Archive for September, 2017
Mary confronted this critter in the garden, whereupon it fled into the compost bin:
She barricaded it with spare tomato cages across the bin’s entrance, I wedged an aluminum sheet behind the cages, and we got the stinkeye for our efforts:
I deployed the hose, watered it for a few minutes, and we left it to consider its options. Groundhogs are pretty much waterproof, but we hoped the wetdown would be sufficiently unpleasant to mark the garden as “Here be dragons” in its mental map.
After an hour, it had vanished. We know from past experience that groundhogs can climb up-and-over the chain link fence surrounding the compost bin (it was a dog pen for the previous owners), although it knocked down the aluminum sheet and may have exited through the garden.
It looks well-fed and ready for winter.
Searching for groundhog will reveal previous encounters with its ancestors & relatives.
The LT1920 instrumentation amplifier now sports two silver-mica caps on its inputs as a differential-mode input filter cutting back strong RF signals (clicky for more dots):
In principle, a DM filter should eliminate RF rectification from out-of-band signals, although I think the attic is quiet enough to not need any help. The caps form a simple RC LP filter rolling off at 5.490 kΩ × 150 pF → 193 kHz, high enough above the 60 kHz signal to not make much difference down there.
The silver-mica caps come from the Big Box o’ Caps, which contained an envelope with a few large 150 pF ±1% caps and a bag stuffed with similar 147 pF ±1% caps. Mixed in with the latter were some smaller 147 pF caps (*) of no particular tolerance (perhaps 5%), from which I neurotically matched a pair to 0.05 pF without too much effort. Doesn’t matter, given the other tolerances and suchlike, but it was amusing.
I’d inadvertently grounded the cold end of the 330 Ω input resistor in the LM353 bandpass filter, now properly tied at the Vcc/2 virtual ground to take the DC load off the LT1920 output: a 100 nF cap (27 Ω at 60 kHz) stores the bias level without messing up the filter shape.
A similar cap rebiases the protected resonator at the LT1010 buffer input:
The new caps aren’t all that visible and the resonator vanishes in the clutter:
Next: find out how well it works!
(*) Yes, there were two envelopes between 150 pF and 147 pF:
I eventually noticed the yellow LED indicating +24 V input from the power supply (previously, a noisy wall wart) was dark. Poking around revealed I’d inadvertently installed a 1 kΩ ballast resistor:
A 1/4 W resistor can’t dissipate half a watt for very long, as shown by the discolored circuit board around the leads and the faint smell of electrical death in the area.
I swapped in a 3.3 kΩ resistor, the yellow LED lit up for a few seconds, then went dark again. This time, the LED was dead; apparently, it’d been overstressed for long enough to fail. I can’t be too annoyed.
Unfortunately, replacing the LED required removing the entire housing with all three LEDs, chopping off the defunct block, reinstalling the attenuated block with the two green LEDs, installing a similar red LED, and finally installing a nice 3.3 kΩ half-watt resistor:
So it goes …
The FG085 function generator shows 60000 Hz and the AD9850 shows 60001.58 Hz, but they’re running at exactly the same frequency:
I trust the AD9850 readout, because I just finished zero-beating it against the GPS-locked 10 MHz frequency reference: it’s dead on. The scope’s frequency measurement is clearly out of its depth at this resolution.
The “user interface” doesn’t amount to much. The DDS starts at 60.000 kHz, as defined by a program constant. Push the joystick left-right to step by 0.1 Hz (actually, multiples of 0.0291 Hz, so either 0.087 or 0.116 Hz, whichever makes the answer come out closer to the next multiple of 0.1 Hz). Push it up-down to step by 1.0 Hz (insert similar handwaving here). Push the button inward to reset to 60.000 kHz.
The OLED displays the frequency (in big digits), the output of the log amplifier (which isn’t hooked up here) in dB (over 4 μV), the DDS clock oscillator temperature, and a few lines of static prompting. The camera shutter blanked the last line, which should read “Button = reset”.
