Author Topic: LM723 die pictures  (Read 8657 times)

0 Members and 1 Guest are viewing this topic.

Online magic

  • Super Contributor
  • ***
  • Posts: 3332
  • Country: pl
Re: LM723 die pictures
« Reply #75 on: March 04, 2020, 04:33:35 pm »
I didn't review every nook and cranny but topology seems equivalent to all the others.

Notably, Q11 is powered from Q4 and O9 from Q5, which means that Fairchild lied on their schematic.

Either that or we need an older Fairlchild die. But no, they lied. It totally makes sense to wire things that way. Q9 and Q11 provide bias current for Q4 and Q5 while Q4 and Q5 improve PSRR for Q9 and Q11. Win win.
 
The following users thanked this post: BravoV, Noopy

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #76 on: March 04, 2020, 06:34:01 pm »
Thanks and credit should goes to you, Noopy, I really enjoy it.  :clap:  :-+

It was me a pleasure!  :)


As I'm noob at interpreting from the die shot, so how this Fairchild UA723 differs from others ?

magic is more familiar with the LM723. He can do better analysis.  :-+

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #77 on: March 09, 2020, 05:31:52 pm »
News! Surprising news!  ;D

BravoV has also sent me a National Semiconductor LM723CH.

With a "C" we would expect a simple die like in the LM723CN or in the LM723CJ:
https://www.richis-lab.de/LM723_05.htm
https://www.richis-lab.de/LM723_01.htm

Wrong! In the LM723CH there is the same die as in the LM723J:



Sorry, the picture is a bit blurred. Didn´t get it better…

Whole story here: https://www.richis-lab.de/LM723_07.htm

Perhaps the die was sorted out because of bad characteristics but it was ok for a C-variant?
Perhaps they had a lot of the complexer dies and it was cheaper to put them also in the C-package?


Overview for those reading only the last post of this thread: https://www.richis-lab.de/LM723.htm
 
The following users thanked this post: BravoV, Wolfgang

Online magic

  • Super Contributor
  • ***
  • Posts: 3332
  • Country: pl
Re: LM723 die pictures
« Reply #78 on: March 09, 2020, 05:43:57 pm »
Maybe they just switched to this new design at some point.

TI still makes National's LM723 so one could see what current production chips look like, but they are only available in metal cans and at stupidly high price. Milking the last remaining aerospace customers or whatever they do.
 

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #79 on: March 09, 2020, 05:49:19 pm »
Maybe they just switched to this new design at some point.

TI still makes National's LM723 so one could see what current production chips look like, but they are only available in metal cans and at stupidly high price. Milking the last remaining aerospace customers or whatever they do.

The LM723CH has a date code 8848.
Possible...
Perhaps I find a newer one to compare…  :)

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 1896
  • Country: mx
Re: LM723 die pictures
« Reply #80 on: March 09, 2020, 05:52:22 pm »
 >:D
Indeed! US$11.10 for the cheapest variant, and US$24.70 for the full rated part!

Besides the military, there must be a few audiophiles which also crave this part.
 

Offline imo

  • Super Contributor
  • ***
  • Posts: 2773
  • Country: li
Re: LM723 die pictures
« Reply #81 on: March 09, 2020, 06:08:48 pm »
That price is not because the 723 chip is great, but because the package (metal TO-100) is not a mainstream anymore. It could be they bond manually 1000pcs from time to time on a machine gathering dust in TI's cellar  :D
PS: I would say the above chip is the oldest design, the artwork done manually into the rubylith foil.
All the others Noopy has analyzed are newer one done with CADs, imho.
« Last Edit: March 09, 2020, 06:16:11 pm by imo »
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 15577
  • Country: za
Re: LM723 die pictures
« Reply #82 on: March 09, 2020, 06:54:46 pm »
More like they made a few hundred thousand, and ran them through the Aerospace qualification process, then took the remainder of the batch after test and placed them inside a controlled atmosphere storage facility, so that when they need a few hundred at some point, they have some already qualified, just needing another round of acceptance testing to verify them. Storage cost built into the price.
 

Offline imo

  • Super Contributor
  • ***
  • Posts: 2773
  • Country: li
Re: LM723 die pictures
« Reply #83 on: March 09, 2020, 07:08:28 pm »
..and placed them inside a controlled atmosphere storage facility,..
For example in Svalbard Global Seed Vault  :D
 

Offline TOTALCHIPS

  • Newbie
  • Posts: 3
  • Country: us
Re: LM723 die pictures
« Reply #84 on: March 16, 2020, 04:03:43 pm »
I recently found this component, the Fairchild Mil version. Just to add to your collection.
 
The following users thanked this post: Noopy

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #85 on: March 16, 2020, 04:22:57 pm »
Thanks a lot!
I will upload it soon!

Interesting, it´s not very different to the "normal" Fairchild LM723 but has some minor distinctions. Have to check that.

Thank!  :)
 
The following users thanked this post: Wolfgang

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #86 on: March 16, 2020, 09:51:05 pm »
The JM38510/102 is online: https://www.richis-lab.de/LM723_08.htm

It seems that the design is not really different to the UA723: https://www.richis-lab.de/LM723_06.htm

Only the big transistor in the right lower corner is not connected in the UA723.
But it is no additional transistor. The UA723 has a smaller transistor placed a little bit higher that does the same job. I don´t know why they changed the transistor size...  :-//

The JM38510/102 is newer but seems to have the older design. Perhaps it was sold for a longer time because it has the MIL-spec.

