Author Topic: Live design flow of 'FX' 10V LTZ1000A-based reference  (Read 3645 times)

0 Members and 1 Guest are viewing this topic.

Online nidlaX

  • Frequent Contributor
  • **
  • Posts: 550
  • Country: us
Live design flow of 'FX' 10V LTZ1000A-based reference
« on: November 08, 2017, 12:15:02 AM »
I've renamed this thread from "TiN livestreaming: layout of new LTZ1000A 10V reference module" per TiN's request since the layout is done and has been sent for manufacturing! :clap:

FX Live Test Data

Layout videos:

Part 1:

Part 2:

Part 3:

Assembly video:

Part 4:

Also check out his posts for different design stage PCB pictures:

First Batch
Second Batch
Third Batch
Fourth Batch (Assembled)

Final Schematic (as of 11-9-2017)
« Last Edit: November 28, 2017, 09:00:42 AM by nidlaX »
 

Online dr.diesel

  • Super Contributor
  • ***
  • Posts: 2003
  • Country: us
  • Cramming the magic smoke back in...
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #1 on: November 08, 2017, 12:21:45 AM »
He'll be working for Fluke before you know it!   ;D

Online Echo88

  • Regular Contributor
  • *
  • Posts: 227
  • Country: de
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #2 on: November 08, 2017, 12:32:47 AM »
I wanted to start designing my own LTZ1000A-board, but considering that digikey claims delay of delivery for the LTZ1000A till April next Year i still have much time for other stuff...  :--
 

Online mimmus78

  • Supporter
  • ****
  • Posts: 591
  • Country: it
    • 118Volt.it
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #3 on: November 08, 2017, 02:46:24 AM »
OK I guess actually I look like my nephew when he watch mine-craft videos on YouTube.

 :popcorn:
-> Check my Store @ Ebay
-> OR my Blog @ 118volt [IT][EN]
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #4 on: November 08, 2017, 06:05:57 AM »
Wow. It's not worthy a separate thread. Just a first try, I'm sure there are lots of issues to iron out first on this rev.

I'll try to get second part tomorrow.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 
The following users thanked this post: dr.diesel, hwj-d

Offline Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1303
  • Country: de
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #5 on: November 08, 2017, 09:29:22 AM »
Illya,
once again, you designed a really nice board!

You mentioned in the video, that you designed in a low pass between the LTZ reference and the AD OpAmp, to suppress the chopper spikes going backwards into the reference circuit.
Well, I doubt a bit that this effect is so important.
You also introduced a lot of solder junctions between the LTZ and the OpAmp, that may ruin your T.C., or you have to balance all these junctions, as it's been done in the 3458A DCI circuit, you remember?

Anyhow, in my in own Spartan single-sided design, I directly (copper-wise) connected the LTZ output to the LT1052 ChopAmp, accepting possible spike feedbacks, but avoiding thermo junctions..

We'll see in 1 year time, what will happen,... and I just started a noise / stability measurement on the 10.000V output, realized with these ordinary BMF from AE.. we'll see tomorrow.

Frank
« Last Edit: November 08, 2017, 09:35:11 AM by Dr. Frank »
 
The following users thanked this post: cellularmitosis

Offline eurofox

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: be
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #6 on: November 08, 2017, 09:43:17 AM »
I wanted to start designing my own LTZ1000A-board, but considering that digikey claims delay of delivery for the LTZ1000A till April next Year i still have much time for other stuff...  :--

https://www.benl.ebay.be/itm/LTZ1000ACH-CAN-8-Precision-Reference/322563133347?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2055119.m1438.l2649
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #7 on: November 08, 2017, 03:39:54 PM »
Illya,
once again, you designed a really nice board!

Will see on that after getting year worth of data :).

LPF between chopper and reference also attempt to REF reduce noise little bit for quieter output. Likely not very effective.
I'm going to revisit that part, by removing R20,R28 and installing same package resistors complimentary to R18,R19, to equalize number of junctions.
I also don't like the grounds, might as well redo that part.

Also I messed up with guarding, guard trace potential is 0.44V so wrapping it around reference output do not good, should be around temperature divider node.
Consequences doing layout at 4am  :-X.

Quote
We'll see in 1 year time, what will happen,... and I just started a noise / stability measurement on the 10.000V output, realized with these ordinary BMF from AE.. we'll see tomorrow.
Looking forward for your data.

