling diode. The model documented here was produced by Motorola in 1988. Four assortments offer different voltage strengths. Without an index, the BDX66 blocks up to 60V. This is followed by the indices A, B and C, where the blocking voltage increases by 20V in each case. The bin C blocks up to 120V. The collector current may be 16A continuously and 20A briefly. At a package temperature of 25°C, up to 150W of thermal power can be dissipated.






The package contains a very large die on a large heatspreader.




The edge length of the die is 5,1 mm.




This is the familiar structure in which the driver transistor is located in a corner. There, the contact of the emitter area surrounds the contact of the base area so that the emitter current can be conducted directly to the base of the power transistor.




A square contour can be seen under the left bondpad, which most likely represents the diode connected in parallel to the transistor.




A pronounced MESA trench isolates the two transistors from each other. Closer inspection reveals that the trench contains a step. It appears that the MESA structure was etched in two steps.




A saw edge is visible on the side. According to the shape, the wafer was cut a little and then the transistor edges were broken.


https://www.richis-lab.de/BipolarA64.htm