Questions about DSO Shell and proposals

Post Reply
Lars
Posts: 6
Joined: Sat Dec 10, 2016 12:23 am

Questions about DSO Shell and proposals

Post by Lars »

First of all I would like to post, that it was easy to build the 15001 kit. Congratulations, you set up a very good kit! Especially I like that you avoided mechanical switches and therefore placed analog multiplexers!

When trying to calibrate and to run the scope, I've had some difficulties and questions, but also two proposals:

1. getting a "good picture" of a rectangle of more than 10kHz
2. calibrating at internal signal generators' amplitude of 0,1V shows no effect
3. start an aquisition in "single" mode
4. holding a signal with the trigger
5. abnormal spikes shown at battery (operating) voltages higher than 9V
6. abnormal waveforms shown in sec/div ranges that are much "slower" than the measured frequency
7. proposal: operation of two 10440 Li-Ion batteries
8. proposal: "easy" second input with other multiplexer

Why can't I see the frequency of the shown signal, like it is displayed at the DSO 138?

But now to the details: Link to the pictures: https://www.dropbox.com/sh/bue3b5v6p0cf ... SUG0a?dl=0

1.
Already at 4kHz and 5Vss, at 50µs/div and 20µs/div, I see overshoots of more than 1V at the end of the edges in shape of a shark fin. These overshoots do not exist in reality! (see pictures 163.jpg and 88.jpg!) Image and Image Image
What element is responsible for these overshoots? More than this, the switching seems to be much slower than in reality. Unfortunately, this is misleading!
What is the maximum "useful measureable" rectangle frequency? I mean, that it still looks like a true rectangle. 20kHz or even higher? What elements are the first responsible/limiting ones? The 082/084? Or the AD speed of the STM? Would it make sense to replace all 082'/084' by 052's/054's with a faster response and higher slew rate? Or is this nonsense (marginal, compared with other effects)? The measuring interval seems to be around 2µs? This would explain the slow increasing slopes of a "sharp" 10 or 20kHz rectangle which is watched at 10 or 20µs/div. The 10 and 20µs/div are misleading me at a 20Khz rectangle. (See pictures 49.jpg and 887.jpg of 20kHz at 20µs Image, Image and even more 644.jpg and 354.jpg at 10µs Image, Image) The last almost realistic looking range seems to be 50µs/div for a 20kHz signal.
I tried to filter away these overshoots with a maximum (30pf) adjusted trimmer: almost no effect. (See picture 233.jpg Image) Only with a 180pF capacitor instead of C5 + C6, I got a "helping tradeoff" at 20µs and 10µs ranges. (see picture 411.jpg Image. Picture 256.jpg is at 1kHz Image) At 50µs, the 180pF seems to be a bit retarding, but is still ok for me. Why is the curve shape at 20µs so much different from 50µs? (see pictures 498.jpg Image and 131.jpg Image)


2.
I was unable to calibrate the waveform at the 0.1V range of the 1kHz signal; the dc offset of that signal is about 3V, so I had to switch to AC coupling (instead of the instruction manual, where "DC" is shown in the picture). When in AC, turning the Trimmer Caps didn't show a visible effect. It looked quite ok before the calibration trial and so does after. I soldered out the TrimCaps to get to know, where is max and min during a 360° turn. But also this didn't help me. Like said under 1., now, I have a 180pF capacitor soldered instead of C5 + C6.
I tried to find the reason for the 3V DC offset in the signal of the internal 1kHz reference generator. Got to know, that R31 (470 in the schematic) has only 22 Ohms in real. Is that ok? Or by mistake? What would happen, if I would change to 470 Ohm? R16 is 1k, like in the schematic.

3.
From time to time, I see no difference between "norm" and "sing" trigger modes or I am not able to start a single aquisition.

4.
For instance a 120Hz signal: sometimes I am not able to hold it with the trigger. It is rolling through from right to left. The same with a 0.1Hz rectangle (at 2 sec/div).

5.
I got to know, that in fact, the voltage must be significantly less than 9V. From 9V on, I saw negative spikes "down to the bottom" with a 10ms width over all V/div. Even with Gnd coupling! They disappeared at less then 9V Vin. What is the reason for this? In don't think it comes from the LM1117 regulator?!
At 1000Hz 5Vss rectangle and 0.5ms/div and 9V operating voltage, every second, I saw long spikes at the trailing edges (not the leading!) of pos and neg swings (slopes). Is this from an interference with the switching frequency of the ICL7660? Or from "high" voltage? (see picture 407.jpg!)Image. In the 0.2ms/div range I can see this behaviour almost all time, at 0.1ms/dic, I see it all time.

6.
When I watch a 20kHz rectangle at 0.1s and 1s/div ranges, I see extremely misleading artefacts. (see pictures 969.jpg Image and 46.jpg Image) I would expect nothing else but a big beam (balk, bar) (see picture 430.jpg on the comparing scope Image)

7.
Further, I have a proposal: I think it would be possible to operate the scope with just two "10440" Lithium-ion batteries (accu). It seems, that there is enough space to place such a battery holder between both "back PCBs" (see picture 797.jpg Image). You only had to change one plastics part: the back side. The other parts could stay the same! (see picture 440.jpg Image)
May I operate the DSO from two 18650 Li-Ion cells ( (6.6) 7.0-8.3V)? What issues do you know at voltages less than 8.0V? Does the DSO show a sign, when the Voltage is that low, that it affects the accuracy? Is a power source inner resistance of less than 0.1 Ohm maybe a problem?

8.
Why did you choose a 4053 for the first multiplexer? When another 4051, with a new firmware, we could maybe enable an optional second input channel (multiplexed)! Or is the HC4051 not fast enough to switch between inputs?

Many thanks in advance for your answer!

Best regards, Lars Krueger
Lars
Posts: 6
Joined: Sat Dec 10, 2016 12:23 am

Re: Questions about DSO Shell and proposals

Post by Lars »

unfortunately, I can' effectively link the pictures from dropbox. Hope it is not too extensive, to find the pictures?
jye1
Posts: 1221
Joined: Wed Feb 20, 2013 4:58 pm

Re: Questions about DSO Shell and proposals

Post by jye1 »

Thank you for posting the questions and suggestions. Since it need some time to answer these multiple questions please allow me to postpone it to a little later when I have larger time slot. I'll get back to this thread as early as I can.
Lars
Posts: 6
Joined: Sat Dec 10, 2016 12:23 am

Re: Questions about DSO Shell and proposals

Post by Lars »

Hi jye,
of course, take your time!
I just ordered another DSO Shell in order to compare both ;-)

Best, Lars
jye1
Posts: 1221
Joined: Wed Feb 20, 2013 4:58 pm

Re: Questions about DSO Shell and proposals

Post by jye1 »

1. The most limiting factor for bandwidth is the sampling rate available. 1Msps is barely up to work for 200KHz bandwidth. The analog channel has a pass band about 300KHz. But this can be improved by re-arranging gains of each stage. When timebase is set to 20us or 10us interpolation is used to make up the samples between real samples. This results in overshoots for sharp edges.

2. There is a bug in the firmware that caused a 3.3V DC overlay on the 0.1V signal. Please use AC couple as workaround. See http://www.jyetech.com/forum/viewtopic.php?f=19&t=1025 .

3. The trigger won't work when timebase is set to 50ms or longer. It was implemented but didn't work properly as it should. This is the same as DSO138. When timebase is 20ms or shorter NORM and SING modes should work.

4. As said in 3 above. In Rolling Mode (timebase is 50ms or slower) the trigger didn't work correct.

5. The analog channel is sensitive to power supply noise. The spikes are more likely being seen when dirty switching mode power supply is used. This can be verify with a linear PS.

6. When timebase is slow the sampling rate is also reduced. So it gives a false image of a signal if sampling rate is close to or less than 2X of signal frequency. This is true for most digital oscilloscopes.

7. A possible issue using two Li-ion batteries is voltage not enough. The negative voltage AV- is strongly dependent on input power supply voltage. With 7V and the protection diode D2 removed the scope would still work. Only the dynamic range of analog channel (the negative end) would be reduced.

8. Not quite clear how you connect the OpAmp and swithch. The current design aimed to provide as wide sensitivity range as possible with minimum parts.

Since many pictures can be seen I tried to "imagine" the waveform by your descriptions. Please freely let me know if I am wrong. Again, thank you for posting the questions and suggestions.
Lars
Posts: 6
Joined: Sat Dec 10, 2016 12:23 am

Re: Questions about DSO Shell and proposals

Post by Lars »

Hi Jye1,

thanks for the detailed answer! Yes, mostly you understood me right and I think, you have chosen the right picture to each issue. But aigain to some points:
1. Mostly, I use the "insulation advantages" of battery powered scopes for measurements of Mosfet's gate signals in switched mode power supplies. Is your DSO138 giving better readings at 20kHz rectangle and 20µs or 10µs/div, compared with the DSO Shell? Is the same interpolation used? Or is there a difference between these two scopes?

2. and 3. Will this be implemented with the next firmware?

6. I compared the 20kHz signal with a simple voltcraft scopemeter. When you look at the link
https://www.dropbox.com/sh/bue3b5v6p0cf ... SUG0a?dl=0
it is the 15th picture:
https://www.dropbox.com/sh/bue3b5v6p0cf ... 0.jpg?dl=0.
There are no misleading "slow" curves to be seen. Although I think, that the VC1008 IS a digital scope. I understand what you explained about most DSO's, but is there no chance to alter the behaviour of YOUR digital scope? Where's the need to reduce the sampling rate in slow timebases?

7. I'll try with a SB130 schottky diode in parallel to the existing D2. Will there be a warning if the voltage becomes too low? (if less than eg 7V?) When not now, maybe with the next firmware? I'll definitely try to build a battery compartment for two 600mAh 10440 Li-Ions in the back side of the case. Though, it would be a very charming little standalone scope, even smaller than a velleman HPS140 ...
http://www.velleman.eu/products/view/?id=384090

8. in http://www.next.gr/meter-counter/meters ... 14169.html is (a very old) example, how more inputs can be switched to a single input scope. My thought was: if you already use the 405x multiplexers, wouldn't it be possible to not only statically switch between 2 gains of just one input, but alternatively switch between 2 separate inputs? Of course you still need an input stage in front of each multiplexer input, which leads to a higher parts count ...

You speak about an improved firmware. Will there be a chance for the existing customers to get this new firmware? Will it be possible (and easy) to flash this new firmware with "normal means", eg an USB-serial or USB-I2C or USB-SPI adapter or Atmel MkII programmer? Unfortunately, I don't know nothing about reading or flashing STM's Arms ...

Thanks, Lars
Lars
Posts: 6
Joined: Sat Dec 10, 2016 12:23 am

Re: Questions about DSO Shell and proposals

Post by Lars »

in 8. I meant "alternatively switch FAST between two inputs"
jye1
Posts: 1221
Joined: Wed Feb 20, 2013 4:58 pm

Re: Questions about DSO Shell and proposals

Post by jye1 »

Hi Lars,

1. The firmware cores of DSO138 and the Shell are pretty much the same. Same interpolation. So for the case you talked about there will be no difference.

2 & 3. We'll do this as soon as possible. Hopefully get it done in the up coming release.

6. The main reason to reduce sampling rate at slow timebase is because of limitation in memory. When timebase is slower the time frame in which we view waveform becomes longer. If sampling rate is not reduced that means it requires more samples to fill up the time frame.

7. D2 can be shorted actually. It acts as protection for reversed power supply polarity. If correct polarity is ensured the scope can run without it.

8. The MCU has two ADCs. They can be used to build a dual channel scope. We are working on this project. Hopefully it can be finished by the middle of 2017.

It shouldn't be too difficult for existing customers to upgrade firmware. Only a USB-serial converter is required. Please see http://www.jyetech.com/Products/LcdScop ... pgrade.pdf . This is for DSO138. The Shell is the same.

Thanks.
Lars
Posts: 6
Joined: Sat Dec 10, 2016 12:23 am

Re: Questions about DSO Shell and proposals

Post by Lars »

Hi Jye,

many thanks! I wait for your upgrade! ;-)

Best, Lars
Post Reply