Difference between LM and LMT op-amps and comparators

[japlytic]

A few days ago, I encountered an LMT339N comparator.
Further searching revealed the LMT versions of the following LM devices:
LM2901
LM2902
LM2903
LM2904
LM324
LM339
LM358
LM393

A replacement list on https://cdn.badcaps-static.com/pdfs/...0e85376559.pdf shows that the LMT devices can be replaced by LM devices.
But what is the difference between LM and LMT op-amps and comparators?

[PeteS in CA]

Re: Difference between LM and LMT op-amps and comparators

Might be a character the vendor added ala Samsung's KA339. A bit of trivia ... National Semi's prefix "LM" stands for "Linear Monolithic". They also have/had "LF", "Linear FET" and "LH", "Linear Hybrid".

[starfury1]

Re: Difference between LM and LMT op-amps and comparators

well I did some googling and could find no reference to what it might mean.
I checked out national but no mention so maybe its some thing specific to motorola's
codes

here the national device markings pdf



but in (pg3 ) table 7 no LMT to be found.

So whatever it does relate to it seems its obsolete from your doc and no one seems to have any info out there on it that I could find.
(except as a part number)

still it would be nice to know

[PeteS in CA]

Re: Difference between LM and LMT op-amps and comparators

The end-of-life document you linked makes two levels of replacement recommendations - direct equivalents and similar parts. The LMT-LM recommendation seems to be direct equivalents.

[Andybr]

Bit late to this but was wondering this myself so given nobody found the answer decided to look for it myself.

Short Answer: "T" is for Toulouse!

Caveat: This isn't confirmed by the higher powers but its just deduced by the following.

Long answer:

In 1993 the Motorola LMT series appeared out of thin air. There was no fanfare, no publicity, but the series appeared on some distributors parts lists and a few datasheets appeared which later on appeared on some websites though strangely not on Motorola's website (as far as I can ascertain on there archived website) and not in any of Motorola's databooks.

Here's the datasheet links for the LMT324 and the LMT358

https://datasheet.octopart.com/LMT32...t-27907009.pdf (dated 1993)
https://datasheet.octopart.com/LMT35...-180795246.pdf (undated except for 1997 fax date)

Note Motorola didn't have a website till 1997. Here's the link to the archived website.

https://web.archive.org/web/19970316...//www.mot.com/

The original website linked to the following page for analog ics.

https://web.archive.org/web/19970328...128_index.html

DL128 in the web page is a reference to the Motorola's analog/linear interface ICs databooks. The links for the databook 1990 and 1993 versions are below.

https://media.searchelec.com/specshe...k1990_text.pdf Rev3 1990
http://www.bitsavers.org/components/..._ICs_Vol_1.pdf Rev 4 1993

Whilst the databooks contain both the LM324 and LM358 datasheets the LMT series parts aren't listed. However there are differences I can see for the 1990 and 1993 databooks for these devices as follows:

1) Package types available changed.

2) Addition of a Vcc 5.0V case for the parameters for the "Output Voltage High Limit" which is an additional added constraint so it isn't a change in the device.

3) Addition of this line on the first page.

"ESD Clamps on the Inputs Increase Ruggedness without Affecting Device Operation".

Item 3 requires a change to the die.

There could of been a "Process Change Notification (PCN)" for this but the web archives don't extend to before 1996.

Prior to 1993 the LM358 and LM324 dies were likely produced at the following two Motorola wafer fabrication (fab) plants. Archives for these indicate they were the only two expressively doing op-amps and whilst other Motorola fabs had the capability they were oriented to logic or power devices.

Phoenix, Arizona, USA.
Bipolar 1 (BP1). Products: Telecom, Op amp, Automative.
https://archive.computerhistory.org/...-05-01-acc.pdf

Toulouse, France.
Bipolar 4 (BP4). Products: Telecom, op amp, reg, auto
https://archive.computerhistory.org/...-05-01-acc.pdf

In 1993 Motorola began using a wafer fabrication plant in Roznov, Czech Republic which was a joint venture with Tesla Sezam (a Czech semiconductor company).