Online magic

  • Super Contributor
  • ***
  • Posts: 3332
  • Country: pl
Re: LM723 die pictures
« Reply #87 on: March 17, 2020, 06:36:59 am »
This "transistor" is the zener of course, where is the collector? ;)
 

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #88 on: March 17, 2020, 08:31:04 am »
This "transistor" is the zener of course, where is the collector? ;)



The collector is green.  :)
OK, it´s not used as a tansistor but basically it is a Transistor.


...I´m still not so familiar with the LM723-topology. Is this part the Vref-Zener?
« Last Edit: March 17, 2020, 08:33:29 am by Noopy »
 

Online magic

  • Super Contributor
  • ***
  • Posts: 3332
  • Country: pl
Re: LM723 die pictures
« Reply #89 on: March 17, 2020, 10:35:20 am »
Yes.  And the pad it's connected to is VREF output.
 
The following users thanked this post: Noopy

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #90 on: March 17, 2020, 10:42:05 am »
I thought so, thanks!  :-+

So they changed the size of the zener reference but kept the old one on the die... Interesting... Hm...  :wtf:

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #91 on: December 29, 2020, 04:04:44 pm »


I have a new LM723! Do you know RIZ, Radioindustrie Zagreb?  ;D They have built the IL72723.
The ceramic package is shifted a bit.  :o




Nice! Nothing special. But there is a failed bond.
You can see that the two parallel power transistors are somehow similar to the Tesla MAA723:
https://www.richis-lab.de/LM723_04.htm
Well both are soviet parts.


More pictures here: https://www.richis-lab.de/LM723_09.htm



I also have updated the Fake ST LM723CN:
https://www.richis-lab.de/LM723_00.htm
Nothing special, I just updated the text.
I´m pretty sure that´s a soviet design because of the particular transistor design.



I also have updated the Fairchild MIL-723:
https://www.richis-lab.de/LM723_08.htm




I was curious about the small difference to the "normal" Fairchild µA723 in the bottom right corner.




In the MIL-723 the Q6 and the D2 have their own collector area. In the µA723 they share the same collector area. D2 is a transistor working in breakdown as a zener reference.
Electrically that should make no difference but there could be a thermal effect. If Q6 gets hotter it conducts more current and so there is less current flowing through D2. Less current gives less voltage but with the higher temperature the zener voltage goes up. Perhaps both effects cancel themself at least partly. And of course better thermal coupling gives better compensation.  :-+


 :-/O

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1117
  • Country: us
Re: LM723 die pictures
« Reply #92 on: December 29, 2020, 04:29:47 pm »
Great work Noopy.

I was talking to a good friend of mine about the 723.  He designed TIs first knock off of the 723 back in the early 70s.  It was news to me...he had never told me that--or I forgot!

I will see him next month on his ranch...will archive this thread to an iPad so we can discuss (no internet at his ranch).

 

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #93 on: December 29, 2020, 04:40:59 pm »
Thanks!  :)

That sounds interesting!  :-+

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #94 on: January 19, 2021, 07:34:01 pm »
Sorry, I posted a UA723-update in the wrong thread...  :palm:

Here you can find news about the UA723:
https://www.eevblog.com/forum/projects/voltage-regulators-die-pictures/msg3421856/#msg3421856
https://www.richis-lab.de/LM723_06.htm




 ;D 8)
 
The following users thanked this post: Wolfgang, imo

Offline imo

  • Super Contributor
  • ***
  • Posts: 2773
  • Country: li
Re: LM723 die pictures
« Reply #95 on: January 20, 2021, 09:42:35 am »
I like your on-chip current measurement based on the emitted light intensity :)

 
The following users thanked this post: Noopy

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 13116
  • Country: us
  • DavidH
Re: LM723 die pictures
« Reply #96 on: January 22, 2021, 06:17:38 pm »
I like your on-chip current measurement based on the emitted light intensity :)

I remember seeing a table once for fusing capability for different bond wire diameters which could be used to make estimates for the minimum current which had been present to destroy a part.
 

Offline TOTALCHIPS

  • Newbie
  • Posts: 3
  • Country: us
Re: LM723 die pictures
« Reply #97 on: June 15, 2021, 03:16:00 pm »
An original LM723CN from ST Microelectronics.
 
The following users thanked this post: Noopy

Offline Noopy

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: de
    • Richis-Lab
Re: LM723 die pictures
« Reply #98 on: June 16, 2021, 03:28:21 am »
An original LM723CN from ST Microelectronics.

Interesting! I have a original ST LM723CN in the queue too. It looks a bit different. Coming soon...
Would it be ok for you if I put this picture on my website?

Online magic

  • Super Contributor
  • ***
  • Posts: 3332
  • Country: pl
Re: LM723 die pictures
« Reply #99 on: June 16, 2021, 08:37:09 am »
For old, pre-merger SGS product you could look for L123. It seems to be a 723 clone.
 
The following users thanked this post: Noopy


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf