Build a 10V-Buffer inside the Teko371 LTZ1000 Box

[hwj-d]

Ok, that's an experiment. I wanted to see if it's possible, to build a 10V buffer together with the LTZ1000 inside this little teko-371 box.

But first, to feed my refs, I build a little compact PS 2x independet 15V based on 2x LT1763:

[hwj-d]

The buffer is build like kj7e's one (thanks to him), with little differences. I used a LTZ1000 non 'A' type, wich i already build inside the teko:









... more to come

[hwj-d]

The result is amazing (measured with 10 megOhm impendance, 100plc) :

[Andreas]

But first, to feed my refs, I build a little compact PS 2x independet 15V based on 2x LT1763:

Hmm,

how do you limit the input voltage to the LT1763 to the maximum rating of 20V with a 15V transformer?
I never use more than 12V transformers for the LT1763.

with best regards

Andreas

[hwj-d]

But first, to feed my refs, I build a little compact PS 2x independet 15V based on 2x LT1763:

Hmm,

how do you limit the input voltage to the LT1763 to the maximum rating of 20V with a 15V transformer?
I never use more than 12V transformers for the LT1763.

with best regards

Andreas

I knew you noticed that. 
Yes. The input voltage is max 20V. But nothing happens, also with max load (300mA testet over the time. Edit: transformer worst, max load LT1763 is 500mA). Now I have another transformer for that. But it seems, the LT1763 out of spec can stand it. I was astonished too, but they doesn't scratch that, not going hot, or have any ripple or something. So, i leave it so, first.

Thanks, you mentioned that, for those who would build that PS too. 

[SeanB]

20V absolute maximum mean that is the voltage at which they guarantee every part will still function. Generally most will survive slightly higher voltages, at ambient temperatures, but over the full operating temperature range the 20V is what all the devices will survive at least for 1000 hours. you can get away with it for your particular samples, but another batch might not be so lucky.

I had a computer where the 5V power supply decided to go out of regulation, applying the unregulated 14V input ( the designers probably wanted a lot of margin in the supply, and it certainly would survive a really big input Ac voltage variation of -50% to 100% of the rated 115VAC 400Hz 3 phase input no problem short term) to the entire board of TTL and NMOS IC's. Only failure was a single hex TTL 30V rated inverter, which did not like having 14V applied to it and went first short circuit, melted the flatpack lid off and then blew it's bonding wires off. The rest of the TTL all survived, and in fact were still operating correctly logically, and still gave valid results. Change the board ( lot of them spare, they rarely gave issues so plenty of new old stock around) and tested after replacing every part in the 5V power supply, including the failed crowbar that should have blown the fuse when the voltage went over 7V. Was not going to look for the failed part, easier to replace every part, including the resistors and the ceramic capacitors instead. only things not changed were the 2 unobtanium block tantalum capacitors in there, 8200uF 25V and 10000 uF 16V, as they were a lovely Vishay part, and lead times on those were in the decade range, seeing as I had orders for that long when I arrived there, and they never turned up. Saw them in the online Vishay catalogue years later as well, with a "contact factory for price and lead time" as the only ordering information. the others were replaced with regular 105C electrolytics, as they would last as long as the rubber seal wet tantalums in service.

[CalMachine]

Do you want your black output feed-through sitting directly above Q1?

[pelule]

SeanB
20V absolute maximum mean that is the voltage at which they guarantee every part will still function.

I am working over 20 years for semicundoctor supplier - so I have to say: your statement may mislead.
As mentioned at the notes of the LT1763 datasheet

Note 1: Stresses beyond those listed under Absolute Maximum Ratings may cause permanent damage to the device. Exposure to any Absolute Maximum Rating condition for extended periods may affect device reliability and lifetime.

20V in real life means 20V + variations (noise, temp-drift...) thus there is a high propability the Absolute Maximum Rating are exceeded not just short term.
My recommendation:
if you want to make a good design and ensure the maximum lifetime of the device, you should never exceed 80% of the Absolute Maximum Rating.
/PeLuLe

[hwj-d]

Do you want your black output feed-through sitting directly above Q1?

No. I can also put that aside. Could this have a better impact under certain circumstances?

I had also thought, to put the transistor at his feet by 90 degrees through a dremeld recess in the housing to the outside. For that, we would also have to change the pcb layout, and rotate the transistor by 90 degrees to make that possible. Could this significantly improve the temperature control?

[hwj-d]

My recommendation:
if you want to make a good design and ensure the maximum lifetime of the device, you should never exceed 80% of the Absolute Maximum Rating.
/PeLuLe

Yes, for a good design, this is absolutely true. 

[kj7e]

Do you want your black output feed-through sitting directly above Q1?

No. I can also put that aside. Could this have a better impact under certain circumstances?

I had also thought, to put the transistor at his feet by 90 degrees through a dremeld recess in the housing to the outside. For that, we would also have to change the pcb layout, and rotate the transistor by 90 degrees to make that possible. Could this significantly improve the temperature control?

Any warmth rising from Q1 can cause a thermal imbalance between the + and - feed through junctions.  This will add some thermal EMF offset to the output.  The effects may be small, but it may be better to move the - feed through to the other side of the + feed through.

[hwj-d]

Any warmth rising from Q1 can cause a thermal imbalance between the + and - feed through junctions.  This will add some thermal EMF offset to the output.  The effects may be small, but it may be better to move the - feed through to the other side of the + feed through.

Ok, I will try to move the - feed. Thanks for the tip.

[vindoline]

Any warmth rising from Q1 can cause a thermal imbalance between the + and - feed through junctions.  This will add some thermal EMF offset to the output.  The effects may be small, but it may be better to move the - feed through to the other side of the + feed through.

Ok, I will try to move the - feed. Thanks for the tip.

It appears the + and - lines are a twisted pair before and after the feed through. Why not just bring the twisted pair through in one feed-through?

[hwj-d]

It appears the + and - lines are a twisted pair before and after the feed through. Why not just bring the twisted pair through in one feed-through?

Because the 'feed through' is a one pole isolated solder tag.
https://www.digikey.de/product-detail/de/mill-max-manufacturing-corp/380-10-001-00-000800/ED10992-ND/7066919

[try]

Hello hwj-d,

with the y-scale of your 34461A being labeled twice "10,000 00" I have no idea which range your readings span.
What are the numerical figures for max and min in there?

Regards
try

[hwj-d]

Hello hwj-d,

with the y-scale of your 34461A being labeled twice "10,000 00" I have no idea which range your readings span.
What are the numerical figures for max and min in there?

Regards
try

The middle line is exact 10V, the the lower one is 10V - 10uV, the upper is + 10uV.