DIY oscilloscope kit build report

[reddesert]

Hi gang,

There have sometimes been questions about cheap DIY oscilloscope kits here. I just built one and will post some pictures and comments. Feel free to use this thread to ask questions, post your own kit experiences, or suggest things for me to try it on to test its usefulness.

The kit I got is a DSO 150 made by JYE Tech: http://www.jyetech.com/Products/LcdScope/e150.php  This is supposed to be an improvement on the DSO 138 model that I've seen a few posts about. Notable changes include a rotary encoder for setting parameters (like volts/div, sec/div) and that it comes with a compact housing.  I got mine from the inartfully named banggood.com, which sells a lot of electronic kits and whatnot. (And drones. Prepare to get marketing emails advertising drones.) JYE Tech says there are a number of knockoff sellers on ebay or aliexpress, so be aware of that.

I got a version that has all the SMD parts pre-soldered. It was just US $20.
I should have ordered an inexpensive function generator kit at the same time.

This is what comes in the box, including 2 sheets of nicely detailed and legible instructions. A main board with display, analog board, bag of components, clip leads with BNC connector, and plastic case.

Edit: you will also need a 9 vdc center positive power supply.

[BrianS]

That looks real interesting.  Plse keep info flowing on your progress and end result.

[reddesert]

The SMD parts are all pre soldered. You have to solder some power connectors, pin headers, and 4 tactile switches onto the larger main board, and various resistors, capacitors, pin header, and the BNC connector onto the smaller analog board. The rotary encoder gets its own mini-PCB and pin header.

The PCBs are double sided and seem good, though they are thinner than a typical DIY pedal PCB like Madbean's, and the pads are quite small. I would say you need a solid B-B+ soldering game. It's not especially difficult, but because the pads are small, you need to be sure that the iron actually contacts and heats the pad enough to draw the solder onto it. I inspected every joint with a magnifying glass and had to redo a few that didn't look properly filled. Also, watch out for little solder bridges between adjacent pins. There's no need for special equipment - I did it all with a 1/8" chisel tip iron - you just need to go slow and be careful. There are not a ton of components, certainly fewer than some medium-size pedal builds.

The directions are pretty good and have a sensible order of assembling the components. It's mostly obvious which side of the board a component should go on, but if you get confused, just try dry fitting them together.

Here are pictures of the soldered-up boards from both sides.  A couple of suggestions: 1. The display on the main board is attached only by a ribbon connector. I put a little masking tape on it to keep it from flopping around and tearing during assembly. 2. The only change I'd make is to make the test connector tab (at top of the main board) stick out a little further, by bending its pins differently.

Edit: The resistors are tiny 1% resistors, and some of them are large values >2Mohm, outside the range of a typical DVM. I suggest being very careful about sorting them by color codes and measuring the ones that you can.

[reddesert]

The next step is to test some voltages for correctness.  First, you solder the little rotary encoder board onto the main board. You then fit the main board and analog board together using the 2x5 pin headers, plug in the power, and test some DC voltages at test points that are marked on the analog board. Mine all came out fairly close to the nominal values - there are a couple that are supposed to be 5.0 V +/- 2% and mine were more like 5% high, but I'm not super concerned about that (it could even be my multimeter). The instructions offer a few troubleshooting suggestions if your voltages are off.

Note that the required power supply is not included; it is 9 VDC but CENTER POSITIVE. In the rest of the world outside pedals and Boss electronics, center positive is much more common. I don't know if you'll burn anything by plugging in a center negative supply, so don't do it.  The current draw is rated 120 mA, so it won't last long on a battery.

[reddesert]

Now you need to trim the compensation of the analog input.  This is accomplished with two small adjustable 25 pf SMD capacitors, C3 and C5 on the analog board schematic. BTW the schematics are available on the JYE-Tech website, which is pretty cool for understanding how it works.

To do this, you attach the probe clip lead to the test signal tab, which puts out a 1 kHz square wave.  Then you follow instructions to trim the capacitors with a small screwdriver until the wave is displayed as square.  (Note: in the previous DC voltage step you had to set the input to GND, grounding the input, but you have to switch it to DC or AC to see the test signal. This is the one thing they forgot to say explicitly in the instructions.)

In the first picture, the compensation isn't quite right, so the test wave doesn't look quite square. In the second picture, it's better, with maybe a little overshoot.  IIRC, basically all oscilloscopes require some compensation to take out the properties of the R-C network  on their input.

[reddesert]

Finally you can remove the power, separate the boards, and then assemble it into the case. This was also not very hard, assuming you soldered all the parts on straight. This picture is a side view as the boards are going together and you can see the pin header that connects them. I think you could disassemble and reassemble it fairly easily, e.g. if you had to adjust the compensation for a different probe.

In this picture it looks like the top panel is slanted/curved. This was bugging me until I realized the outer case is curved to match it.

[reddesert]

OK, let's look at a signal!  Like, hmm... well there is a Duovibe circuit sitting here.  This is the square wave of the LFO, on the DC input setting of the scope, at 1 V/div and 0.2 sec/div.  (All the other clip leads in the pic are just how I wired the Duovibe board to its pots.)

[alanp]

That is insane value for $20. O'scopes are extremely useful things for dealing with waveforms

[reddesert]

It sure is!  I also think it's a well thought out kit. The instructions are clear and there is supplementary information available on their website.

Now a cautionary remark. Oscilloscopes are relatively complex test instruments and you will get more out of one if you read up on how to use it and practice a little. For an example of the kind of simple mistake you can make, when I initially connected the scope to the Duovibe LFO, I switched it to AC input.  The LFO is AC, right?  Well, I got the attached weird looking waveform. AC on a scope uses a blocking cap on the input - here it's 0.1 uf, from the schematic they kindly provide - and the input resistance to ground is about 1 M, so there is a highpass RC filter at a few Hz. Here the blocking cap is charging/discharging with the square wave. For LFO work you need to use the DC input. That applies to any scope, not just this one.

[sharpan.sv]

I bought this scope to calibrate bbd choruses, it's working great.

I've tried center negative adapter (accidentally) and there is a diode protection, so nothing were harmed.

Отправлено с моего Lenovo A536 через Tapatalk

[stringsthings]

Excellent build report!

[reddesert]

I bought this scope to calibrate bbd choruses, it's working great.

I've tried center negative adapter (accidentally) and there is a diode protection, so nothing were harmed.

That's good to know about the diode protection.

Here is another picture showing something maybe useful. I have here a cheesy Rocktek Analog Delay pedal, which uses a CD3207 (not much of a delay), and I'm using the scope on a pin of the CD3102 clock generator to look at the clock pulse. It's a bit more than two divisions at 20 usec/div, so the clock frequency is a bit under 25 kHz. This pedal doesn't have a lot of trim adjustments, but if you did need to adjust the clocks the scope would clearly be useful.

At this point I thought it would be even nicer to have a dual-trace scope so you could look at both clocks together (although I don't think you absolutely need to). For some types of circuits where you care about relative phase or delay, a dual trace scope would be nice. But this little thing is quite functional, and easy to put away. Maybe the closest alternative is this cool Gabotronics board with a USB interface: http://www.gabotronics.com/oscilloscopes/xprotolab-plain.htm. There are commercial USB oscilloscopes, but they cost more and I'd be concerned that anything with a proprietary software interface will eventually be a paperweight a few OS versions from now. Vintage Tektronix oscilloscopes are awesome, of course, but also take a lot more room.

[MacHeath]

Great build report, there's one coming my way too now 😀

[reddesert]

A forum member asked me some questions about the use of the scope, and I figured this ramble was worth sharing so others can read it:

An oscilloscope is basically like a time-variable voltmeter ... so you hook it up to a circuit as if you were measuring voltages. Typically, you put the black lead on ground and the red lead on whatever signal you want to test, although I can imagine less-common circumstances where the black lead would be somewhere other than ground, like if you measured the voltage across a component. Also like a voltmeter, you're fairly unlikely to break anything by probing the circuit, especially at 9V. Just try not to accidentally short two parts of the circuit together - this can be easy when probing IC pins. A real oscilloscope probe, which has a very narrow spring clip on the positive end, is useful.

The main subtlety of an oscilloscope is getting it to trigger properly. This means where it starts the sweep across the screen. Imagine you are looking at a repetitive signal like a sine or square wave. If the scope triggers at the same point on the wave every cycle, then the trace on the screen will look stationary. But if it isn't triggering at the same place, the trace will move or wobble, like an old analog TV that is getting bad reception (it's the same principle).  This scope has an auto trigger, and a manual setting where you can adjust the trigger level and whether it triggers on an up or down slope. You have to kind of fool around with it to get used to triggering.

If your input signal is transient, like a guitar strum, or has a lot of harmonic content, then it often won't trigger cleanly, but you should still see the signal rise and decline.

What kind of signal you put in depends on what kind of circuit and what parts of it you are trying to test.  For example, to check that the LFO of a phaser is actually varying, you would put the red test lead on the place where the LFO outputs to the rest of the circuit. You don't even need an audio input signal for that, since the LFO should be running regardless of audio input.

If you had a dual-channel scope, you could send a waveform into the audio input and use the scope to look at the input audio signal and the phase-shifted audio signal to watch the phase-shift between the two, but that is a level-up of sophistication.

On the other hand, for a compressor, there is usually a sidechain that rectifies the input signal to generate an envelope, and uses that as a control voltage to control the gain of an amplifier. So there you might want to use an actual guitar strum as the input audio signal and use the scope to probe the control voltage. It should vary with the attack of the guitar strum. If you put in a constant waveform at the audio input, the control voltage will probably be constant. But if you turn the amplitude of the waveform up and down, you will hopefully see the control voltage from the sidechain also vary.

For applications like the LFO or the compressor envelope, where you're looking at a signal that varies maybe a few times a second, use the DC mode of the scope, but for looking at audio frequency signals, you can use AC mode (which blocks slow signals with a capacitor).

[joltjrs]

for 20 bucks you can't beat that with a stick. i'm from the auto world so i'll just break out my snapon vantage pro. cool shit though.