-
I see a ref01 , what are the other 2 cans.
Some sort of matched fets (marked K146), looks like two TO-92's with a little aluminum hat.
edit: looks like the K146 is a pair of 2SK146 https://www.semicon-data.com/transistor/tfet/k0/K146.html -
cellularmitosis
-
Ok, here is an initial effort to find the R4/R5 temperature set-point "offset" of an LTZ1000 (non-A).
There are two ways to find this offset: either use a fixed R4/R5 and place the circuit into an oven, slowly ramping the temperature until the voltage falls out of regulation, or use a fixed temperature (ambient) and adjust the R4/R5 ratio until fall into and/or out of regulation. Here, I'm using the latter technique.
The setup: R4 of a px-ref board has been broken out to a decade box, so that I can manipulate the temperature set-point. The meter is a Keithley 2015 at "medium" speed with a 10-reading moving average filter.
I've spliced the footage from the three cameras together, but I haven't edited the result into smaller chunks, so I'll just list a timeline of here:
0:00 Camera sync-up.
0:30 Approaching the ambient set-point from below (100-ohm resolution). 10.4k to 10.5k to 10.6k create steps of about 29uV. We are not yet in thermal regulation.
1:00 The 10.6k to 10.7k step diverges from the 29uV behavior.
1:30 10.8 to 10.9 to 11.0 create steps of about 77uV. We are now in thermal regulation.
1:45 Decreasing the set-point while so close to the regulation limit results in very slow response.
2:10 Switching to 10-ohm resolution. When just under regulation, adjusting the resistance has no effect (10.85 through 10.89).
2:30 10-ohm steps create steps of about 8uV when in regulation (probably, the steps are actually 7.7uV).
3:07 Using a fan to create forced cooling, to approach the ambient set-point from above.
4:00 10.86k is in regulation.
4:15 10.85k is in regulation.
4:24 10.84k is in regulation.
4:35 10.83k is in regulation (using the fan).
4:45 10.82k is in regulation (using the fan).
4:57 10.81k is in regulation (using the fan), but self-heating due to zener-current slowly pushes it out of regulation.
5:22 Self-heating drift of 8uV up from the 10.81k set-point, which also means the 10.82k set-point isn't with regulation at equilibrium.
5:39 Drift reaches 16uV, so ambient regulation must be above the 10.83k set-point.
5:55 24uV drift; 10.84k is not within regulation at equilibrium.
6:12 31uV, so 10.85k isn't regulated.
6:45 39uV, so 10.86k isn't regulated.
7:15 7.7uV times 6 steps would be 46uV. The drift is really slowing down as we cross over 47uV, so 10.87k isn't in regulation, but we are very close now (at ~23.5C).
7:25 At 48uV of drift, we verify that 10.87k isn't in regulation, and we verify that 10.88k is in regulation. We then decrease back to 10.87k, and the drift starts falling again, indicating the chip is cooling off, which verifies that 10.88k is in regulation at 23.46C.
So, at 23.46C, we drop out of regulation somewhere between an R4 value of 10.88k and 10.87k.
https://youtu.be/mbSmjaNA8xc
-
Having gone through this exercise and thought about this a bit more, I now understand why the LTZ1000A needs a higher thermal set-point than the non-A chip. If the thermal resistance to ambient is ~5x greater, then the same (~4mA * 7V) 28mW of zener power dissipation will result in more self-heating in an "A" chip than a non-A chip.
Thus, for a given ambient environment, you must operate the A chip at a higher thermal set-point to maintain regulation. -
Having gone through this exercise and thought about this a bit more, I now understand why the LTZ1000A needs a higher thermal set-point than the non-A chip. If the thermal resistance to ambient is ~5x greater, then the same (~4mA * 7V) 28mW of zener power dissipation will result in more self-heating in an "A" chip than a non-A chip.
Thus, for a given ambient environment, you must operate the A chip at a higher thermal set-point to maintain regulation.
