AI4QJ wrote:
"Bryan" wrote in message
...

Suzy wrote:

Looking for a website that gives comparative performance of diodes (max
frequency for UHF use)

A recollection from decades back are the 1N21, 1N23, 1N32, 1N53, 1N60, and
1N78 point-contact silicon diodes that were used as mixers well into the
microwave spectrum. Googling 1N23 produces 8000+ hits... including a
"Simple 2.4 GHz SWR Meter": (http://www.qsl.net/n9zia/24swr/index.html). I
see RF Parts (http://www.rfparts.com/) sells the 1N21 and 1N23.
MicroMetrics (http://www.micrometrics.com/) makes them, as well as various
schottky diodes suitable for sampling.
73,
Bryan WA7PRC


Interesting but this is not what was requested.




I don't think you'll find a website that compares diodes for RF
performance.

A couple reasons spring to mind:
1) Most RF designers have their favorite parts that they've used before,
so they will tend to just use them again. If their favorite part
doesn't work, then they'll start hunting through the pile of datasheets
and books or ask around for suggestions (e.g. just like on this list)

2) The performance is highly application specific. You can look at the
basic numbers (PIV, forward resistance, capacitance, etc.) to get an
idea, but it's really going to come down to how it works in YOUR
circuit, and that's usually a breadboarding or modeling exercise (most
parts have pretty good SPICE or other models available). Just because
manufacturer X's 1n5711 works great in my mixer doesn't mean it will
work great in your frequency tripler or someone else's RF switch.

3) There are significant differences in performance between
manufacturers for the same JEDEC partnumber (e.g. 1n5711). In a lot of
high performance applications, you actually depend on some non-datasheet
performance property. For instance, there's a small company in England
that makes a low noise JFET that is unique in the world and prized by
folks building charge/voltage converters, even though it's got a
standard 2N number. And then, they still get a batch of them and hand
select.

4) Parts selection is something of competitive value. Someone who
designs RF stuff for a living (i.e. someone who would be likely to know
all the trade offs between various part types and mfrs and could
generate a comparative website) is going to be encouraged by their
employer to not publish this for the whole wide world.

On Feb 5, 9:29 am, Jim Lux wrote:
AI4QJ wrote:
"Bryan" wrote in message
m...

Suzy wrote:

Looking for a website that gives comparative performance of diodes (max
frequency for UHF use)

A recollection from decades back are the 1N21, 1N23, 1N32, 1N53, 1N60, and
1N78 point-contact silicon diodes that were used as mixers well into the
microwave spectrum. Googling 1N23 produces 8000+ hits... including a
"Simple 2.4 GHz SWR Meter": (http://www.qsl.net/n9zia/24swr/index.html). I
see RF Parts (http://www.rfparts.com/) sells the 1N21 and 1N23.
MicroMetrics (http://www.micrometrics.com/) makes them, as well as various
schottky diodes suitable for sampling.
73,
Bryan WA7PRC

Interesting but this is not what was requested.

I don't think you'll find a website that compares diodes for RF
performance.

A couple reasons spring to mind:
1) Most RF designers have their favorite parts that they've used before,
so they will tend to just use them again. If their favorite part
doesn't work, then they'll start hunting through the pile of datasheets
and books or ask around for suggestions (e.g. just like on this list)

2) The performance is highly application specific. You can look at the
basic numbers (PIV, forward resistance, capacitance, etc.) to get an
idea, but it's really going to come down to how it works in YOUR
circuit, and that's usually a breadboarding or modeling exercise (most
parts have pretty good SPICE or other models available). Just because
manufacturer X's 1n5711 works great in my mixer doesn't mean it will
work great in your frequency tripler or someone else's RF switch.

3) There are significant differences in performance between
manufacturers for the same JEDEC partnumber (e.g. 1n5711). In a lot of
high performance applications, you actually depend on some non-datasheet
performance property. For instance, there's a small company in England
that makes a low noise JFET that is unique in the world and prized by
folks building charge/voltage converters, even though it's got a
standard 2N number. And then, they still get a batch of them and hand
select.

4) Parts selection is something of competitive value. Someone who
designs RF stuff for a living (i.e. someone who would be likely to know
all the trade offs between various part types and mfrs and could
generate a comparative website) is going to be encouraged by their
employer to not publish this for the whole wide world.



And of course there's the other side of the coin on number 4: a
manufacturer isn't going to provide a comparison with other
manufacturers' parts for at least a couple reasons. They want to
focus attention on their own parts, and they have no control over any
changes another manufacturer may make in their parts and thus no way
to insure accuracy of such a comparison.

The tough part of Suzy's quest is finding an acceptable diode in non-
surface-mount, it seems to me. I did find that a common 1N4148 non-
Schottky performed in my test detector circuit at 450MHz better than I
expected, and better than some other diodes I tried like 1N34 and
several other germaniums, and an axial-leaded Schottky or two in the
high voltage, high capacitance signal diode class that I believe Wim
warned against.