I've scheduled finishing up part 2 for today, 10pm UTC+8.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline zucca

  • Supporter
  • ****
  • Posts: 992
  • Country: it
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #8 on: November 08, 2017, 09:02:20 PM »
Preorders?  ;) ::)
« Last Edit: November 09, 2017, 12:42:05 AM by zucca »
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 3514
  • Country: de
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #9 on: November 09, 2017, 03:52:36 AM »
The LTZ circuit usually does not need guard traces, as it is mainly low impedance (e.g. 1 K). It is only part of the temperature regulator that uses high value resistors, but this is with rather high signal levels an not much sensitivity. Also the 70 K the collectors are at point that is not that sensitive.
Chances are the guards do more bad than good due to thermal effects - though I don't think this is that important.

Before doing much work on the layout it is a good idea to get the schematics right before. A first idea of the layout is Ok, to get an idea of the placement of the main parts, but no final layout before the circuit is finished. AFAIR there where still some points not finished about the output stage.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #10 on: November 09, 2017, 05:59:01 AM »
Already removed unwanted guards, as even if there would be any benefit, guard source need to be low impedance, not the way it was done. That was a flop.

PDF-schematics.

I've add provisions for compensation cap around output stage opamp, from our discussion before.

Animated GIF all layers (5 sec/layer):



Changelist today:

* Replaced input filter RCR to balanced network using MELF metal film resistors.
* Routed GND_Signal return as point-point signal, no shapes.
* Rearranged placement for output chopper for more balanced routing

I'll do some final checks and finetune tomorrow, and get board for fab :). It's better to test any issues on practical setup, than spending pages on forums on theory.  :scared:

Part 2 vid:

.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 
The following users thanked this post: dr.diesel

Online dr.diesel

  • Super Contributor
  • ***
  • Posts: 2003
  • Country: us
  • Cramming the magic smoke back in...
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #11 on: November 09, 2017, 11:21:27 PM »
Preorders?  ;) ::)

Once everything is finalized, perhaps a group order?

Offline zucca

  • Supporter
  • ****
  • Posts: 992
  • Country: it
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #12 on: November 10, 2017, 12:32:35 AM »
Once everything is finalized, perhaps a group order?

Bare board will be cool but without Fluke 2464707 kind of sad business.
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #13 on: November 10, 2017, 01:21:57 AM »
Maybe somebody is in friends with VPG to get a group buy on divider? That should be easy, compared to Fluke custom network  :).
Same reason why as of now I see no point to making FX a group order device.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline zucca

  • Supporter
  • ****
  • Posts: 992
  • Country: it
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #14 on: November 10, 2017, 01:25:50 AM »
Maybe somebody is in friends with VPG to get a group buy on divider?

That would be sooo soooo cool.  +1
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #15 on: November 10, 2017, 06:54:59 AM »
PCB layout complete.

PDF-schematics.

Animated per-layer GIF PCB.



Changelist:

* Added R33 for equal amount of resistors in negative feedback.
* Added optional C48,C49 at the output force
* Redraw schematic symbol and changed trim connections on Fluke resistor network for clarity
* Added power LED D18,D30
* Added I2C U7 EEPROM, for storing module-specific DC output voltage, calibration values, tempco/drift values, configuration parameters, etc.
* Changed connector type to 20pin 0.5mm pitch XP2 for onboard temp sensor and I2C ROM. Power for sensor and ROM are 100% isolated from rest of the reference.
* Added extra C46 cap near heater NPN.

Part 3 video:

.

xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline branadic

  • Frequent Contributor
  • **
  • Posts: 732
  • Country: de
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #16 on: November 10, 2017, 07:34:35 AM »
Illya,

What do you plan for the housing?
How do you fix the pcb inside the housing?
What connectors are you planing to use?

From a temperature point of view the LTZ connections are rather asymmetrical. I also wonder about the shape of the board. Just another fancy looking LTZ board like the KX? Sorry, but I'm sceptical that this approach does any good at all. But I appreciate that the ground layer is no longer fragmented.

-branadic-
Prema 5000 | Prema 5017 SC | Keithley 181 | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, HDO6054, Keysight 53230A and other goodies at work
 
The following users thanked this post: TiN

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: TiN livestreaming: layout of new LTZ1000A 10V reference module
« Reply #17 on: November 10, 2017, 03:38:24 PM »
branadic,

It will be box very similar to this 112 x 60mm Hammond, places next to battery compartment in Fluke 792A power pack.
There is no room to make it any bigger, so it will be all very tight fit.

PCB will be suspended inside the housing, without rigid mechanical coupling, as box eventually might see transportation, however it's not quite the travel standard due to weight.

There will be no connectors at the board, other than auxiliary 20 pin for digital interface and onboard temperature monitor.
Power input and reference output will be routed out of the box with copper shielded twinax, soldered to output port J3.

External world connectors will be Low Thermal 2758 or LEMO EGG, not decided on that part yet.

Yes, just another fancy looking board with some bling added to it. Board sent to fab, will see it in few weeks. Should be enough time to get some parts. I already have LT1013's, LTZ1000ACH's and VHD 20K/7.663K dividers.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 3514
  • Country: de
Re: Live design flow of 'FX' 10V LTZ1000A-based reference
« Reply #18 on: November 10, 2017, 08:48:57 PM »
I had not look at the layout in detail before. The animated gif is kind of hard to watch if one want to look at details. For me it looks like the connections to R1 not so well made. My feeling is there is way to much trace resistance in series with the 120 Ohms. One might be able to fix this with extra wire in parallel, and / or a bridge.

A more minor point with the output amplifier is that C41 and D43 would be better towards ground instead of the the +- supply.

The filtering cap C36 might also be better differentially instead of toward ground - but this might be only a minor effect.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 2939
  • Country: tw
  • Volt-bolt
    • xDevs.com
Re: Live design flow of 'FX' 10V LTZ1000A-based reference
« Reply #19 on: November 10, 2017, 09:39:28 PM »
I'll post PDF-version later.

Connection to R1 is 40 mil wide trace, 1340 mils long from LTZ, and 1150 mils to GND star point.
xDevs.com | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Flinstone

  • Regular Contributor
  • *
  • Posts: 160
  • Country: be
Re: Live design flow of 'FX' 10V LTZ1000A-based reference
« Reply #20 on: November 11, 2017, 04:05:53 AM »
Hello - the last line of the marking on the top should hold the date code and from what I observe this is not in line with LTC's marking system -
- buyer be cautious - most likely this is counterfeit part:

-The metal case is glossy and should be matte
-The TAB indicating pin 1 is at wrong position
-The date code is not inline with LTC's marking system

Best regards
Flintstone

https://www.benl.ebay.be/itm/LTZ1000ACH-CAN-8-Precision-Reference/322563133347?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2055119.m1438.l2649
« Last Edit: November 11, 2017, 09:44:21 PM by Flinstone »
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 1653
  • Country: de
Re: Live design flow of 'FX' 10V LTZ1000A-based reference
« Reply #21 on: November 11, 2017, 06:31:59 AM »
- buyer be cautious - most likely this is counterfeit part.
not only the datecode also the alignment of the marking versus package is fishy
so for me this is 100% counterfeit. (at least re-marked)

with best regards

Andreas
 

Online mimmus78

  • Supporter
  • ****
  • Posts: 591
  • Country: it
    • 118Volt.it
Re: Live design flow of 'FX' 10V LTZ1000A-based reference
« Reply #22 on: November 11, 2017, 09:36:05 AM »
    Hello Andreas - I observe 3 issues:

    The metal case is glossy and should be matte
    The TAB indicating pin 1 is at wrong position
    The date code is not inline with LTC's marking system
    [/list]

    Best regards
    Flintstone


    Pin 1 is not in the wrong position, this is special version of LTZ1000 and must be mounted on the other side (back) of the PCB :-)

    PS: I'm sure only the few will get it ...
    « Last Edit: November 11, 2017, 09:38:21 AM by mimmus78 »
    -> Check my Store @ Ebay
    -> OR my Blog @ 118volt [IT][EN]
     

    Offline eurofox

    • Frequent Contributor
    • **
    • Posts: 673
    • Country: be
    Re: Live design flow of 'FX' 10V LTZ1000A-based reference
    « Reply #23 on: November 11, 2017, 09:52:18 AM »
    - buyer be cautious - most likely this is counterfeit part.
    not only the datecode also the alignment of the marking versus package is fishy
    so for me this is 100% counterfeit. (at least re-marked)

    with best regards

    Andreas

    Many others are available but up to you and wait 6 months  |O
    https://www.benl.ebay.be/sch/i.html?_from=R40&_sacat=0&_nkw=LTZ1000&rt=nc&LH_PrefLoc=2&_trksid=p2045573.m1684
     

    Offline Edwin G. Pettis

    • Frequent Contributor
    • **
    • Posts: 291
    • Country: us
    • The plural of anecdote is not data.
    Re: Live design flow of 'FX' 10V LTZ1000A-based reference
    « Reply #24 on: November 11, 2017, 01:53:21 PM »
    To my knowledge, there is no 'special' layout from LTC, I think what you are referring to was an incorrect board layout and he had to mount the LTZ on the backside of the PCB to make it work.  It wasn't discovered until after the boards were made, it happens.
     


    Share me

    Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
    Smf