Teardown : Agilent 4142B semiconductor DC test mainframe/SMU

[TiN]

I have something special today to show you, dear test gear nuts



Let's see your scherlock skills.
Guess the equipment model by little piece on the photo? There is enough information around to make exact guess

EDIT:

It's 4142B with some modules I got for 200$+230$ shipping.



This beast is 29kg, full 4U size, so not easy to haul it .

Peek inside it's rather special power supply:



GNDU SMU (1.6A sink, 0V) + ADC



Isn't this BB DAC gorgeous?



SMU module 41421B:



Some high-impedance teflon action:



Everything seem to work just fine.

Added some wires to check main 10V reference:



Measured just fine:

[Earendil]

HP 4142B?

[Smith]

Some HP SMU? At least is says SMU, and the silk seems to be HP

[ywara]

41421B 100V/100mA module for the 4142B. Looks to have fairly impressive current resolution.

[TiN]

Bingo!

I took lots of photos of all modules and supply boards, and posted them in article on my site

Enjoy. Even though unit was made in 1992/2000 (different boards different date codes), it does not have a single spec of dust. My best guess - it was in clean room of semiconductor facility, hence pristine condition. Not a smudge of dust even on fan blades, and trust me, those 120x38mm 24V fans are no joke.

[nctnico]

Neat! One of the hidden gems you can stumble upon and if you can take the noise from such a machine then why not?
Quite some electronics porn inside as well.

[Vgkid]

Nice unit, what is the internal vref?

[TiN]

REF102CM, as I shown in article.

I received broken 41420A SMU module off ebay, this one definately need some repairs

python ./cvr.py
HEWLETT PACKARD,4142B,0,4.30
0,0;HP41421B,3;0,0;HP41424A,1;0,0;HP41424A,1;0,0;HP41420A,3
8423,0,0,0

According to HP-IB manual reference error 423 (8 is just meaning it's slot 8 module) is :

SMU is not functioning properly. SMU failed I Monitor Amp. Offset Error Test of Self-test.

What ever that means..

[TiN]

Here's 41420A photos for those who care:



DAC is same as in 41421B's.



More similar hybrids:



And rest of 'em:





Back of the output board still have flux. Not pretty, but it's not the reason why module failed..



Triax outputs:





Moar green transistors from NEC Japan :

[free_electron]

that's a yokogawa-hp machine.

[TiN]

Here's front panel looks after some H2O2 procedures:



Having so tight plastic wrap is actually not such a good idea, as that made barely visible areas of different tint due uneven UV exposure. Works for first try, but lesson for next time.



Overall happy with the process. Bodged quick guide article about it. Keithley panel is out now this time..

[nidlaX]

God, why have I been running around salon / beauty supply shops when I could just buy that stuff from Amazon.

[lukier]

You don't add Tetraacetylethylenediamine (TAED) as an activator? This I believe is mentioned in the original Retr0bright receipe. It doesn't make a difference?

[TiN]

No, i did not add anything.

[TiN]

It's been awhile since this thread was created. I got just "few" additions today.

* Bad 41420A got long ago, fails self-diagnostics. Error code is 423, according to service "I monitor opamp fail" on output board A2.

How to fix? Buy more donors! 

Today arrivals:

* Two supposed to be good 41420A HPSMU (40uV-200V, 1nA-1A ranges, 20fA best resolution). Installed into box, both run w/o errors.
* Third, supposed to be bad 41420A HPSMU (40uV-200V, 1nA-1A ranges, 20fA best resolution). Installed, running w/o self-test errors. Hm?
* One more 41425A monitor (it's like a fast analog comparator between SMUs levels, handy for tracking measurements). No self-test errors. I've already had one from before .



Red zero means no errors detected.

There are few more modules for cheap on ebay, so SMU-nuts, beware of EEVBlog/xDevs effect. Bunch of hi-res photos in my old teardown article still available. Plenty of AE foil resistors and japanese green FETs insides, with fancy shielded reed relays.

And finally with Keithley 6485 I can actually hook it all up and measure some low currents. Wonder who is more noisy, Keithley picoammeter (lowest range 1nA) or JHP SMU?

[plesa]

And finally with Keithley 6485 I can actually hook it all up and measure some low currents. Wonder who is more noisy, Keithley picoammeter (lowest range 1nA) or JHP SMU?

I do not know HP SMU, my guess is - Keithley 6485 is more "noisier" meter

[TiN]

As SMU-nuts may learned already, triaxial cable set for SMUs can often cost as much as SMU itself.
I received today a jackpot, perfect match for my 4142B:




This cable designed to connect 4142B with four SMU modules and one VM/VS module to the HP 4085B semiconductor analyzer switch matrix.

[HighVoltage]

As SMU-nuts may learned already, triaxial cable set for SMUs can often cost as much as SMU itself.
I received today a jackpot, perfect match for my 4142B:

Very interesting!
I bought the very same set a few month ago and have run out of Triax cables / connectors again !

[Berni]

As SMU-nuts may learned already, triaxial cable set for SMUs can often cost as much as SMU itself.
I received today a jackpot, perfect match for my 4142B:

This cable designed to connect 4142B with four SMU modules and one VM/VS module to the HP 4085B semiconductor analyzer switch matrix.

Dang! That is a score indeed. I was forced to make my own and it still cost a fortune!

[plesa]

As SMU-nuts may learned already, triaxial cable set for SMUs can often cost as much as SMU itself.
I received today a jackpot, perfect match for my 4142B:

It looks like you are planning prober for Christmas