diy solar

diy solar

$50 Oscilloscope Game

svetz

Works in theory! Practice? That's something else
Joined
Sep 20, 2019
Messages
7,201
Location
Key Largo
No mad scientist's treasure horde is complete without an oscilloscope! Thanks to @BiduleOhm posting info on the video below we now know how to get one:

To "win" the game, you either need to buy or report on ≥ dual channel 20 MHz Analog Oscilloscope that is < $50.
What do you win? Well, the bargain oscilloscope you just found!

The problem is, a lot of folks on eBay have watched the video... so just finding one less than $50 isn't hard, but usually they'll tack +$$$$ onto the shipping.

I learned a lot about shopping on eBay from the video... here are my take-aways:
  • Don't use the filters, professional sellers use those and are looking to make money, not dump "junk" they don't know how to use.
  • You're looking for ones that have typos and are mis-categorized.
  • Names to look for are Hitachi v212, kenwood, LS8050, kikusui, panasonic, Phillips, gw instek
  • Names like Tektronix are there, but harder to find bargains with
  • Don't buy anything from before the '70s
  • Don't buy one unless you see a trace on the screen, no trace = probably doesn't work
  • If the price is $70-100, and they don't have a lot of sales (e.g., not a professional), offer them $40.
  • Don't touch Digital storage
Good Luck & Happy hunting!
 
The best deal out these is one of those USB scopes like the Hantek 6022B. They are about $65 shipped and it comes with two probes. It will measure directly volts, time and frequency. Any image can be captured as a jpg and shared with others. Old scopes have a myriad of problems generally related to the switches and pots.

If you choose a USB scope be aware it the extreme top or bottom is perfectly flat without any noise, it is possibly overloaded. I wish they would turn that flat section red to indicate an overload. Measuring anything over 50V and you should buy a fixed 100X scope probe. It is easy for X10 switch to slip to X1 on the standard probe. It is easy to blow out the input if that happens on high voltage.
 
I'm playing more at the other end of the food chain.

I started out with a Tek analog scope, $200 for 250/350 MHz Tek 485 with four probes (left top in my lab photo)

1595954815184.png

Later I bought a 784D for $2000, so I could automate testing. That started having errors and I bought another for $1600.
Hard drive, 4 GS/s, 1 GHz, 8 Meg memory depth, ITAR controlled. That was a $25,000 scope new.
Also 6000V differential probe.
8 bit ADC but over-sampling gives about 12 ENOB. I used it to linearize my 750 Vpp 11 MHz amplifier (for a mass spectrometer) well beyond 12 bits by averaging multiple readings in the PC. Had to toggle between amplitude settings so time-averaged power in the probe remained constant, didn't heat it up during an 8-hour calibration.

Note that the kilovolt probes for scopes and DMM are not for utility grid with unlimited current. Just for things like CRT supplies.

1595954437759.png 1595954686530.png

I don't buy these things when they're new (new old stock if I'm lucky), but my employers are sometimes on the bleeding edge:


As they say, "If you have to ask, you can't afford it"
But I asked, anyway:



1595954332751.png
 
Bandwidth of an oscilloscope is something like the -3db or half-power point, as higher input frequencies are attenuated more.
These days, sample rates can extend well above that point of the analog input.
What Keysight has done is to run FFT on the trace, restore amplitude of each frequency component according to the attenuation saved in calibration, then reconstruct the time-domain waveform.
That lets them show time-domain waveforms flat out to 110 GHz even though analog bandwidth is lower.
We can do that too, although only up to whatever our sample rate allows.
Maybe I could squeeze almost 2 GHz from my 4 GS/s. But data transfer due to software on my PC is far slower than the 1 MB/s of GPIB.
Maybe if I learned how to write programs that execute locally on the scope.

Performance can be stretched in a different direction as well.
When 8 bit ADC samples many times between two different values (e.g. slow signal falls from binary 64 to binary 63) you can get more bits with a moving average or other filter. The ADC reading toggles back and forth due to noise, and when above 63.5 it returns 64 more often than 63.
When I sample my 11 MHz sine wave at 4 GHz, this gives me several more bits than just 8 bits. That feature is built into the scope, but I also averaged about 100 readings to improve linearity. I want steadily increasing analog amplitude to deliver monotonically increasing 14 bit values, not bouncing up and down due to noise.

That was DIY rather than having my employer get a different product with 14 bit ADC for considerably higher price.
They complained about the 4 to 8 hour cycle time, but I pointed out that the annual production rate it supported was significant.
Several ways to improve throughput, anyway.

I've been able to use the scope to capture periodic data communication between my SMA inverters and monitoring equipment. Sunny Boy Control reads from Sunny Boy about once a second. What I have never captured is the message from Sunny Island that tells Sunny Boy "on grid" or "off grid". I see Sunny Boy's display show which state its is in, but was never able to trigger on the signal. Digital scope might only be ready to trigger part of the time and miss a single event (especially if also triggering on other events). Analog scope might capture that, but look fast before the phosphor fades. Digital can trigger once and hold the waveform for you.
 
Of course if you can afford a digital one it's nicer but a 2x 20 MHz analogue one should be good enough as a first scope. If you need more then you already know it and you'll buy something nicer ;)
 
I've got an old Goldstar OS-9100 D 100MHz scope that works perfectly fine sitting here. If anyone in Finland reads this and wants it - PM me and you can come and pick it up. It'll be too much trouble to ship it...
 
A decent USB-based oscilloscope is the Owon VDS1022I. It is currently ~USD100 on amazon (but I have seen it as low as $80). It is dual channel 25 MHz only but that covers pretty much what any hobbyist may probably need. An important feature (signaled by the "I") is that the analog part is electrically insulated (or at least that is the claim) from the USB data+power connection. The software is somewhat clunky and windows only (but there are workarounds to make it work in Macs and Linux since some of the software version are really in Java).
 
Here's a product we've started to use at work, Saleae Logic Analyzer:


8 or 16 input, $400 to $1000, up to 500 MHz sampling, analog and digital inputs.
It can digitize up to 125 Hz digital signals so long as high and low time are something greater than 25% of period (at least one sample reliably in the middle of the high and low levels.)
Analog up to 5 MHz, 50 Msample/second, 12 bit, +/-10V
So there's a slow USB-based mixed signal scope with up to 16 channels.
Comes with serial port parsing for RS232, RS485, I2C, SPI. People also write their own for other bus standards. Digital is 1.2 to 5.0V so you might need to attenuate for full RS-232 levels.

It comes with a GUI, and Python scripts are available. I wrote Matlab to kick Python and return a capture file.

USB isolation is an issue. A ground loop between PC and DUT can blow something up. Best to have laptop floating, or get an isolated USB. I've seen damage in previous jobs using other equipment because a switching power supply for DUT made an 8V transient when first plugged in (capacitive coupling of its isolation transformer.) My solution was a quality bench supply instead.
 
If I get one as a hand-me-down, does that count? It has a layer of dust on it that might be permanent. I would have no idea what to do with it. But it would be cool!
 
s-l225.webp
Almost less than $50 ... $51.17 (price includes shipping)
 
Back
Top