Homemade oscilloscope to detect PWM DIY guide
For Android there's an app called "HScope". It supports various Hantek devices, including the 6022BE. My first impression is that it's great software. It detects frequencies much better than OpenHantek does and generally seems very polished. We might be able to create a portable solution with an Android device, a USB On-The-Go cable, a power bank, 6022BE + probe, and the HScope app.
HScope App: https://play.google.com/store/apps/details?id=com.martinloren.hscope
Guide: http://hscope.martinloren.com
- Edited
JTL Hantek says up to 48MS/s. The HScope guide says 48Ms/s are supported by the app: http://hscope.martinloren.com/oscilloscope-tables.html
martin No more than 240 Hz yet. I find the soundcard solution to be much more responsive. It even picks up my ceiling's 100 Hz incandescent flicker from afar while the Hantek oscilloscope receives a flat line. Maybe the photodiode needs an additional external voltage. The audio cable somehow might supply that voltage.
martin my setup is enclosed in a case so it's not easy to photograph. I started off with a cheap Amazon USB scope but found that I was hitting the limits pretty easily. My present scope was rather expensive and does an actual 500 MS/s so it won't ever be the weak link.
The special sauce is having a really clean power source and then adding a 100,000 ohm (or higher) resistor to amplify the fluctuations coming out of the diode. Also, be certain that the diode's switching rate is sufficiently fast to capture the frequency you're looking for.
@Wootever @TechSensitive The OPT101 PDF (first page) says "Bandwidth: 14 kHz". Does that mean no more than 14 kHz can be detected? Any thoughts?
http://www.ti.com/lit/ds/symlink/opt101.pdf
I got all the components, and did very bad solder work. To my surprise, the oscilloscope detects light input. But it's no better than the BPW34 yet. I can't get the 9V battery to work. I hold one end at VCC and the other end to any of the two probes, but the signal won't get amplified. Also the first signal was negative, so I had to swap hook and grapple. I don't know how to make progress. It seems impossible for an average person. I believe without the battery it won't detect any higher frequencies.
I found this device in internet. Could be useful. https://www.quarta-rad.ru/en/catalog/luxmetr/luxmetr-pulsmetr-yarkomer-radex-lupin/
- Edited
valex13 The price looks good but they don't say anything about the sample rate, which seems suspicious. It could be anything. Am tempted to buy it, but where's the detailed data sheet?
Edit: Found some offers on eBay in which the sellers claim the sensor operates at 6000 Hz. That would be way to little for our purposes.
Wootever How would one know the correct orientation when inserting the OPT101 into the CJMCU-101? I have another set of parts but the next time I want to do everything right. And how did you know where to put the red and black battery cable before soldering them? It seems those two tasks, apart from decent soldering, are the most important ones.
I just fried my 2nd OPT101 by connecting the 9V battery. The sensor got pretty hot and then ceased to function. Pretty frustrating. Don't know what to do now. Maybe this approach is a dead end.
- Edited
If anyone still reads here, in a final attempt to measure anything useful I bought the handheld flicker meter Fauser LiFli ($300), and maybe the most important thing I found out so far is that flicker-free monitors do flicker, although very little, and even at 100% brightness. My two BenQ EW2740L flicker at 2% and the totally unusable Dell U2515H flickers at 4%. Still doesn't explain why one of the BenQs is not perfectly usable. Maybe very small differences in percentage make the difference. But 2% is the least you can measure with this device. My cheap Hantek 6022BE oscilloscope which I connected to the LiFli seems to suck too much to pick up such small differences. Currently I'm not willing to spend any more money for a better oscilloscope. I have an LED bulb that flickers at 4% when you turn it on and then it slowly continues down to 0% over some minutes. During the first seconds I get PWM symptoms. Later this LED is much better. So it seems the safe percentage for me might be somewhere between 0% and 2%. Which is insane since we're talking of frequencies between 20 kHz and 400 kHz. However, even this 0% LED is still buzzing slightly, made audible by the LiFli device's speaker. So there still is some ripple. It seems true flicker-free is not possible with wall-plugged devices. My small LED battery-driven flash light is at 0% without any buzzing.
I didn't measure flicker percentage (a-b)/(a+b) but ripple percentage (a-b)/a, which makes more sense to me.
Where a is max and b is min voltage.
https://www.fauser.biz/li/lifli_e.htm
Edit: With my new oscilloscope setup (a few posts below), it turned out that the low percentages (like 0-2 %) the device measured was probably the display's 60 Hz refresh rate flicker. I did not even know that LCD matrixes would pulsate at each refresh when displaying a constant image, but apparently they do.
- Edited
The 6022BE has much trouble with low voltages:
In the picture you can see a ripple of roughly 20%.
To me it looks like it is impossible to spot amplitude differences of 1% and even below, which seems necessary. Can any oscilloscope do this actually? If anyone knows how to amplify the signal before it enters the oscilloscope (it is a 3.5 inch audio to BNC cable), please let me know. Preferably with some purchasable device.
Just stumbled upon this post:
Like us he was interested in measuring PWM and compared different photodiode setups and their efficiency. Great work, must read!
http://budgetlightforum.com/node/61254