I really appreciated Wim's posting on this topic, and also yours,
Jim. Thanks.

Cheers,
Tom

K7ITM wrote in
:

....
The tough part of Suzy's quest is finding an acceptable diode in non-
surface-mount, it seems to me. I did find that a common 1N4148 non-
Schottky performed in my test detector circuit at 450MHz better than I
expected, and better than some other diodes I tried like 1N34 and
several other germaniums, and an axial-leaded Schottky or two in the
high voltage, high capacitance signal diode class that I believe Wim
warned against.
Presuming this is for the 70cm reflectometer, I have already pointed
Suzy to two Australian retailers with diodes that should work.

My assumption is that she will design the detector to work above the
square law region, and the usual practice is then to design the coupler
so that the dc voltage is much greater than the diode voltage drop, and
the scale shape is relatively independent of the diode drop.

Another retailer is minikits.com.au. Mark has 1N5711 for 0.66, and it
will probably cost $7 for postage. Last time I tried to buy diodes from
Mark, he wanted to know what I wanted them for, whether it was a worthy
project, and would deem to sell me only a small quantity. I subsequenctly
bought a couple of hundred 1N5711 on Ebay for $15.

Ebay is a huge source of schottky diodes (if you must prefer schottky).
Just search for schottky, sbd, barrier, hot carrier etc, then get the
specs for each one and see if they are suitable. Thing is that most the
signal ones will be SM packages of one or two diodes, but they can be
used with pigtails attached.

There are 9 listings on Ebay right now for 1N5711.

Owen

K7ITM wrote:
On Feb 5, 9:29 am, Jim Lux wrote:

AI4QJ wrote:

"Bryan" wrote in message
news:GK2dnawmofw9DjvanZ2dnUVZ_tqtnZ2d@giganews. com...

Suzy wrote:

Looking for a website that gives comparative performance of diodes (max
frequency for UHF use)

A recollection from decades back are the 1N21, 1N23, 1N32, 1N53, 1N60, and
1N78 point-contact silicon diodes that were used as mixers well into the
microwave spectrum. Googling 1N23 produces 8000+ hits... including a
"Simple 2.4 GHz SWR Meter": (http://www.qsl.net/n9zia/24swr/index.html). I
see RF Parts (http://www.rfparts.com/) sells the 1N21 and 1N23.
MicroMetrics (http://www.micrometrics.com/) makes them, as well as various
schottky diodes suitable for sampling.
73,
Bryan WA7PRC

Interesting but this is not what was requested.

I don't think you'll find a website that compares diodes for RF
performance.

A couple reasons spring to mind:
1) Most RF designers have their favorite parts that they've used before,
so they will tend to just use them again. If their favorite part
doesn't work, then they'll start hunting through the pile of datasheets
and books or ask around for suggestions (e.g. just like on this list)

2) The performance is highly application specific. You can look at the
basic numbers (PIV, forward resistance, capacitance, etc.) to get an
idea, but it's really going to come down to how it works in YOUR
circuit, and that's usually a breadboarding or modeling exercise (most
parts have pretty good SPICE or other models available). Just because
manufacturer X's 1n5711 works great in my mixer doesn't mean it will
work great in your frequency tripler or someone else's RF switch.

3) There are significant differences in performance between
manufacturers for the same JEDEC partnumber (e.g. 1n5711). In a lot of
high performance applications, you actually depend on some non-datasheet
performance property. For instance, there's a small company in England
that makes a low noise JFET that is unique in the world and prized by
folks building charge/voltage converters, even though it's got a
standard 2N number. And then, they still get a batch of them and hand
select.

4) Parts selection is something of competitive value. Someone who
designs RF stuff for a living (i.e. someone who would be likely to know
all the trade offs between various part types and mfrs and could
generate a comparative website) is going to be encouraged by their
employer to not publish this for the whole wide world.




And of course there's the other side of the coin on number 4: a
manufacturer isn't going to provide a comparison with other
manufacturers' parts for at least a couple reasons. They want to
focus attention on their own parts, and they have no control over any
changes another manufacturer may make in their parts and thus no way
to insure accuracy of such a comparison.

The tough part of Suzy's quest is finding an acceptable diode in non-
surface-mount, it seems to me. I did find that a common 1N4148 non-
Schottky performed in my test detector circuit at 450MHz better than I
expected, and better than some other diodes I tried like 1N34 and
several other germaniums, and an axial-leaded Schottky or two in the
high voltage, high capacitance signal diode class that I believe Wim
warned against.

I've used a fair number of 1n914 and 1n4148 diodes for detectors over
the years. It all depends on what you want to get out of it.

The 1N2x and 1N34 diodes work well in their special mounts (like as a
waveguide microwave detector), but I think that's a packaging and
connector capacitance issue, mostly.