Another TESLA ref diodes TKZD13/D

[iMo]

I published a post with the teardown of the KZZ82 - the ancient TESLA TC compensated diodes - and this is a picture of the "newer" diodes from their 4.5-6.5 digits meters, produced in 80ties (.. most precisely the diodes were used in 4.5-6.5 digits meters produced by the company "Metra Blansko" in former Czechoslovakia, btw. - still existing with 110 years long tradition).
Below the picture - 6 digits on the sticker, but none indication at which current (the diodes potted in a small plastic box).
I cannot find a DS for it, any info is welcomed..

PS: a nice 10y old thread on the Metra 4.5-6.5digits meters from 80ties - https://www.eevblog.com/forum/reviews/czechoslovakian-measure-equipment-collection/

[jaromir]

AFAIK TKZD references were made in two variants - TKZD13 and TKZD25, numerical value indication the nominal operating current. I believe those were manufactured up to early 90's, nowadays I would consider it difficult to source historical rarity.

I obtained two specimens of TKZD25 reference diode and ran a short temperature sweep test with them. Diodes were placed in DIY temperature oven, swept by 8°C, connected in series and powered by DIY SMU set to 25mA. Voltage on the diodes was switched by Prema 2040 switch and sensed by Keithley 2010 DMM.

Both units do have reference voltage around 8,45V. As apparent on the graph, one reference looks quite OK, the second one is not that stellar, with regards to both tempco and drift. Longer test would be needed to settle it down, I assume the references weren't powered for decades. Still not as unconvicing as KZZ8x references - https://www.eevblog.com/forum/metrology/kzz82-voltage-reference/msg4253485/#msg4253485

[iMo]

@jaromir: I made a quick measurement with the black one above, without playing with themerature, just observing the knee and I wrote 12.5mA there on it (in the red), without knowing the "13" means the TC=0 mA current
Btw. aprox 0.75ppm/C with one of your samples is not bad, indeed..

[saturnin]

Some time ago, I disassembled a faulty TKZD13/D (it was too noisy if i recall correctly). Inside, I found three Zener diodes (likely 2x KZ260/6V8 and 1x KZ260/7V5 - its real Zener voltage was 7.1V, so you got approx. 8.5V in total: 2x0.7V+7.1V) and one wire wound resistor (see attached schematic for the internal connection). The diodes were sealed by some kind of epoxy resin. The resistor values varied between samples (had four of them). I measured following values: 1.48 kOhm, 2.10 kOhm, 4.22 kOhm, and the last one was open (> 20 MOhm).

I was surprised they used KZ260 which is a general purpose Zener diode, nothing fancy (see attached datasheet). I suppose they had to age references and select the best ones with low long-term drift.

[iMo]

Thanks!
A pretty expensive construction, similar to the KZZ82..
Btw. - isn't the resistor wired to the 2 upper forward biased diodes only?? (like in the KZZ82)..

[Gyro]

Is that a finned (TO220?) heatsink that they've used as an isothermal block? The threaded hole doesn't seem to line up with the mounting hole - direct thermal connection between the two would probably be a bad idea anyway.

[DavidKo]

Based on this article I have understood that the parallel resistor was used to change the current in compensation diode(s) to tune the temperature coefficient.