Author Topic: EEVBlog #426 - HP 3457A Multimeter Teardown  (Read 31653 times)

0 Members and 2 Guests are viewing this topic.

Offline CarlG

  • Regular Contributor
  • *
  • Posts: 155
  • Country: se
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #25 on: February 20, 2013, 07:18:36 pm »
Great video as usual!

And what a coincidence! Would you guys believe that my company were about to throw away not one, but TWO 3457A (together with a bunch of other meters, counters, supplies, etc) some month before Christmas?! The production unit had determined to clean the shelfes from not-in-use-anymore...Luckily the larm went off, and a squad from the electronics department hurried down the stairs to the electronics waste area and "cleaned the table" so to speak. I got my hands on one of the 3457A's, together with a pair of Princeton Applied Reasearch 186A lock-in amplifiers, an hp3438A 3 1/2 digit bench DMM along with some more stuff.

This 3457A has Sno 3114A13573, probably manufactured spring-summer '92 (some date codes 9213). The "analog main board" is still basically the same, but there are some minor changes.

The opto coupler pairs are replaced with some hp device marked 901 366 (could not find this type designator), DIP-8 part. (No, it's not HCPL-901, that's a 16pin part.)

Oddly, the broken wire at the cut in the board by the optocouplers is still there, with the 1k resistor in place. Don't get the meaning if that? Anybody knows why the 1k + broken wire is good for?

The ROM sticker says Rev 2, but the part number's the same. The backup battery is Panasonic Lithium cell measuring 3.38V, probably replaced at least once. Otherwise it mostly looks the same as Dave's.
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #26 on: February 20, 2013, 07:23:27 pm »
I've got two questions regarding the PCB design in this unit, hopefully someone with more knowledge than I have can answer these:

1) It seems a lot of traces vary in width all the time. I suppose the smallest width is calculated as minimum for the current that could flow through. But why vary in width so much?

2) I've seen this in other (very recent) precision designs as well: instead of 'solid'copper pours as "ground"-planes they use a mesh. Here it's placed between quite wide traces. What is the reason to use a mesh structure instead of a solid reference plane? Does is have something to do with signal integrity, or the physics of current flow in a reference plane maybe?
(Maybe this should be asked in a separate topic, but for now it's directly related to the HP 3457 teardown)

<speculation mode on> this may have to do with thermal balancing more than anything else. you want to keep the heat local and not spread it around too much
<speculation mode off>

Now, as for lattice planes on INNER layers : this has to do with pcb fabrication technology. When they stack the layers the isolator adheres better to itself than it adheres to copper. so if you have large slabs of copper there is a risk of popcorning or full delamination when you send the board through the wave soldering flow ( we're talking thru-hole here ). Especially in areas where heat is transferred vertically like on a pad with a component pin in it... it creates extra stress. you can alleviate that with wagon-wheels ( thermal reliefs ) but it helps having additional 'anchor points'. so you simply make the plane a lattice. all this stuff is low frequent enough that current flows everywhere.

You can indeed use this mesh to 'force' the current path. electrons take the path of least resistance. they're not going to take a detour if thereis a lower impedance path , this low impedant path can longer than the mesh. the mesh is always going to be 'worse' than a solid slab'
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #27 on: February 20, 2013, 07:29:18 pm »
The ROM sticker says Rev 2, but the part number's the same. The backup battery is Panasonic Lithium cell measuring 3.38V, probably replaced at least once. Otherwise it mostly looks the same as Dave's.
there are 2 revisions of that motherboard. the difference is around the GPIB area. if you got the intel 8291 and another big 40 pin ( i think it's an AMI ) that is revision 1 ( like dave's )  they reshuffled also the position of the cpu and rams.

Now that i think of it , it may be the other way around. they originally used 4 bit wide rams or smaller rams and needed two sided by side and then upgraded to the 6264/62256 style  rams. the same happened in the 8116 and 8112 generators as wel as es the 3488 switchboxes.
at a certain point the rams became obsolete and they had to do a respin.  That AMI chip was a HP design and is a simple gpib bus driver.

That harris chip at the front in dave's machine got replaced as well i believe when they respun the board. It is some weird uart with an additional timer and i/o.

the service manual lists the board changes.

-edit-

page 8-9/8-10 in http://cp.literature.agilent.com/litweb/pdf/03457-90011.pdf
page 8-66 in in http://cp.literature.agilent.com/litweb/pdf/03457-90012sg.pdf

dave's got the newer rev.
« Last Edit: February 20, 2013, 07:49:00 pm by free_electron »
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline CarlG

  • Regular Contributor
  • *
  • Posts: 155
  • Country: se
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #28 on: February 20, 2013, 08:19:03 pm »
free_electron, thanks a lot for all the time you spend on helping me and others, and explaining so many details about what's going on inside :) Really appreciated!
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #29 on: February 20, 2013, 08:25:26 pm »
+2  :-+

Offline Christe4nM

  • Supporter
  • ****
  • Posts: 252
  • Country: nl
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #30 on: February 20, 2013, 08:59:56 pm »
free_electron, thanks a lot for all the time you spend on helping me and others, and explaining so many details about what's going on inside :) Really appreciated!

+another 10 or so :-+
 

Offline krish2487

  • Frequent Contributor
  • **
  • Posts: 500
  • Country: dk
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #31 on: February 20, 2013, 11:19:14 pm »
another +10 from me too....
 :-+

Thank you for the extremely detailed explanation free_electron..
If god made us in his image,
and we are this stupid
then....
 

Offline Fox

  • Regular Contributor
  • *
  • Posts: 54
  • Country: de
  • Electronics Engineer
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #32 on: February 21, 2013, 09:10:21 pm »
The ripped up trace beneath the optocouplers is acting as a spark gap, see service manual Figure 8-5.
A closed Switch should have zero Ohms or less!
 

Offline ModernRonin

  • Contributor
  • Posts: 44
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #33 on: February 23, 2013, 10:54:12 pm »
Anyone who's interested in still more explanation might want to have a look at this:

http://en.wikipedia.org/wiki/Integrating_ADC

Is it just me, or could one build a crude integrating ADC with a 555 timer?
 

Offline chickenHeadKnob

  • Super Contributor
  • ***
  • Posts: 1054
  • Country: ca
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #34 on: February 23, 2013, 11:12:07 pm »

Is it just me, or could one build a crude integrating ADC with a 555 timer?

I suppose so but I avoid 555 at all possible opportunity ;), I just don't like them. It is an old imbedded trick to make a single slope ADC using at timer compare/capture and comparator with external RC if you are stuck with a real low end micro that doesn't have an onboard ADC.
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #35 on: February 24, 2013, 01:09:39 am »
ik now people who have breadboarded one using a precision comparator and a good opamp + a cpld. they got better than 51/2 digit on the first try...
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline andyturk

  • Frequent Contributor
  • **
  • Posts: 895
  • Country: us
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #36 on: February 24, 2013, 06:21:50 am »
So if you're going to build one of these, one of the first things is to have your reference voltage available in both polarities. How do you take something like an LM399 or a LTZ1000 (which provide positive references) and flip 'em around for a negative reference?
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #37 on: February 24, 2013, 06:33:17 am »
Opamp with good DC stability and running off +/15V rails, and give it a gain of -1.
 

Offline andyturk

  • Frequent Contributor
  • **
  • Posts: 895
  • Country: us
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #38 on: February 24, 2013, 07:30:22 am »
Opamp with good DC stability and running off +/15V rails, and give it a gain of -1.
That's got to be quite an opamp in those 6.5+ digit meters.
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #39 on: February 24, 2013, 04:01:58 pm »
Not necessary. Remeber. You are not after absolute precision. Its the long term drift that matters.

Take service manual of the meter. It'll probably be a precision opamp like an ad706 or even an lf356.
You amplify the zener voltage with a specific factor, then send it through an inverting amp in unity gain. That is all you need.
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #40 on: February 24, 2013, 05:42:31 pm »
Thank freee- for a concise description of HP's multislope II algorithm.  Note, it was first used in the 3456a.  The oldest in the series, the 3455a, was built before the invention of multislope, is about 5x less accurate in DC over 1 year. 

In an old post, rfloop summarizes an issue I also share, if your primary need is the voltmeter section:

https://www.eevblog.com/forum/chat/bench-multimeters/msg11322/?PHPSESSID=c3622f6e6975828583fa6c4d6be72c4d#msg11322


 


Best Wishes,

 Saturation
 

Offline Christe4nM

  • Supporter
  • ****
  • Posts: 252
  • Country: nl
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #41 on: February 24, 2013, 06:00:22 pm »
Some German students had a go at it. See schematic for their board here:
http://www.emsp.tu-berlin.de/menue/studium_und_lehre/mixed-signal-baugruppen/digitalvoltmeter/
 

Offline Amarbir[Lynx-India]

  • Frequent Contributor
  • **
  • Posts: 431
  • Country: in
  • Indian Dealer
    • Lynx-India - Visit Us For Not So Boring Electronic Instruments
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #42 on: March 03, 2013, 05:14:52 pm »
Dave ,
     This was a kick Ass brilliant tutorial which was elevated to the moon with free-electrons inputs and many others  .Its so hard to find this kinda stuff in india  .I see a lot of mordern era multimeters in the range of USD 500 .How do they stock up .Specially GWinstek  ? . 
Regards

Amarbir Singh Dhillon [ Lynx-India ] , Chandigarh [ India ] - > www.lynxdealerstore.com , www.lynx-india.com
Indian Distributor For  [ Autoelectric , Sofitech , IDEOfy ,Peak Electronic Design [UK ] , Anatek And Creatronica ]
My Electronics Blog - > www.lynxchandigarh.com
 

Online bingo600

  • Super Contributor
  • ***
  • Posts: 1976
  • Country: dk
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #43 on: April 01, 2013, 11:47:53 pm »
I just bought a 3457A on the 'bay with a 44491A Relay Module in it.

It's a 87' , and i was wonderingif anyone can point me to a battery replacement that will work & fit 100%.

Could it be the same model as the one i just replaced in my 3478 ?

/Bingo
 

Online bingo600

  • Super Contributor
  • ***
  • Posts: 1976
  • Country: dk
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #44 on: May 13, 2013, 04:57:47 pm »
Nice walkthrough of the HP3457A , from HPJ-86-2.

Had to chop it up (rar) due to the 2MB limit here.
Had to add a "fake" .zip extension , as rar isn't allowed here

One strange thing though ... They talk about a 8051 MCU , but i'm quite sure there was a Motorola in my unit.
But i think there was something about a new CPU board , or maybe there is a separate 8051 on the analog board ....


/Bingo
« Last Edit: May 13, 2013, 05:40:46 pm by bingo600 »
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8515
  • Country: us
    • SiliconValleyGarage
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #45 on: May 13, 2013, 08:03:19 pm »
motorola on the earth referenced section.
8051 in the meter section. ( there's actually 2 cpus in the meter section. 8051 does the comms) there is a nec8 bitter that handles the multislope control.
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #46 on: August 12, 2013, 11:57:53 am »
Does anybody know the denotation of the opto-coupler? I have found similar products such as Omron EE-SJ3 series or CNY36, but they are just similar, not equal.
« Last Edit: August 13, 2013, 06:54:17 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #47 on: August 13, 2013, 06:55:04 pm »
Noone out there who can give a statement to that question?
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline chickenHeadKnob

  • Super Contributor
  • ***
  • Posts: 1054
  • Country: ca
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #48 on: August 13, 2013, 09:32:33 pm »
Does anybody know the denotation of the opto-coupler? I have found similar products such as Omron EE-SJ3 series or CNY36, but they are just similar, not equal.

Hello:

Looking at the HP service manual gives this line in the parts list:

schematic denotation# A2u509-A2U510
HP# 1990-1075 <- this is mostly useless unless you have a master cross reference list
Manufacturer: OPTEK KT3172 IC-LED opto-isolator gate IF=20MA-MAX

seems strange HP did not use in house made opto as this application does not appear to be particularly special or demanding.

Hope it helps.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: EEVBlog #426 - HP 3457A Multimeter Teardown
« Reply #49 on: August 13, 2013, 09:47:44 pm »
Quote
Manufacturer: OPTEK KT3172 IC-LED opto-isolator gate IF=20MA-MAX

Seems to be unobtanium. Is there a recommendable replacement available? Or at least a datasheet of it to find some?
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf