Author Topic: Ultra Precision Reference LTZ1000  (Read 874436 times)

0 Members and 1 Guest are viewing this topic.

Offline JBeale

  • Regular Contributor
  • *
  • Posts: 206
Re: Ultra Precision Reference LTZ1000
« Reply #75 on: February 16, 2013, 12:59:50 am »
I might be interested in a LTZ1000 reference board, but my interest level would depend on the price :-)   I am just starting to test a batch of LM399 references and finding that each one is quite individual as far as microvolts of noise/drift in any given frequency range. I care most about frequencies of 1 Hz and lower, so maybe better to call it "drift".

"...now working on getting the ltz1000 chips from linear."   From this link http://www.linear.com/purchase/LTZ1000 and then
http://shopping.netsuite.com/s.nl?ext=F&c=402442&sc=2&category=&search=LTZ1000
it looks like "Linear Direct" currently has 14 of the LTZ1000ACH#PBF parts in stock, but none of the other types.  But they also note "a minimum buy may apply".
Then again, they also said that on the order form about the LM399's I got, but then shipped me the six parts I requested anyway.
« Last Edit: February 16, 2013, 01:09:25 am by JBeale »
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #76 on: February 16, 2013, 06:10:17 am »
The device I use as my reference is a Thaler VR305A hybrid IC. 5V 0.01% initial accuracy. 0.6 pmm/C without an oven or on-chip heater. If you want to put them in an oven, it includes a diode junction for as a temp sensor. That means they are better then total 0.001% stability as the temperatures varies from 17 degC to 33 degC and no warm up time - good enough for me.  The LTZ1000 can obviously get an order of magnitude better temperature stability with its built in heater but you do have to give it over 5 minutes to stabilize after turn on.

The Thaler references use a buried zener based reference,  and I think resistors and thermistors on the ceramic substrate are then laser trimmed to zero the temp coefficient.

Long term stability is worse then the LTZ1000 - 6ppm in 1000 hours. Does it stabilize after 1000 hours? It seems to, but I don't know. I don't have any figures in how they actually perform long term after they have burnt in. There is much less in a  LTZ1000, so I guess there are much fewer components that can drift.

The problem with the Thaler references is they used to be very affordable - I think I got 5 for about $20 each from Thaler over 10 years ago.

Thaler were bought by Cirrus Logic (Apex Microtechnology) and since then, every time I look, the price has gone higher. A couple of years ago, they were $70 and Digikey have them currently for $93. Apex Microtechnology website does seem to offer samples. http://www.apexanalog.com/products/voltage-sine-wave-references/
 

Offline JBeale

  • Regular Contributor
  • *
  • Posts: 206
Re: Ultra Precision Reference LTZ1000
« Reply #77 on: February 16, 2013, 04:20:33 pm »
I made a table of the specs on all the precision references I found referenced on the "volt-nuts" list. I lost interest in the VRE305 after I saw the price on DigiKey. It is by far the most expensive one on the list; twice as expensive as a LTZ1000 from Linear Tech (although to be fair, the LTZ1k needs several other expensive parts in the standard circuit).  Both the Apex product page and the data sheet actually call the VRE305 a "low-cost" part, which is actually true by reference to some of their other parts... Digikey lists the VRE102CA at $262.   I would be impressed if they were willing to sample any of these gold-plated parts to a hobbyist.
« Last Edit: February 16, 2013, 04:22:48 pm by JBeale »
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #78 on: February 16, 2013, 04:38:14 pm »
I made a table of the specs on all the precision references I found referenced on the "volt-nuts" list. I lost interest in the VRE305 after I saw the price on DigiKey. It is by far the most expensive one on the list; twice as expensive as a LTZ1000 from Linear Tech (although to be fair, the LTZ1k needs several other expensive parts in the standard circuit).  Both the Apex product page and the data sheet actually call the VRE305 a "low-cost" part, which is actually true by reference to some of their other parts... Digikey lists the VRE102CA at $262.   I would be impressed if they were willing to sample any of these gold-plated parts to a hobbyist.
The datasheet Apex used is the Thaler datasheet with the name Thaler replaced. Nothing else was ever changed except for a fivefold increase in price. No new devices have been designed. Sad really.
 

Offline fmaimon

  • Supporter
  • ****
  • Posts: 165
  • Country: br
Re: Ultra Precision Reference LTZ1000
« Reply #79 on: February 16, 2013, 04:39:36 pm »
There is someone on ebay selling VRE305KS for less than $30 but there is no picture of the device. I've sent him a message asking for a photo of the IC.
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #80 on: February 16, 2013, 11:10:56 pm »
There is someone on ebay selling VRE305KS for less than $30 but there is no picture of the device. I've sent him a message asking for a photo of the IC.
If it is the ebay sale for US$22, it looks a good deal. They are the devices that can go down to -40C. Mine are only rated down to 0C.

The IC's are surface mount versions of the DIP8 package, like mine. Here is a photo, if it helps:



Also I just found there very  cute 8mm x 8mm peltier devices for $4 inc postage that can shift up to 1/3W of heat. I thought they could be good for making a small temp controlled "oven" for voltage references, crystals, etc. It means an operating temp of something like 30 degC can be chosen, rather then having to pick something well above normal ambient, like 50 degC. The total heat you have to dissipate from the peltier is about 1.2W so that can be a small heatsink, or directly to a metal box. I have ordered a few to play with.


http://www.ebay.com.au/itm/TES100701-DC-0-8V-0-33W-Peltier-Cooler-Thermoelectric-Cooler-Cooling-8-x-8-mm-/110900388714?
« Last Edit: February 16, 2013, 11:21:10 pm by amspire »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: Ultra Precision Reference LTZ1000
« Reply #81 on: February 17, 2013, 12:02:39 pm »
LM399H and LM399AH can be bought at www.electronic-search.de in germany. I did that a few months ago and made a small board with this milled slots to keep away mechanical stress and minimize thermal drop between the pins resulting in thermo voltages. Keep in mind that LM399 is available from LT and National Semiconductor, both with some small difference in spec. You could also ask them for the LTZ1000.

I'm sensitized with references at ebay. I bought a few MAX6350MJA (Buried Zener in a CERDIP package) from ebay and hoped that they were also used and recycled, because MAXIM don't stock them anymore. CERDIP because of the humidity influence to references plastic package is sensitive to.
The brand printing was the first indication, didn't look like MAXIMs original logo. The package marking on the bottom showed some code Linear Technology uses for their products (N12345) instead of a 7 digit letter combination plus a 4 digit date code, what is common. And also measurements showed that there is nothing but a bandgap reference inside, presumably a Ref02 or similar.
MAXIM now knows about that asian seller.

BTW: If some of you have a few (used) MAX6350MJA in private stock and want to sell them send me a message.
« Last Edit: February 17, 2013, 01:12:02 pm by branadic »
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #82 on: February 17, 2013, 01:42:07 pm »
branadic,

Have you looked at the Thaler/Cirrus/Apex VRE305KS that was mentioned above?

http://www.ebay.com/itm/Low-Cost-Precision-Reference-IC-VRE305-VRE305KS-NEW-/221065482973

SOP-8 Ceramic package and almost compatible with the Maxim device. Most specs as good or better then the Maxim chip. Major difference is that the VRE305KS has a seperate GND and Reference GND, so that the reference output is not affected by the IC supply current. Other differences are that the VRE305 required 13.5V minimum compared to Maxim's 8V and consumes slightly more current, but has better long term stability figures. The VRE also has a temp sensor junction in case you want to use it in an oven.
 

alm

  • Guest
Re: Ultra Precision Reference LTZ1000
« Reply #83 on: February 17, 2013, 02:00:35 pm »
Any idea how it compares to the much cheaper LM399 in real world figures? The spec for long term stability of the LM399 is much worse (20 ppm / 1k hrs typ), but we know from the many devices designed around it (including the far majority of the 6.5 digit DMMs) that it can perform much better. Is the same true for the VRE305? I guess it should be better since it runs at a much lower temperature, everything else being equal. But of course everything else is not equal.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: Ultra Precision Reference LTZ1000
« Reply #84 on: February 17, 2013, 02:52:30 pm »
branadic,

Have you looked at the Thaler/Cirrus/Apex VRE305KS that was mentioned above?


Yes, I once did, but I only payed 5.80€ for an new LM399H, you can guess what I prefered to buy.

"...Fairic is a Hong Kong based company with China offices in different provinces..."

ebay + China + Precision Voltage References --> all my alarm bells are ringing. Thanks for the link, but I would never again move any finger to click on the "buy it now" button. The last seller also had nearly 100% positive feedback. I'm not sure if the guys who bought the MAX6350MJA too where able to measure their devices? They should have noticed that the device behaves contrary the datasheet.

This is a typical example of such a fake:
http://www.ebay.com/itm/Maxim-MAX297-MAX293MJA-MAX294MJA-MAX297MJA-8-Cerdip-IC-/260653697527?pt=LH_DefaultDomain_0&hash=item3cb02bc9f7
Notice the brand, this is how the MAX6350MJA fake was looking like.
« Last Edit: February 17, 2013, 03:11:35 pm by branadic »
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline fmaimon

  • Supporter
  • ****
  • Posts: 165
  • Country: br
Re: Ultra Precision Reference LTZ1000
« Reply #85 on: February 19, 2013, 01:11:59 pm »
The guy sent me this photo. It has some strange things.
  • Its mounted in a board. If he is selling these things, he would need to unsolder it at the time of sale. Why didn't he do it already?
  • It was listed as a VRE305KS. The picture is from a VRE305A
  • The date code is 0001. Although it is possible that the IC was produced on the first week of 2000, it is strange.
  • The markings are somewhat different from Richards
So, what do you think?
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #86 on: February 19, 2013, 02:45:28 pm »
The ThC logo is the older logo of Thaler Corp, and my Thaler datasheet is Rev D 2001. I have posted it below.

The USA marking and omitting the "S" off the end is in line with the way Thaler do it. I thought I had purchased mine in the early 2000s, but it looks like I got it in 2007, so I have a late model with the newer TC logo.

The case style is unusual, so there may not be any other reference ICs in an identical case. Mine has no visible ceramic cement between the upper and lower half of the case, but that many be a refinement after 6 years. The metal contact construction in this 2001 IC are slightly different to my 2007 IC.

Why has he got a photo of a more accurate "A" model when he is offering a "K" model? I do not know the answer to that question, except for the fact that to quote the part as a VRE305KS, he is not reading it off a chip - that may be the code of a tube, box, bag, or somewhere else. The IC will have the part number VRE305K.

Given the market price, it is possible that someone has got a REF02A chip (or similar) and put it in the SOP-8 ceramic package to make a fake. It would work in the same circuit, except the initial accuracy would be over 10 times worse, and the long term drift over 20 times worse. In the photo, the IC is sitting on a PCB, and may be soldered onto the PCB. It is odd.

I did an image search on Google, and no match for that photo was found, so the seller may have actually taken the photo.
« Last Edit: February 19, 2013, 03:00:27 pm by amspire »
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: Ultra Precision Reference LTZ1000
« Reply #87 on: February 23, 2013, 10:47:32 am »

He said he will probably try listing them on ebay soon. Maybe by next week, but i'm not sure. I honestly don't know if anyone will want to buy them, but we will see.





Any news about that? Couldn't find the offer yet, but was interested in the price too. Have them seen a burn-in test or are they just new and directly soldered to the board?
Strange, can't find any "LTZ1000" with it as keyword on e**y, only by searching for "HP 3458" and similar I get a few hits. What's going on there?

I made a table of the specs on all the precision references I found referenced on the "volt-nuts" list.

You'd like to share your table with us?
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline JBeale

  • Regular Contributor
  • *
  • Posts: 206
Re: Ultra Precision Reference LTZ1000 (table of Vref data)
« Reply #88 on: February 27, 2013, 07:31:59 am »
I don't know if I can attach the spreadsheet here...  when I print a PDF it does not look good (long rows are split across too many pages). Anyway you can find my list of voltage reference specs here:  http://www.bealecorner.com/pcb/vr1/Vref-table-Feb2013.xls

I did this just for my own purposes, it may not be accurate, and is surely not complete. But if you do see any mistakes, let me know.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 29960
  • Country: au
    • EEVblog
Re: Ultra Precision Reference LTZ1000 (table of Vref data)
« Reply #89 on: February 27, 2013, 07:43:43 am »
I don't know if I can attach the spreadsheet here...  when I print a PDF it does not look good (long rows are split across too many pages). Anyway you can find my list of voltage reference specs here:  http://www.bealecorner.com/pcb/vr1/Vref-table-Feb2013.xls

I did this just for my own purposes, it may not be accurate, and is surely not complete. But if you do see any mistakes, let me know.

Reposted as a handy PDF.
Nice work.

Dave.
 

Offline JBeale

  • Regular Contributor
  • *
  • Posts: 206
Re: Ultra Precision Reference LTZ1000
« Reply #90 on: February 27, 2013, 07:50:27 am »
By the way, I just got two of the LTZ1000ACH (direct from Linear Tech- the ebay versions did not look promising) and am trying one out using the simplified +Vref schematic that uses a single LT1006 opamp, and does not use the heater. 

Both this and the more complex heater-stabilized "7V Positive Reference Circuit" schematic on p.6 of the LTZ datasheet use a 1N4148 diode from the opamp output to the Vref out (LTZ pin 3, zener cathode).  Can anyone explain what that diode is for? This is a single-supply opamp circuit so there is no chance of the output going negative. Is it some protection for the opamp in case of a badly-behaved external load, or something about power-up or power-off behavior?

I have tried this circuit with the diode and without, and as far as I can tell, the only effect of the diode is to require the opamp supply voltage to be 0.7 V higher than it would otherwise need to be to prevent a dropout (unloaded Vsupply threshold of 9.1 V instead of 8.4 V , in the case of a LT1013 and my particular LTZ1000A which provides a Vout of 7.2406 V, with the temperature compensating R1 at 25 ohms and Iz = 4.27 mA).

EDIT: I suppose the circuit is (meta-)stable with both inputs and the output sitting at 0.000 V, so there might be a startup problem. If so, does that diode alone actually fix it?  There is a version with an additional 10k pullup on the output (Figure 65 on p.15 of AN42, http://cds.linear.com/docs/en/application-note/an42.pdf ) suggesting that is the issue.

EDIT2: Thanks alm for pointing to the answer; I should have seen those threads before.
« Last Edit: February 27, 2013, 07:23:35 pm by JBeale »
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 6203
  • Country: 00
Re: Ultra Precision Reference LTZ1000 (table of Vref data)
« Reply #91 on: February 27, 2013, 08:08:01 am »
JBeale, thanks for the spreadsheet, really handy.  :-+

Btw, Intersil also carries quite some number of Vref ICs as well.

Just read the link in your sheet -> http://www.febo.com/pipermail/volt-nuts/2010-September/000447.html , quite interesting result about MAX6350.

Quote pics :



« Last Edit: February 27, 2013, 08:52:51 am by BravoV »
 

alm

  • Guest
Re: Ultra Precision Reference LTZ1000
« Reply #92 on: February 27, 2013, 07:12:14 pm »
Both this and the more complex heater-stabilized "7V Positive Reference Circuit" schematic on p.6 of the LTZ datasheet use a 1N4148 diode from the opamp output to the Vref out (LTZ pin 3, zener cathode).  Can anyone explain what that diode is for?