Yep, per the data sheet ~ 10 deg C higher set point for the 'A' version is recommended. -
Completed the rest of the R4 / temperature "ambient offset" measurements, both for lowimpedance's LTZ1000's as well as some LTZ1000A chips (and an LTZ1000) which I have on hand. All measured in the same (socketed) board, using the same decade resistor setup shown in the above video.
I found, on average, that the A chip needs set-point which is 228 ohms higher than non-A chips. Assuming a slope of 20 Ohms / C near 23C, this means the A chip needs an set-point which is 11.4 degrees higher than the non-A chip. Pretty close to the datasheet values.
-
Approaching this from another angle:
I measured the voltage across the 120R zener resistor at 0.541V, which makes the zener current 4.5mA. With Vref at 7.077V, that makes Vz 6.536V, which means zener self-heating is 29.5mW.
For a non-A chip with 80C/W thermal resistance, that's a 2.357C temperature rise.
For an A chip with 400C/W, that's a 11.787C temperature rise.
That's a difference of 9.429C -- pretty close to the 10C datasheet value. However, I'm not accounting for the heating due to current in the two transistors in the LTZ1000, which would push my figure a bit closer to 10C.
Taking my 228 Ohm average difference, that makes a slope of about 24 Ohms / C at ambient.
-
However, using the formula (which I think Dr. Frank derived), the R4 operating point agrees pretty closely, but the slope near 23C is more like 40 Ohms / C.
-
Ohhh ho ho ho look what arrived in the mail today!
As far as who to give them to... perhaps I should draw six random names from the list of members who are currently signed up for cal club round 2? And if one of them wishes to decline, I'll draw another name instead. If you have any ideas on this matter, feel free to share them!
That would be amazing! . I've been waiting to pull the trigger on building my own ltz ref. Now may be the time!
-
lowimpedance has been very generous and has donated these parts "to the cause"
I think its time we had a few more members on this forum with LTZ references! I'll add a bit to his generosity and throw in a circuit board with each.
As far as who to give them to... perhaps I should draw six random names from the list of members who are currently signed up for cal club round 2? And if one of them wishes to decline, I'll draw another name instead. If you have any ideas on this matter, feel free to share them!
Man, I cannot believe almost a month has already slipped by. Unfortunately hobby time has been very scarce lately.
I had hoped to have a working demo of my latest PX ref board ready by now, so that I could offer a board + the resistor chip along with the LTZ from lowimpedance, but I have held up this thread for 3 weeks already, and I don't want to stall any further.
So, here are 6 random names drawn from this go-round of the cal club:- TWMIV
- Svgeesus
- orin
- GEOelectronics
- Vacuuminded
- bitseeker
Congrats to you guys!
Along with the LTZ I will include a PX-ref board of your choice:
A v2.4.2 board, which is the same as 2.4.1 but with some slight tweaks to the thermal copper ring which surrounds the LTZ. This board fits inside of a TEKO 371.16 steel enclosure.
A v2.5 board, which is a longer version of 2.4.2, which is meant to mount the input and output filter ferrites directly onto the board (see photo). This fits in the slightly longer TEKO 372.16 steel case.
A v2.6 board, which uses an OHMTEK multi-resistor chip which I found on taoboa for divider and the two 70k resistors. A 120R resistor needs to be supplied separately, and the footprint for that will accommodate a metal foil or wire-wound resistor. Note however that I soldered up an example of this board and for some reason it rails the heater control fully on, so probably best not to choose this option until I get the issue figured out.
I'll message each of you to arrange shipping and such
-
Quote
Just curious, has anyone ever made any correlation to a particular vintage being more stable than others? Would NOS be desired for any reason over the newer parts? Does aging only occur when the part is in an active circuit or does some ageing occur over time even in storage?
QuoteTo be a fair comparison that should be the case. Still it would be interesting to see long term drift measurements with one of these oldies.
Kind of sad that we now won't see a fair comparison between the vintage specimen as they will now be seperated, tested under different conditions on different board designs and with different instruments. Could have been an interesting investigation if there is any hint on a correlation
-branadic-
-
Today I found some in dumpster too!
Old days Linear was better than today, that's for sure. Nice package with ESD-safe sponge. Modern orders they just toss poor LTZ's into bag and ship it, flying and non-secured inside the bag.
Some speciol chips here too.
Added into my 1000A stash (new arrivals = row on top.)
LM394H and cute dual matched JFET (?) pairs (I'd assume that is from Keithley 146?) will be interesting feat too.
-
Quote
Today I found some in dumpster too!
Can you please guide me to what kind of dumpster I have to look for such a find?
-branadic- -
Quote
Today I found some in dumpster too!
Can you please guide me to what kind of dumpster I have to look for such a find?
-branadic-
I sustain this, I want one as well, if not a full carton
Cheers,
DC1MC -
Can you please guide me to what kind of dumpster I have to look for such a find?
Tin sold his soul to Satan a long time ago. Those parts came out the dumpster behind some low-rent brothel in hell.
-
Can you please guide me to what kind of dumpster I have to look for such a find?
Those parts came out the dumpster behind some low-rent brothel in hell.
How did you know 
. Now that has made my day
AND........
From now on y'all keep ya greasy mits outa my dumpster......
-
Today I had some Australian dumpster find, too. LTZ1000CH with date code 8628, 8716, 8830 as well as 9015. Seems like some worldwide reference infection is going around.
-branadic- -
Hello,
Suitable to the 1986 devices the data sheet from the LT 1986 databook (printed in 1985).
So the 86 datecode are very early (preliminiary?) devices.
with best regards
Andreas
-
lowimpedance has been very generous and has donated these parts "to the cause"
I think its time we had a few more members on this forum with LTZ references! I'll add a bit to his generosity and throw in a circuit board with each.
As far as who to give them to... perhaps I should draw six random names from the list of members who are currently signed up for cal club round 2? And if one of them wishes to decline, I'll draw another name instead. If you have any ideas on this matter, feel free to share them!
I had hoped to have a working demo of my latest PX ref board ready by now, so that I could offer a board + the resistor chip along with the LTZ from lowimpedance, but I have held up this thread for 3 weeks already, and I don't want to stall any further.
So, here are 6 random names drawn from this go-round of the cal club:- TWMIV
- Svgeesus
- orin
- GEOelectronics
- Vacuuminded
- bitseeker
Congrats to you guys!
Along with the LTZ I will include a PX-ref board of your choice:
A v2.4.2 board, which is the same as 2.4.1 but with some slight tweaks to the thermal copper ring which surrounds the LTZ. This board fits inside of a TEKO 371.16 steel enclosure.
A v2.5 board, which is a longer version of 2.4.2, which is meant to mount the input and output filter ferrites directly onto the board (see photo). This fits in the slightly longer TEKO 372.16 steel case.
A v2.6 board, which uses an OHMTEK multi-resistor chip which I found on taoboa for divider and the two 70k resistors. A 120R resistor needs to be supplied separately, and the footprint for that will accommodate a metal foil or wire-wound resistor. Note however that I soldered up an example of this board and for some reason it rails the heater control fully on, so probably best not to choose this option until I get the issue figured out.
I'll message each of you to arrange shipping and such
Thank you, lowimpedance and cellularmitosis for the generous giveaway! And thanks to Andreas for the period-matching databook. -
Hello,
This has taken forever and a day but I have finally tested the 1/f noise of these LTZ's. I'll be shipping them out shortly!
I've followed the same setup as I used last time: https://github.com/cellularmitosis/logs/blob/master/20180324-ltz1000-1f-noise/README.md
The probe multiplier has been set to 100x, and the 1/f noise amplifier has a gain of 10,000x, so just mentally drop down one SI unit class in your mind when looking at these scope shots (i.e., if it says mV, that's actually uV).
Here's the LTZ with date code 8628, on which I've placed one dot:
-
Here's the chip with date code 8628, which I've marked with two dots:
-
Here's the chip with date code 8830:
-
Here is the chip with date code 9015, which I've marked with one dot.
-
Here's the chip with date code 9015, which I've marked with three dots:
-
And here's the chip with date code 9015 with four dots.
(the 6th chip, 9015 with three dots, was mailed out to Svgeesus before I had the chance to test the 1/f noise, sorry!)
