If you happen to have a resin printer, you¡¯ll have one less step. Otherwise, just send it in to be professionally printed. It¡¯s such a tiny part that having 10-20 of them printed using high resolution industrial resin printers will still be affordable.
I¡¯d find leadless SMD parts to cobble this together from, install them in a split resin body, then use conductive epoxy to connect them internally, then close the body up if desired. Of course all this has to be done under a stereo microscope with various custom tools to manipulate the parts, dispense the epoxy, etc.
Printing the body is the least of the problems. You¡¯ll need bent capillaries to vacuum-hold the parts while moving them, and to dispense the epoxy. Maybe one can buy such already bent for the use in mind - 10 years ago when I played with it, I didn¡¯t see any. I¡¯ve had some success filling the thinnest of syringe needles with silver solder (drawing it in with vacuum), bending them in a custom wound steel spring (probably unnecessary but it helped with not nicking the outside), then blowing them free of solder with compressed air. All of that over a cobbled together temp-controlled hot air gun (feedforward PI instead of an on-off thermostat). Maybe overkill but I could bend those thick hairs (essentially) with 5-8mm bending radius while keeping their aspect ratio well within 90% of the original (according to the micrometer).
Another thing I had to do when working on similar sized assemblies (not for RF applications, but sizes were same) was to make custom files to expose more of whatever SMD part that I repurposed, so that the leads could be made smaller. I got some thin hard steel stock, ground it down on a tool grinder down to size, finishing with a 3000 grit wheel, put photoresist on, and etched the patterns. Worked well, but was a pain to get set up reproducibly. In the end I had to make a run of a 100 files, each a 1x2 mm cross section. Those things were sharp and cut the typical IC encapsulant plastics as if they were bitter.?
Cheers, Kuba
toggle quoted message
Show quoted text
7 sep. 2020 kl. 4:04 em skrev Mark Kahrs <mark.kahrs@...>:
?
Salut encore J-C: You might consider an intermediate step from beam-lead (I'm glad to see that Craig also thought of that).? Macom makes packaged diodes in small packages:
In any event, you'll have to fabricate the mechanical package as discussed.? I still wonder about using a 3-D printer.
Hello Craig,
???? Glad to find a guy who knows about 1817A sampler. This state
of the art equipment (for then time it was marketed) has fallen in
the museum class but works perfectly when the rather fragile
sampler diodes are in good condition. Your suggestion to use a
beam lead microwave schottky diode is one of the solutions I
envisaged to fix my 1817's. I also tried some very tiny microwave
SMD diodes (0.3 x 0.6 mm), which are supposed to be very fast (SMS
7621-060 from Skyworks).? If I have the binocular microscope, I
don't unfortunately possess the micro tweezers (clamps) necessary
to manipulate the diodes and put them at the right place on the
holder, and my attemps to getb them in place? and to fix them with
soldering paste have failed up to now. May be beam lead diodes
would be easier to handle !
??? Another solution I'll try will be to machine an new holder,
fitting in place of the original one, and using an old DO34 housed
Germanium point diode, which also presents very small
capacitance and offset voltage. It is slightly bigger than the
genuine one, but that seems workable. I recall you that the
assembly is composed of the diode, a series 2.2 K resistor, and a
2 pf capacitor towards 7mm line body. The new part would fit, but
electrical performances (rise time, overshoot) might be degraded !
??? Thank You anyway to help me renewing one of the most advanced
equipment ever marketed by former HP firm. (joined two drawings of
one of the sampling diode part - there are 2 opposed polarity
diodes ).
???? Jean-claude
?On 07/09/2020 09:19, Craig Sawyers
wrote:
I
was thinking in terms of a beam lead microwave diode ¨C but
held back from suggesting it given the difficulty in
handling, mounting and soldering/conductive epoxy/lead
bonding such physically tiny devices. But given the
bandwidth of the 1817A sampler, that would seem to be the
only easily procurable (and cheap!) part family. And they
have the correct sort of capacitance (0.2pF) and inductance
(<250pH) needed. Just don¡¯t sneeze when handling them ¨C
and a good stereo microscope is essential.
?
Craig
?
?
The 1817A sampler contains the
incredible Grove sampler (see the US Patent, US
3,278,763 for the associated patentese, also see the
Proc. IEEE paper by Grove for the theory of operation
and detailed analysis).? IMHO, the Grove sampler was the
last hurrah for non-monolithic microwave sampler
construction.? Since this sampler was found in many HP
instruments of the era, you can find repair instructions
in the 1430C manual (and others).? Basically, the diode
is a pill that is shimmed into contact with the
stripline.? To replace this with a modern diode, you're
going to have to fabricate a matching mechanical
contrivance.? I suggest (without having done this), that
you might be able to "print" a suitable diode holder and
then insert a replacement diode (it would be tempting to
use a beam-lead diode or find a suitable pill diode).
For extra homework, I suggest
replicating Grove's analysis in the Proceedings.
?
Hello,
???? I'm trying to repair several old 1817A sampling
heads. They all show the same failure : the sampling
diodes are fused. Since they are fitted in a very tiny
structure, inserted into the wave coaxial guide, it is
rather tricky to attempt replacing the diode by a modern
SMD Ga As shottky device. The original are obviously
point germanium diodes, though not written on the
manual. Has anybody an idea where to get such sampling
assemplies (which includes a resistor, a 2 pf capa and
the diode in a very specific assembly), or an idea to
fit a modern component in place of the old one ?
??? Thanks for replay !
??????? bernjc
