Page 4 of 5

Re: DSO138 Oscilloscope + NBTV signal test

PostPosted: Wed Feb 22, 2017 10:10 pm
by Steve Anderson
Which shipping method did you select? I can't find any except the real expensive options for delivery to Thailand...seems the same for Australia.

Steve A.

Re: DSO138 Oscilloscope + NBTV signal test

PostPosted: Wed Feb 22, 2017 10:18 pm
by Andrew Davie
Steve Anderson wrote:Which shipping method did you select? I can't find any except the real expensive options for delivery to Thailand...seems the same for Australia.

Steve A.


Shows free shipping to me?!!

Re: DSO138 Oscilloscope + NBTV signal test

PostPosted: Wed Feb 22, 2017 10:20 pm
by Andrew Davie
Steve Anderson wrote:Which shipping method did you select? I can't find any except the real expensive options for delivery to Thailand...seems the same for Australia.

Steve A.


To make sure it wasn't my account getting "specials" I logged out, reloaded the URL, and selected Thailand as the shipping destination.
STILL shows free shipping. No idea what's happening.

Re: DSO138 Oscilloscope + NBTV signal test

PostPosted: Wed Feb 22, 2017 11:50 pm
by Andrew Davie
Andrew Davie wrote:Here's the item I purchased. Currently the cheapest price (including shipping!) that I can find. Just over US$21
DSO Shell Oscilloscope from AliExpress


Have had it running of a 9V battery for the last hour or so - at some point I'll work out the power draw.
Also considering purchasing this LiPo battery as it's almost the exact dimensions of the case. could just glue on the bottom and give long-duration power-brick free usage. I know it's 7.4V and the unit requires minimum 8V, but LiPo are in my experience around 4.2V/cell - hence this will be about 8.4V for quite some time during usage.

Re: DSO138 Oscilloscope + NBTV signal test

PostPosted: Thu Feb 23, 2017 9:00 am
by gary
It's free shipping for me too (60 days though), but I just looked at the original $15 kit and it's no longer available :-(


This fully assembled version is available:
https://www.aliexpress.com/item/Fully-A ... 25959.html

Re: $15 Oscilloscope!

PostPosted: Sat Feb 25, 2017 5:57 pm
by Steve Anderson
Well my two $15 specials arrived today, I've only opened one package and it's a fake. No serial number/sticker and the fonts are different as JYETECH indicate.

The LCD display is loose to, just flapping around on its flexible PCB connector, not fixed to its PCB. Nothing a spot of super-glue won't fix though.

I hope to get around to building them anyway during the course of the coming week - I'm up to my neck at the moment with 'real work'.

Steve A.

Re: $15 Oscilloscope!

PostPosted: Sat Feb 25, 2017 6:27 pm
by Andrew Davie
Steve Anderson wrote:Well my two $15 specials arrived today, I've only opened one package and it's a fake. No serial number/sticker and the fonts are different as JYETECH indicate.

The LCD display is loose to, just flapping around on its flexible PCB connector, not fixed to its PCB. Nothing a spot of super-glue won't fix though.

I hope to get around to building them anyway during the course of the coming week - I'm up to my neck at the moment with 'real work'.

Steve A.


Looks like the AliExpress ones are (mostly?) fake. The one I got from BangGood was genuine.
However, the BangGood one had the "flapping around" LCD whereas the AliExpress one was solid.

Re: $15 Oscilloscope!

PostPosted: Sun Feb 26, 2017 9:29 pm
by Steve Anderson
I started on the first '$15 special' this afternoon. I soldered in all the resistors, small chokes, diodes, crystal and the USB connector - all the low-lying stuff. I took my time, I measured every resistor apart from reading the colour-code. All went well. Taking my time, crossing things off the build instructions in their order, it took me about 90 minutes to get to this stage. I'm in no rush.

There were two resistors that were not required (surplus), this may worry some into thinking they've missed something. All resistors were marked as being 1% components but some were outside that tolerance, not by much admittedly, but I'm assembling it as supplied.

Tomorrow I'll do another hour or so.

Steve A.

Photo of how the board looks at this stage attached. The suspect-looking joints have been cleaned up now after the photo was taken.
Note: The resistors are 1/6W and are about half the size of the usual 1/4W variety, so they are quite small. The three larger looking resistors are actually chokes. The diode in the top-right corner is a usual 1N4004, the diode top-centre is a same-size component.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 5:57 pm
by Steve Anderson
I took an hour away from 'real work' and continued on...I added all the disc ceramic capacitors, the TACT switches and the LED. I will omit the next item, the power supply connector as I plan to use flying leads, also the other power connector when I get to it (Step 1, section 15). Both are in the extreme top-right of the PCB.

Again there was an extra 100n capacitor supplied which wasn't required.

Steve A.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 7:39 pm
by Steve Anderson
I've just realised that the screen backlight LED is fed from the input (unregulated) supply. This is probably where most of the input current goes, by altering the value of R36 you can alter the screen brightness and/or save some current in battery-powered applications. Higher values than 180R = lower brightness and currents.

For more efficiency in battery-powered applications use a CMOS 555 with variable PWM to save power in the backlight. More in due course.

For the time being I'll continue to build it as supplied.

Steve A.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 7:48 pm
by gary
Steve Anderson wrote:I've just realised that the screen backlight LED is fed from the input (unregulated) supply. This is probably where most of the input current goes, by altering the value of R36 you can alter the screen brightness and/or save some current in battery-powered applications. Higher values than 180R = lower brightness and currents.

For more efficiency in battery-powered applications use a CMOS 555 with variable PWM to save power in the backlight. More in due course.

For the time being I'll continue to build it as supplied.

Steve A.


Hacking it before building it? Great stuff Steve.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 8:03 pm
by Steve Anderson
gary wrote:Hacking it before building it? Great stuff Steve.

Just considering possiblities as I get more familiar with the device. This first one will be built as advertised.

Something else I have considered...this micro can run at higher speeds, replacing the crystal with a 16MHz version should double the bandwidth. But with the current software the horizontal scales will be a factor of two out. Where it might say 500us/div it actually will be 250us/div. Again, something for the future.

At above the current 200kHz limit (their specification) the TL084 quad op-amp may run out of steam at the higher gain settings. Always a comprimise in this game.

However, it begs the question, why haven't the original designers (whoever they may be) done this? Perhaps to protect their higher-end products, "We can't have these el-cheapo products becoming as good as our top-end stuff can we?"

Steve A.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 8:46 pm
by Andrew Davie
Steve Anderson wrote:Something else I have considered...this micro can run at higher speeds, replacing the crystal with a 16MHz version should double the bandwidth. But with the current software the horizontal scales will be a factor of two out. Where it might say 500us/div it actually will be 250us/div. Again, something for the future.


PERFECT cue for me to step in. I asked if the company would let me modify the software for the UI and they have told me that yes, they are releasing the UI software to public domain. I have already logged a few bugs which if they don't fix, I will. The "core sampling" software will not be released, but if it's just the scales that are out then that's something I could fix. I get the feeling it's a very small company, but they are very responsive.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 10:03 pm
by Steve Anderson
Andrew Davie wrote:I get the feeling it's a very small company, but they are very responsive.

If it is a small company, like my own, every little outside suggestion does help. Some - the most - are worthless, but sometimes there's the answer we have been looking for in house. Not often, but it has happened.

Who knows where this may lead?

Steve A.

Re: $15 Oscilloscope!

PostPosted: Mon Feb 27, 2017 10:17 pm
by Andrew Davie
Steve Anderson wrote:
Andrew Davie wrote:I get the feeling it's a very small company, but they are very responsive.

If it is a small company, like my own, every little outside suggestion does help. Some - the most - are worthless, but sometimes there's the answer we have been looking for in house. Not often, but it has happened.

Who knows where this may lead?

Steve A.



I have already figured out some bits of the core sampling software. It has a 1024 sample buffer (which is the full width of the captured signal). I believe the onscreen display is a subset of 256 of those samples, so you scroll left/right over those 1024 samples. Now due to a bug I found when you press V/DIV while the display is halted, it gets its voltage Vmax calculation wrong. That means that it's not storing the absolute sampled value but instead a converted value, based on the V/DiV setting. That is, the buffer is already converted to the units selected when it was captured, and the display system just dumps the buffer to the screen as-is. Were it me, I would change that - I'd store the actual value in the sample buffer, and adjust it at display time based on V/DIV so that the voltage calculations could/would always be correct and the highest resolution.

Another issue is the subsampling problem caused when you have a high frequency signal and a lower frequency sample. I'd be trying to drive the sampling at the highest frequency possible - regardless of the SEC/DIV setting and 1024 sample buffer. Then you could average the samples and at least get some idea of what's there that you can't see at the moment. I showed an example of this a few days ago - the sync pulses coming at 2.5ms were displaying as 10ms frequency because I had the SEC/DIV to 10ms and the timing was such that only every 4th sync pulse was being caught/displayed.

Anyway, yes let's hope that we can have a play and make some improvements :)