"1993 MOTOROLA Ordered first Chips of Bipolar Analog IC's"
https://indico.cern.ch/event/67311/c...Freundlich.pdf (translated)

"In the end, the company Motorola turned out as the only perspective partner. Through its enterprise in the French Toulouse, this company provided TESLA with its experts and consultants. Afterwards, there were produced qualification sets of dice that were packaged in the Motorola plants and submitted to the careful reliability tests. Those tests were completed without any rejection. In 1993 Motorola decided to close a contract for a supply of the semiconductor components in the agreed amounts. There were bipolar transistors on the 4" Si wafers. The first types produced were voltage controllers of the series MC 78xx, which were the three-terminal integrated circuits. From the beginning of this cooperation, this assortment was being gradually added with the other types of components, not only by the voltage controllers but also by the operating amplifiers and by the customer integrating circuits."

https://web.archive.org/web/20010306...ry_en.html#h01 2001

In 1999 products were given as "Standard analog products: operational amplifiers, regulators"

https://www.sec.gov/Archives/edgar/data/1097864/000091205700000782/0000912057-00-000782.txt 1999

"In the TESLA SEZAM Wafer Fab Line, various types of integrated circuits on the Si wafers are being produced. These are divided into the several application families. Nowadays, it is possible to define those groups as follows:

Control circuits for the switching power supplies
+/- Voltage stabilizers
Adjustable voltage controllers
Simple and multiple operating amplifiers
Motor control circuits of motors"

https://web.archive.org/web/20010306...cz/cipy_en.html 2001

From the above it seems it is a possibility that Motorola started producing LM324 and LM358 op amps from 1993 in Roznov, Czech Republic.

Looking at the LMT324 and the LMT358 datasheets they don't include ESD protection unlike the LM324 and the LM358 in the Motorola 1993 databook. However otherwise they conform to the same parameters. In fact they seem to be the same as the LM324 and the LM358 in the Motorola 1990 databook.

One thing of note is that the Motorola databooks include a LM324A device for which there is no LMT equivalent. However there is a new device in the LMT324 datasheet which is LMT324B. Like the LM324A device the only difference is a reduced maximum input voltage offset however this is only 4.5mV at 25C for the LMT324B whereas it's 3.0mV at 25C for the LM324A. Why is there a difference? Whether the LMT324B is a different die or a standard LMT324 device tested to be within LMT324B limits I don't know.

So given the timeline that the LMT series without ESD protection was introduced at the same time as ESD protection was added to the LM324/LM358 devices I'm betting that the LMT devices are simply the old LM devices without ESD. LMT324B remains a mystery though.

If the LMT series is really the old LM devices without ESD why was this line kept going? I can think of three possible reasons as follows:

1) The fabs constructing the LMT devices weren't capable of making the new ESD protected LM devices.

2) There was enough of a difference between the new ESD devices and the old LM devices that some customers didn't want the new ESD devices

3) The new ESD devices cost more than the old LM devices and some customers didn't want to pay the extra.

Given the Czech Tesla fab came into use at the same time as the LMT devices were introduced the "T" in LMT could possibly either be "T" for "Toulouse" or "T" for "Tesla". The following paragraphs look into this a bit further.

On 08-Feb-1999 Motorola issued "Product Discontinuance" PCN4323 which discontinued the LMT series as of 08-Aug-2000.

https://www.onsemi.com/pub/docs/pcn/4323.pdf

This was followed up by "Product Bulletin" PCN10062 issued by OnSemi (which took over part of Motorola in July 1999) on 03-Nov-1999 which immediatedly discontinued the LMT series (and the related MCT and SCT series). This PCN noted that the reason for this immediate discontinuance was that that they were "unable to secure the foundry die used in their construction".

https://www.onsemi.com/pub/docs/pcn/10062.pdf

Of importance here is when OnSemi took over part of Motorola they didn't get all the op amp wafer foundries included. In fact out of all the three op amp foundries (BP1@Phoenix, BP4@Toulouse, TESLA@Roznov) only the TESLA@Roznov foundry was included. Any dies still sourced from the other foundries were subject to negotiation.

Since the wafer foundry TESLA@Roznov was included in OnSemi the PCN above means that this foundry couldn't be the source of the LMT devices.

On 27-Sep-99 OnSemi issued a "Product and Process Change Notification" PCN4813 titled "Transfer from BMC/BP4 to TESLA of Analog Devices". This was followed by another "Product/Process Change Notification" PCN10041 titled "Transfer from BMC/BP4 to TESLA of Analog EPI82, 85, 92 Devices".

https://www.onsemi.com/pub/docs/pcn/4813.pdf
https://www.onsemi.com/pub/docs/pcn/10041.pdf

The "BMC" foundry reference name here is the collective name for the bipolar foundries at Phoenix which included BP1. BP4 as mentioned before was the Toulouse foundry. This confirms that OnSemi was moving op amp production from the foundries it didn't own (BP1/BMC@Phoenix, BP4@Toulouse) to the foundry it did own (TESLA@Roznov).

Notably in the list of devices given LM324 and LM358 weren't mentioned. This would seem to confirmed that production of LM324 and LM358 had already moved to TESLA. Indeed the following line in the PCN seem to confirm this.

"The Tesla site has been qualified for standard analog products since 1993 and is currently running high volume analog devices."

As the LM324 and LM358 are popular devices they would fit the criteria for "high volume" devices.

On 08-Mar-2000 an "Update Notification" PCN10147 was issued title "Update Notification for PCN's 4813 and 10041". This PCN effectively states they were able to delay the transfer of BMC (Phoenix) devices to TESLA but not for the BP4 (Toulouse) devices.

https://www.onsemi.com/pub/docs/pcn/10147.pdf

Given the above it's fair to say that the "T" in the LMT series stands for "Toulouse".

But like always I could be wrong. Either somebody who knows for sure needs to confirm or I can simply compare the dies (which I may do).

Cheers
Andy B

[Andybr]

I managed to find a die shot of LMT324 which is the quad version of the LMT358 (which is a dual op amp).

The link to the die shot is below. On the linked page click on the die image for a higher resolution image.

https://www.siliconpr0n.org/archive/...ola:lmt324dopt

This is on the siliconpr0n site which is a great site for IC die shots and has a number of contributors including Antoine Bercovici who took this shot.

Antoine also has an X page with more die shots.

https://x.com/siliconinsid

On perusing this die shot I noticed the marking "75902" on the righthand side of the die. There's also a "1" on the bottom which confirms to me that the "75902" marking is the right way up.



Normally on dies you get the IC device code plus occasionally the revision id directly after. Sometimes they just give the IC device number. Other markings can include the manufacturer, the mask code, copyright symbols, and pad alignment marks ("1" in this case).

Antonie provides some Motorola and OnSemi LM324 shots as follows.

https://www.siliconpr0n.org/archive/...ola:lm324adopt Marking "LM124N".

https://siliconpr0n.org/archive/doku...onal:lm324nopt Marking "LM124".

Note LM124 is the base code for the die used in LM124, LM224, and LM324 parts. "N" in the first link is the revision.

Obviously "75902" is not "LM124". Neither is it LM75902 or MC75902 (MC is a common Motorola prefix).

On a bit of searching I found two parts that match: Toshiba TA75902 and Korea Electronics Co (KEC) KIA75902.

Korea Electronics Co was originally a subsidiary of Toshiba so this is the same part.

The link to the Toshiba datasheet is below.

https://www.alldatasheet.com/datashe.../TA75902P.html

On comparing this datasheet I found that the simplified schematic basically matches the Motorola LM324 schematic and the parameters seem to be the same except the temperature range matches the automotive version of the LM324 which is the LM2902.

Furthermore National Semiconductor in their "1988 National Linear Databook: Volume 1" linked below includes a cross reference from the TA25902 to the LM2902.

https://archive.org/details/bitsaver.../2up?q=TA75902

So it seems that the Motorola LMT324 in the die shot and the Toshiba TA25902 IC are the same IC.

While some may be inclined to think that the LMT324 shown in the die shot is a counterfeit, Toshiba and Motorola (and OnSemi) have a long collaboration including co-owning semiconductor fabs in the past (and also presently).

At the time this IC was produced Toshiba and Motorola didn't co-own a fab but were collaborating on other things.

My inclination is that Motorola decided in 1993 to upgrade the LM324 to the ESR version and for customers wanting a non ESR version they used the Toshiba TA25902 marked as LMT324.

So it seems "T" based on this case and probably the other ICs (not checked) is "T" for "Toshiba".

Cheers
Andy B