Oh No! I got infection.

[zhtoor]

hello,

1N823A is not a burried zener, it is simply a glass sealed/passivated temperature compensated zener.

regards.

-zia

[Vtile]

That's nice. I usually grind out isolated pads with a burr, but the small pieces of circuit board are way better.

Heh, I would have done so also, but I still do not have the tool for making those islands. ...And I'm too cheap to buy one. 

I wonder if there is any benefit for these pads as they create a small capacitor to every circuit node. On the other hand I do not have equipment to measure the noise anyway.

[GeorgeOfTheJungle]

..But manhattan style ref. board looks nice. 

 Olé!

[Vtile]

hello,

1N823A is not a burried zener, it is simply a glass sealed/passivated temperature compensated zener.

regards.

-zia

Hmm... I had strong impression that the 1n82x series were using subsurface technology (like every zeroTC reference zener). I'm probably in wrong impression. I'm not using A-series and only 1n823, because they were cheap and in stock at local electronics supplier. 

[David Hess]

I wonder if there is any benefit for these pads as they create a small capacitor to every circuit node. On the other hand I do not have equipment to measure the noise anyway.

If the capacitance serves to make the prototype a more accurate representation of the production model then it does.  I have seen the same construction technique used with strips cut to a width to make 50 ohm microstrip transmission lines for RF work.  If low capacitance and leakage is required, then just air wire those nodes.

How did you connect the pads to the copper clad board?  Glue?  Solder?

I was thinking of laying out tiny boards with various surface mount footprints with pads only on the top and maybe pads for common things like local decoupling.  Corner or edge plated through hole pads could be soldered from the top or side for attachment to a copper clad board below.

[Vtile]

I wonder if there is any benefit for these pads as they create a small capacitor to every circuit node. On the other hand I do not have equipment to measure the noise anyway.

If the capacitance serves to make the prototype a more accurate representation of the production model then it does.  I have seen the same construction technique used with strips cut to a width to make 50 ohm microstrip transmission lines for RF work.  If low capacitance and leakage is required, then just air wire those nodes.

How did you connect the pads to the copper clad board?  Glue?  Solder?

I was thinking of laying out tiny boards with various surface mount footprints with pads only on the top and maybe pads for common things like local decoupling.  Corner or edge plated through hole pads could be soldered from the top or side for attachment to a copper clad board below.

I did use glue as the copper is 1-sided, old stock ... cheap ... did I mention already that I'm cheap.

I just used a bandsaw to cut one board to strips and then with side cutters I do cut them to the right length.

The coupling to the ground plane (the intact copper surface) from the strips might reduce some noise, but again it might create oscillations as it just "happen to be there". I do not have equipment to measure down to the noise floor so for me the noise is pretty irrelevant atleast for now... Like I said I need better meter now, since I did hit to my measurement limit with this board (The wonders of bridge measurements.).       

The reference is in star-ground btw. While the supply filter capacitors aren't.

[branadic]

Did someone had a look on 1N4569 (0.5mA), 1N4574A (1mA), 1N4579A (2mA) or 1N4584A (4mA)? They are temperature compensated zeners with 5ppm/K. Sure expensive, but available at Mouser and/or Digikey and fabricated by Microsemi.
I wonder how their noise performance looks like, since there is no specification given in the datasheet.

-branadic-

[Vtile]

Did someone had a look on 1N4569 (0.5mA), 1N4574A (1mA), 1N4579A (2mA) or 1N4584A (4mA)? They are temperature compensated zeners with 5ppm/K. Sure expensive, but available at Mouser, ... and fabricated by Microsemi.
I wonder how their noise performance looks like, since there is no specification given in the datasheet.

-branadic-

I think the Conrad Hoffmans original topic of this traditional zener reference circuit did use 1N4569...


Edit. No... it were 1n4568A : https://www.eevblog.com/forum/metrology/old-fashioned-zener-10v-reference/

[zhtoor]

njsemi makes temperature compensated zeners with tested noise specs.
check out the following with a guaranteed 0.2 uv/rtHz noise.

http://www.njsemi.com/parts/1N4908A

regards.

-zia

[Vtile]

njsemi makes temperature compensated zeners with tested noise specs.
check out the following with a guaranteed 0.2 uv/rtHz noise.

http://www.njsemi.com/parts/1N4908A

regards.

-zia

Is that actually two junctions in series as the 12.8Vz is pretty close of 2 * typical ref. Zener optimum.

[zhtoor]

the low noise suggests that it is two zeners in series with another couple of diodes in compensation mode,
since avalanche breakdown (> ca 5.6v breakdown) has higher noise.

regards.

-zia

[cellularmitosis]

Vtile that circuit is more an architectonic masterpiece.

Thank you.

It is the byproduct of the enormous amount of resistor I had to use to get the obscure ZeroTC current value for the buried temperature compensated zener and then the gain set for the OP.

..But manhattan style ref. board looks nice. 

I’m a fan!

https://www.eevblog.com/forum/beginners/pcb-cutting-health-concern/msg369099/#msg369099

[zhtoor]

That's nice. I usually grind out isolated pads with a burr, but the small pieces of circuit board are way better.

Heh, I would have done so also, but I still do not have the tool for making those islands. ...And I'm too cheap to buy one. 

I wonder if there is any benefit for these pads as they create a small capacitor to every circuit node. On the other hand I do not have equipment to measure the noise anyway.

speaking of "cheap", the POOR MAN's island cutter:-

a steel nail, some filing, some bamboo and epoxy later (=USD 0.1) :-

regards.

-zia

[tggzzz]

That's nice. I usually grind out isolated pads with a burr, but the small pieces of circuit board are way better.

Careful dremelling can allow SMD devices with >=3 pins to be directly accomodated.

The glued-on PCB tiles can be used with 2-pin SMD devices, e.g. decoupling capacitors tilted "precariously" at 45degrees to the plane.

[Vtile]

Nice tools and tricks!

One project that did jump to my mind is to make one of these devices, but with glimmer light and ADA4700 operational amplifier. Might be interesting contraption. 

[Vtile]

That's nice. I usually grind out isolated pads with a burr, but the small pieces of circuit board are way better.

Careful dremelling can allow SMD devices with >=3 pins to be directly accomodated.

The glued-on PCB tiles can be used with 2-pin SMD devices, e.g. decoupling capacitors tilted "precariously" at 45degrees to the plane.

I don't like the epoxy&fiber dust what Dremel method makes. I wonder how those metal engraver tools (DIY or bought) would work in this job (goodle image search: metal round graver tool). If that tool would work then in the right hands the electronics art would get a whole new meaning. 

[Vtile]

Well three months and the manhattan project reference still shows exactly 10.0000 V. It takes about an hour to reach that and then it stays on it, when I turn on both the meter and the reference.  Those three months the reference have been powered like 2 weeks total.

After the first week I needed to adjust the tension by about 100 uV up. After that it have staid on 10.0000 at times blinking 9.9999 V

I'm pretty pleased.  ... and the manhattan reference is still nice to look at.