Re: like butta

john larkin <JL@gct.com> wrote:

On Thu, 23 Jan 2025 03:43:03 -0000 (UTC), Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:
 

Simon and I are working on a TDR based soil moisture sensor for
agriculture.  Like many such things, it uses two parallel tines made of
18/8 stainless, that form a balanced transmission line.
 
Ours has a slide hammer for pounding it into really difficult soil, e.g.
hardpan. The measured shock from that is around 1.6E5 m/s**2, i.e. 16000
gees, and over its lifetime it might see around 1E6 blows. Challenging.
 
Doing <200 ps TDR on a balanced line obviously needs a very wideband balun.
We’re using a plain ferrite design based on Ferroxcube 61 sleeves on 1.25mm
coax. The resulting 50-? differential mode goes into a machined tapered
structure that maintains 50 ohms while spreading out the mode to match the
25-mm tine spacing.

 
We've got excellent results with micro-coax on pot cores.
 
https://www.dropbox.com/scl/fi/o2r1m53es9rb3e7nbsnsr/Pot_Core_TXline.JPG?rlkey=t1g3rp0erz72tqchb99fupfu7&raw=1
 
https://www.dropbox.com/scl/fi/ie7gzmdwuw4gqgy1pxpjl/TX_1.jpg?rlkey=xllwjn2cg0a0t3yjh90om9ap1&raw=1

I have that one in my tricks file, for sure.

We’re using two ferrite sleeves as a 1:1 balun, which is the sideways
version.

 

 
For survivability, the whole thing is being potted in very hard epoxy with
a dielectric constant of 3.5ish. This means that it’s hard to iterate—we
get one try per apparatus.
 
Sooooo, I tried making a soft material with that epsilon, which turns out
to be nontrivial. My initial thought was to use alumina lapping powder
(9.3) for filler and vaseline (2.0) for the matrix, but it turned into
clumps by the time it got to 3.0.
 
Casting about for alternatives, I remembered the butter in the lab fridge,
so I measured that: 3.8, better than good enough.
 
We were able to verify the design without doing anything irrevocable, and
there was some left over for the bagels.
 
Fun.
 
Cheers
 
Phil Hobbs

 
Are you doing a reverse convolution to beautify the step waveform? You
can tolerate a really ugly TDR if you can make a FIR filter to pound
it flat.

The step is actually very nice-looking—clean edges, flat pulse tops, no
worries.  We did have to make a couple of layout adjustments to get there,
but the result shows TDR edges surprisingly similar to the SD-24’s.

Ours is much less fancy, and needs a few pulses per delay value to get good
convergence in the sampling loop, but in a not-too-scientific comparison on
a 24-inch RG-188 cable, both the SD-24 and our gizmo showed 60-ps TDR edges
(10-90%).  Of course the SD-24 is more like 30 ps on a less lossy cable,
but for $16 per completed board, including MCU and data converters, I’m
pretty happy with it.

Cheers

Phil Hobbs

 
 

--
Dr Philip C D Hobbs  Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics  Optics,
Electro-optics, Photonics, Analog Electronics