There’s no amplitude adjustment, other than the DDS current-control twiddlepot and the buffer amp’s gain-setting jumpers, but I (think I can) gimmick up an adequate inductive kicker for the fake preamp antenna circuit.
The Arduino source code as a GitHub Gist:
This took entirely too long to figure out:
That’s with the scope probe ground clip connected to the wall wart coax connector barrel and the scope probe tip on the ground clip. It’s not the noise on the 24 VDC supply, it’s the noise injected into the ground connection!
Huh. Makes it tough to sort out low-level signals, it does indeed.
Consider one of my bench power supplies at 24 V:
Nice & quiet, the way power should be. One might quibble about the residual noise, but at least it’s not blasting out horrific bursts at 120 Hz.
For completeness, the PCB inside the offending SMAKN 24 V wall wart:
“High Quality Commercial Grade” my aching eyeballs.
[Update: Edits based on eagle-eyed observations in the comments. ]
Not as many missing components as I expected, though, if the truth be told. The missing
transformer common-mode choke seems odd and, AFAICT, the resistor inductor angling out from the R1 callout doesn’t connect to anything, connects directly to the AC line because C5 is missing and the pad joining them doesn’t go anywhere else it replaces the jumper (?) to the bottom-left pad and the missing parts. The red LED in the upper right isn’t visible through the black case, although it might serve as a voltage regulator.
Over on the far right, beyond the transformer and between the two capacitor cans, is a component marked C9 with an oddly angled part. Seen from the other end, it’s a ferrite bead:
I don’t know why that spot has an inductor symbol with a capacitor part callout.
The other side of the PCB looks clean:
It’ll probably serve well in a noise-tolerant application, maybe an LED power supply.
FWIW, the UL mark seems conspicuous by its absence:
Not sure what I’ll replace it with, although a small 24 V power supply brick may suffice.
Limiting the resonator drive to about 1 μW in the face of wildly varying RF from the antenna (or the occasional finger fumble) requires brute force. A nose-to-tail pair of Schottky diodes seems to do the trick:
The 100 Ω resistor blunts the drive from the LM353 op amp (implementing a bandpass filter) when the signal peaks exceed 200-ish mV in either direction from the Vcc/2 bias stored in the 10 μF cap.
The 11.5 kΩ resistor downstream of the resonator isolates it from the Vcc/2 bias, with the 100 nF cap sinkholing the signal and the 4.7 kΩ resistor preventing feedback into the bias supply. The cap looks like 26 Ω at 60 kHz, so the feedback runs -52 dB from the output and the bias supply knocks it down a bit more. The preceding amps apply 40-ish dB of gain from the antenna terminals, so the loop gain looks OK.
It’s another few components on the board:
The blue twiddlecap should allow pulling the tuning fork’s series resonance upward to exactly 60 kHz.
Applying way too much signal to the antenna terminals in order to get 1 Vpp from the LM353 shows the limiter in action:
The resonator sees no more than 200 mV in either direction from the bias level, so it’s all good.
On the low end, the diodes have no effect:
Pay no attention to all that noise.
My first thought was to put the diodes across the resonator, a Bad Idea: straight up, doesn’t work. The 1N5819 datasheet shows they have about 300 pF of junction capacitance at zero bias and a pair of ’em will swamp the resonator’s internal 0.8 pF parallel capacitance and punch it out of the circuit.
An array of tiny eggs appeared on the outside of our bedroom window:
The patch measures 12 mm across and 14 mm tall. From across the room, it looks like a smudge, but it consists of hundreds of eggs, each on a tiny stalk glued to the glass:
The bottom image is two days later than the top one, both are scaled to about the same size and contrast. The critters look about the same, although I think the lines have more prominent ripples or bumps.
We have no idea what they’ll turn into, but they certainly look like they have two eyes and wings …