Ultra Precision Reference LTZ1000

[PA4TIM]

Hmm, i think I am not good enough in Chinese (or Japanese ?)  to read the links from nukie.

[nukie]

Use Google Chrome there's auto translate. I can't read Chinese but the schematics and diagrams teaches a lot.

[muvideo]

nukie, the LTZ1000 is pretty costly, probably it
would be fun and cheaper form me to experiment
with cheaper references like the lm399, that will
also set a starting point to make comparisons.

What do you think about the SZA263?
How do they compare with LM399? I've seen them (the 263)
used in old fluke standards, and meters, there are
the schematics online, but it seems that
the 399 is simpler to use, am I correct?
Is there any cheap source for SZA263 references,
other than dismantling old fluke meters?

Thanks,
Fabio.

[nukie]

Most sza263 you can find today are mostly reclaimed from recycling Flukes. You can contact Fluke to see if they stock any surplus, they do it for military program maybe?

As for LM399 vs 263, I would settle for the LM399, its widely used and lots of example around. I bought 2 more than 15 months ago and has been burning in since, performance seems to be pretty stable after 4000hours. I am planning a 5x LM399 in parallel when I have more time.

The LTZ1000 I have are also from used equipment. They are sourced directly from a electronics recycling market in China. There are people who are willing to go out and source for the right equipment that carries the specific component.   The components are removed carefully and resold. They are not expensive. Same case with the foil resistors.

There are fake LTZ1000 floating around even the Chinese locals have encounter them. There are also companies that repackage these chips with new gold leads and new case, I suspect that's what you get from EBay.

[quarks]

Use Google Chrome there's auto translate. I can't read Chinese but the schematics and diagrams teaches a lot.

I read all through it. That's is really helpfull. Could you or anybody else download the schematics and can share it? Have you bought the boards and/or can you share where to get them?

[muvideo]

The LTZ1000 I have are also from used equipment. They are sourced directly from a electronics recycling market in China. There are people who are willing to go out and source for the right equipment that carries the specific component.   The components are removed carefully and resold. They are not expensive. Same case with the foil resistors.

There are fake LTZ1000 floating around even the Chinese locals have encounter them. There are also companies that repackage these chips with new gold leads and new case, I suspect that's what you get from EBay.

This is interesting, can you share more details, also in PM if you want.
I'd like to know where to ask for these parts (LTZ1000 and metal foil resistors)
and the order of magnitude for the price I could expect for these parts.
Talking about ebay, what you say is that used parts are more trustable
than new ones?

Thanks,
Fabio.

[Christe4nM]

You can read more about it here
http://home.51.com/jj3055/diary/item/10053954.html

Schematics here
http://bbs.38hot.net/read-htm-tid-36472.html

And finally here how to convert 7V to 10V
http://www.crystalradio.cn/thread-229749-1-1.html

Looking at the pcb I can't stop wondering why these slots are milled. The voltage isn't that high that you need the extra creepage distance, and right now it seems to me it's only weakening the pcb's structural integrity. Worst case it introduces unknown offsets as the pcb will bend a bit either when moved, or with temperature changes. Am I overlooking something here?

[Rerouter]

i would imagine to maintain the accuracy required the leakage currents (even if absolutely tiny) have to be mitigated somehow, and the slot probably is one of the easier ways to accomplish it,

[SeanB]

More there for thermal isolation, as the references are all enclosed in an insulating cover and this removes the path for heat leakage out. The thin connecting traces also reduce heat transfer down them.

[PA4TIM]

Slots in pcbs are for limiting mechanical stress. A pcb is screwed in a cabinet so there is mechnical stress, due to movement, gravity, thermal stress ect. That is why the often make a sort of U slot around a reference.  ( i ould not see it sharp on the pictures so maybe you alk about other slots, because i work mobile)
There is app note about references, can not remember of it was about the LM399 of from LT. They lso describe the best locations on the pcb in relation to stress and return current paths.

[Christe4nM]

I indeed talked about the L-shaped slots around the LTZ1000. Thanks for your reply's.

As far as appnotes go I found at least these ones:
Linear AN42 Voltage Reference Circuit Collection
Intersil AN177 Voltage Reference Application and Design Note

Yet it seems they are not the one you (PA4TIM) mentioned. Might this be the one?
Linear AN82 Understanding and Applying Voltage References

@everyone: the literature references mentioned at the last page of application notes are a great way to delve deeper into any subject.

[PA4TIM]

Yes AN82 is the one that talks about the slots

[muvideo]

Also here linear talks about mechanical stress problems pag 19-20:
http://cds.linear.com/docs/Datasheet/1460fc.pdf

I learned about the strain problems watching pics of internals
of the datron meters, there is an example on these
pics (excellent ones) from tekfan, 4th pic, there are the
two matched zeners:
https://www.eevblog.com/forum/reviews/what-do-you-want-to-see/msg41015/#msg41015

Fabio.

[keen101]

For the board, I haven't seen any for sale, only the ridiculously priced meter boards like Dave showed. Ideally I want some kit or finished board like what's been on the Chinese forums because getting all the precision parts around is single quantities must be PITA. So if anyone wants to make small run, I'm interested for one or two.

I'm helping my dad produce a few replacement boards for the HP 03458-66509 boards. He is working on his prototypes now. I don't know much about it, but he went to a lot of trouble ordering a bunch of precision resistors and stuff. He is now working on getting the ltz1000 chips from linear. He worked for HP/Agilent for 45 years and he calibrated instruments like the HP 3458A. He is recently retired. He is thinking of sell these for cheaper than what the official HP boards would cost. I am helping him make the boards and he is doing the complicated stuff. These boards should be pin compatible. Is there any interest in them here?

[saturation]

Just thoughts: a worthwhile project, but a very limited market.  Owners of working 3458a are not likely to pop in a new board  for fear or having to re-calibrate it, which is not cheap.  Owners of ones requiring service and thus are likely to experiment with new boards are much fewer, reason being if you own this DMM chances are you have it in working order and calibrated, and if you are storing it for the day you'll repair it, meaning you also don't need it right now, its worth $3000+ not working, so its an easy sell on eBay.

You'd have to prove that your aftermarket voltage reference boards are >= the factory boards in terms of accuracy and at least equal in quality.  Until that is known, I'm not sure an owner will risk having their units drift out of spec due to unforseen issues with the boards stability, since just checking the unit via a cal lab is $500-600, and the gold calibration runs past $1300.

For the board, I haven't seen any for sale, only the ridiculously priced meter boards like Dave showed. Ideally I want some kit or finished board like what's been on the Chinese forums because getting all the precision parts around is single quantities must be PITA. So if anyone wants to make small run, I'm interested for one or two.

I'm helping my dad produce a few replacement boards for the HP 03458-66509 boards. He is working on his prototypes now. I don't know much about it, but he went to a lot of trouble ordering a bunch of precision resistors and stuff. He is now working on getting the ltz1000 chips from linear. He worked for HP/Agilent for 45 years and he calibrated instruments like the HP 3458A. He is recently retired. He is thinking of sell these for cheaper than what the official HP boards would cost. I am helping him make the boards and he is doing the complicated stuff. These boards should be pin compatible. Is there any interest in them here?

[krivx]

You'd have to prove that your aftermarket voltage reference boards are >= the factory boards in terms of accuracy and at least equal in quality.  Until that is known, I'm not sure an owner will risk having their units drift out of spec due to unforseen issues with the boards stability, since just checking the unit via a cal lab is $500-600, and the gold calibration runs past $1300.

Wow, is it really that expensive? I had no idea. I assume this is because of the high requirements on the calibrator/source equipment? How does this compare to having, say, a 3.5 or 4.5 digit meter calibrated?

[ve7xen]

Just thoughts: a worthwhile project, but a very limited market.  Owners of working 3458a are not likely to pop in a new board  for fear or having to re-calibrate it, which is not cheap.  Owners of ones requiring service and thus are likely to experiment with new boards are much fewer, reason being if you own this DMM chances are you have it in working order and calibrated, and if you are storing it for the day you'll repair it, meaning you also don't need it right now, its worth $3000+ not working, so its an easy sell on eBay.

I think the idea is more for volt-nut types to build a decent house standard, not people trying to repair 3458's. Depends on cost, but if he can get it significantly cheaper than what the HP boards go for on the used market I expect he'll sell some. Not a large market, but large enough to make it worthwhile I think.

[saturation]

The 3458a is very popular metrology grade top of the line DMM; I wouldn't be surprised if nearly all metrology labs use one in lieu of its competitors, making it a defacto standard top line DMM.

Its thus, a bit more difficult to test the extremes of its range.  As a ball park, calibrating a 5.5 digit DMM is $100-200  in the US, and $200-300 for 6.5 digits, Agilent prices.  A reason for the range is that fully automated DMMs like say the 1272a are just plug-and-play calibration, while older DMM may need more labor.

The $500 calibration is a typical one but for its best specifications, the $1500 cal comes only from Agilent, AFIAK.  Even at $500 most folks will send it to Agilent to insure its done properly.

https://service.tm.agilent.com/infoline/Product-Service.aspx?pn=3458A&countryid=114

https://service.tm.agilent.com/infoline/product-service-query.aspx


You can type your Agilent device above and the website will quote you a charge, it varies by country.

The cost is proportionate to the labor, and at the lowest cost its basically mostly the cost of the calibrator and or references, buying and maintaining it.

You'd have to prove that your aftermarket voltage reference boards are >= the factory boards in terms of accuracy and at least equal in quality.  Until that is known, I'm not sure an owner will risk having their units drift out of spec due to unforseen issues with the boards stability, since just checking the unit via a cal lab is $500-600, and the gold calibration runs past $1300.

Wow, is it really that expensive? I had no idea. I assume this is because of the high requirements on the calibrator/source equipment? How does this compare to having, say, a 3.5 or 4.5 digit meter calibrated?

[keen101]

I think the idea is more for volt-nut types to build a decent house standard, not people trying to repair 3458's. Depends on cost, but if he can get it significantly cheaper than what the HP boards go for on the used market I expect he'll sell some. Not a large market, but large enough to make it worthwhile I think.

Okay i was just curious. I thought maybe something like this project might work for some of you, but like i said before i don't know much about it.

http://www.maxmcarter.com/vref/

[alm]

This is exactly the volt-nuts kind of application that ve7xen was referring to. A good implementation of the LTZ1000 is non-trivial and requires several fairly exotic components. For the right price, I think you will get some buyers on this forum. No idea how many. I would certainly consider it.

I think the street price for a used 3458A reference was under $100 last time I saw one for sale on eBay. One advantage that the real Agilent units would have is that the LTZ1000 would be selected and burned in. I seem to recall a fair amount of variation in noise of the LTZ1000. This assumes that the references for sale on eBay were not factory rejects.

If he has access to a calibrated 3458A or similar DMM, than he might be able to offer calibration, as a high-end equivalent to the Geller Labs / Voltage Standard products. The reference would need to be burned in for bests results, which is likely to expensive/complex to make it worthwhile.

[quarks]

Use Google Chrome there's auto translate. I can't read Chinese but the schematics and diagrams teaches a lot.

I read all through it. That's is really helpfull. Could you or anybody else download the schematics and can share it? Have you bought the boards and/or can you share where to get them?

So far nukie has not responded. Does anyone else know, where to get the boards nukie uses and can share the schematic?

[quarks]

I'm helping my dad produce a few replacement boards for the HP 03458-66509 boards. He is working on his prototypes now. I don't know much about it, but he went to a lot of trouble ordering a bunch of precision resistors and stuff. He is now working on getting the ltz1000 chips from linear. He worked for HP/Agilent for 45 years and he calibrated instruments like the HP 3458A. He is recently retired. He is thinking of sell these for cheaper than what the official HP boards would cost. I am helping him make the boards and he is doing the complicated stuff. These boards should be pin compatible. Is there any interest in them here?

that looks good. Do you and your Dad want to offer/sell these boards?

[keen101]

that looks good. Do you and your Dad want to offer/sell these boards?

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.

[JBeale]

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.

[amspire]

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/