As you already guessed, it is for startup. See this thread on EEVblog and this thread on volt-nuts for more information.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: Ultra Precision Reference LTZ1000
« Reply #93 on: February 27, 2013, 10:50:16 pm »
Andreas Jahn determined in his measurements, that references in plastic packages are very sensitive to humidity compared to their brothers in ceramic package. This sensitivity results in additional drift next to tempco. This is also a reason why I was searching for the MAX6350MJA, it's the CERDIP version.
It seems, that I now found a trustable source for that part, but it's even more expensive (16,10€) compared to the LM399AH (9,95€).
The LTC6655 is now available in a ceramic smt package (LS8) with excellent specs, but up to now only available as a 2.5V reference. This could be a worthy alternative to the good old stuff, awaiting the 5V version. And as by a wonder it is again some Linear part.
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline JBeale

  • Regular Contributor
  • *
  • Posts: 206
Re: Ultra Precision Reference LTZ1000
« Reply #94 on: March 13, 2013, 05:10:39 am »
Quote
The LT1021-7 also comes to mind-- with fewer internal resistors, this part will show exceptionally low long-term drift.  Again, you can find the hermetic part from various vendors, and you would be wise to ovenize it (along with any circuitry you might add to gain it up to 10V).

Interesting you should mention the LT1021-7. I have three of those (plastic DIP) that I've been running for several weeks. I am seeing several ppm of short-term drift (10-100 sec) but just a few inches away, a pair of MAX6350 parts (also plastic DIP) is stable to less than 1 ppm over the same time period.
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: Ultra Precision Reference LTZ1000
« Reply #95 on: March 13, 2013, 09:53:06 pm »
Quote
As far as long term drift goes, probably the Burr-Brown REF102 (now the TI REF102) is the best available voltage reference that is not ovenized

Well, we need to have a look on the new Linear Technology LS8 packaged voltage references, such as the mentioned LT1236-LS8 or LTC6655-LS8.
It's interesting that manufactors seem to have "unlearned" what was already done in the LTZ1000. I wonder if there will be a heated, 5V output buried zener in LS8 package or similar in near future, supplied with moderate voltages around 5V, a (still) typical value in todays analog circuits.
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 6203
  • Country: 00
Re: Ultra Precision Reference LTZ1000
« Reply #96 on: March 14, 2013, 02:05:03 am »
At Intersil's voltage reference they're using the term so called "floating gate" technology, is this the same as buried zener ?

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #97 on: March 14, 2013, 02:32:22 am »
At Intersil's voltage reference they're using the term so called "floating gate" technology, is this the same as buried zener ?
No - very different. The floating gate depends on a stored charge as the voltage reference - a bit like a eprom memory cell.
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 6203
  • Country: 00
Re: Ultra Precision Reference LTZ1000
« Reply #98 on: March 14, 2013, 02:56:40 am »
At Intersil's voltage reference they're using the term so called "floating gate" technology, is this the same as buried zener ?
No - very different. The floating gate depends on a stored charge as the voltage reference - a bit like a eprom memory cell.
Ah thanks, no wonder they're worrying that the trapped charge will somehow "leaked" caused by xray like their appnote -> X-Ray Effects on Intersil FGA References

Curious if there are any advantages of this floating gate over other old & proven technology like buried zener ? Cheap to produce maybe ?  :-//

PS : I've been "imagining" to use this xray effect to deliberately fine tune the trapped voltage to get high absolute accuracy, and once finished, sealed it in thick lead box.  ;D
« Last Edit: March 14, 2013, 04:26:58 am by BravoV »
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3783
  • Country: au
Re: Ultra Precision Reference LTZ1000
« Reply #99 on: March 14, 2013, 11:06:55 pm »
I have never seen anyone use a floating gate reference in lab gear. I have never used one myself either so I have no experience, but I prefer standards somehow constructed as hardware - like a buried zener - rather then something that sounds as vulnerable as a stored charge that can be affected by radiation and probably by heat as well.

I wonder how much the reference changed when if goes though soldering?

Buried zener references have a good record of maintaining accuracy after 20 years or more. Wouldn't have a clue about the floating gate reference performance over 20 years.
« Last Edit: March 14, 2013, 11:10:18 pm by amspire »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf