DSO NANO - how we users can impact next hardware design !!!

[traugott.guenther]

Hi all -

as we all have been fascinated by the DSO nano and its potential so far (I assume)
we still might have lots of ideas , proposals , recommendations which can make this baby even more flexible, useful and successful
by performing a hardware redesign. Right now the DSO nano still has some way to go to become a real measurement instrument BUT
it can grow to this "size" if we are interested to get it there and support the SEEEDSTUDIO people to move in that direction (as they now
have finally moved to this fine new office they can take care for the nano, can't they ? ).

So here is the topic to get yourselves involved in this HW redesign - I open it up in alignment with the SEEEDSTUDIO guys

PLUS

I heard some rumours that they are really curious about our inputs and as more as we make them excited as more they might honour the related proposals


Lay down your ideas in the following posts as I will do with the next post - hope we can make this little beast even more attractive

All up to us

Cheers from Traugott

[traugott.guenther]

As mentioned in the initial post I will start to add some ideas/proposals for the next HW design.

All my comments are based on the dismantled DSO nano (3 modules : main board with ARM Cortex, plugin board with probe connector/card connector/USB connector/switches and display board).

As of my impression the DSO nano is based upon a multimedia player device in housing and plugin board. This was very good thought to keep cost low but also means some compromises
if one looks out of the perspective of a measurement instrument.

So here are my first proposals :

1) Design a new plugin board as a 4-layer-PCB to enable good signal shielding (at least one ground layer);
move the OpAmps, prescaler chip (yet : HC4051) and related R's, C's, L's as close as possible towards the probe connector.
The stereo plug/jack connection can stay but might be prepared to be used as diffential probe connection (as the jack/plug provide 3 connections).
This would allow to make a move towards AC coupling easier. As long as we stay with the ARM internal ADC we can not exceed 500 kHz bandwidth anyhow.
For the PCB design provide full symmetric layout at least until the inputs of the Op Amps to lower noise impact. As the housing of the DSO nano is not really
an EMC shielding it might make sense to add a small metal shield over the signal input section - this would also help "against" impacts from the USB port and/or card slot.
Such shielding material is widely used in the mobile phone industry - so it should be availbale on the market.
btw : this does not automatically mean a redesign of the main board but would be a good idea instead of using "flying wires".

2) Use a better OpAmp; even if the TL082 is a really good and well known workhorse (and therefore also cheap) it is used at its low end of operating voltage in the present
design; it's also more of a general purpose chip and not so much trimmed for measurement purposes. I have read of sync problems with the DSO nano - those can be based
on production lot variations of the chips which will show up very much if the device is driven close to its operating limits (here : close or maybe below the lower Vop limit of 6V acc. the datasheet I have from ST)
There is a large selection of high precision, high slew rate, low input current etc.pp. OpAmps out there.
So here is my quote : Texas Instruments OPA 2141 - for sure would need a new calculation of the input voltage divider chain - but I think it's worth it in regards to measurement
accuracy and flexiblity. I assume that the other passive SMD components have been selected with a very low tolerance on values.

3) In regards to add extensions to the DSO nano I recommend to bundle the yet unused I/O port connections of the ARM Cortex and add something like a "panelmate connector" (TM of Molex) for them
to the mainboard. This opens the opportunity [together with some SW enhancements!] to add other ADC boards, special probes, data interchange capability with other
DSO nanos, Logic analyzer functions...)

4) Add full realtime USB interface capability to the DSO nano (is more of a SW issue I think) - this would allow to integrate the DSO nano into an automated measurement setup which might
be based on the National Instruments LabView suite(s).

In regards to the production of the DSO nano I recognized quite a bit of solder tin on the connectors of the ARM chip <=> points to hand soldering. This should be done via (hot air) reflow
to lower the risk of shortages. Even lead free tin can still "flow" some µm if the chip gets heated up and cause shortages or increased crosstalk.

Hope my inputs can help and are interesting to the (desireably) growing DSO nano fan community

All the best

Traugott

[slimfish675]

I'd like to comment out some of the ideas posted by Traugott and to add some more:

1. This is a low range Oscilloscope, but input stage could be improved like traugott said. The sampling rate can also be enhanced by using both integrated ADCs to do interlaced sampling, so theoretical sampling rate tops at 2Msps. Battery measurements are marginal, as they only need to be measured once per second or less.

2. Full realtime USB interface is unattainable as far as the USB is 1.1. In numbers: 1Msps * 12bit = 12Msps (top limit for USB 1.1). Maybe with lower sampling rates (say 100K) and less precision (8 bit)... . MicroSD can be used to store longer captured signals...

3. Battery charger: lithium, charger & diode are never in the same sentence. It works for a couple maybe three times as long as the user don't try to charge batteries longer than three hours, otherwise battery is usually harmed. If you want to have a good product and keep the price low, you could use a low cost alternative of LTC4054 (like pin compatible Microchip MCP73811 - 0.5$) and get rid of the diode. I have the Nano for more than four months and the battery works more or less like the first day. But after buying it, the first thing i did was to remove the diode (even before trying to charge it).

4. Power management: what happens if a user forgets on/off switch ON -> battery is completely depleted. Display ilumination is regulated by a XC6206... replace it by a cheap (2N7002, BS270) transistor and use a PWM pin to toggle it (light intensity control/off). Using Sp3232 to generate +-5V is a neat trick, but it can't be disconected. Another transistor... and voila. By the way, SP3232 is very cheap but it needs a "lot" of space. If the opamp used is rail to rail, it can be replaced by a tiny and also cheap LM828 (SOT23-5) inverter for the negative rail.

And suggestions doesn't end here...

[hobbes]

Dual channel! I don't think there is a lot to be gained from going to interleaved sampling to get 2 MS/s on one channel. 1 MS/s is plenty to cover basic audio-frequency use and 2 MS/s isn't going to suddenly make it suitable for high speed logic or RF.

What I would love to see is 2 input channels plus a function generator output. That way I could apply a test waveform to a circuit and compare the output. Ideally there would be a built-in sin / square / triangle generator plus the ability to load a arbitrary waveform from flash. I now use my laptop sound card as a function generator. That works fine, but the maximum frequency is limited. Also, since the nano is single channel only I can't display both the test waveform and the output.

I know the enclosure is part of keeping it cheap, but I am not thrilled with the d-pad control. I very frequently hit left/right when I mean to hit up/down or vice versa. A set of more standard "soft buttons" on the right side of the screen plus value up-down controls would be a big improvement. If that is possible within the target price range I would buy a second one.

[peterjohnston]

Either that or better firmware which clearly indicates what each button does. Perhaps popping up when a button is pressed and disappearing again after a delay.

[toalan]

Put in a digital com port that is accessable to/by other TTL/CMOS devices; SPI, I2C, UART. I would like to mod the firmware to use the DSO nano as a display for some of my automotive products. The digital com port would also allow to add external peripherials to increase the functionality of the scope; logic analyzer, external SPI ADC, etc...

[dnordenberg]

Put in a digital com port that is accessable to/by other TTL/CMOS devices; SPI, I2C, UART. I would like to mod the firmware to use the DSO nano as a display for some of my automotive products. The digital com port would also allow to add external peripherials to increase the functionality of the scope; logic analyzer, external SPI ADC, etc...

You could include a usb->serial adapter driver in your custom firmware

[felixpoor]

As I just have to explain a Nano to an employee who wants to measure only a few things, and as I'm even not so fit in electronics, here is a proposal which would bring wider acceptance and higher volumes:

Implement an additional "simple mode", where the Nano works as easily as a voltmeter, but displaying an oscilloscope signal curve also. Somewhere in a corner of the display the digital numbers for voltage, maybe Vrms and may be frequency. Readable from a distance of about 1 meter.

EVERY technician, no matter whether working with cars, washing machines, office machines and so on needs something like that, and every electronics engineer also if he has to solve simple tasks.

Like 50 years ago, a voltmeter and an oscilloscope from HP, but now on 5x11 cm² for less than 100 €. That's about it!

PS: Nowadays you can buy such a multimeter with oscilloscope for 200 € to 300 €, but if you could do it with this little thing for the shirt pocket for about 100 €, that would be pretty nice.

[dnordenberg]

Wow, that almost sounds dangerous, nano != multimeter. I hope you are atleast using a X10 probe for your higher voltages multimeter adventures. Thats the thing with multimeters, you can plug them in almost everywhere without worries, most works with all voltages up to 1000 volts or so and any serious multimeter have all kinds of protective things that should protect you no mather how wrong you are using it, it blow up and hurt you etc. A oscilloscope you don't plug in unless you atleast know what voltage range you are supposed to be in. My multimeter I would never substitute with my nano.

Be careful!

[felixpoor]

Yes, you are right, totally. But I did not want to make my post longer so I waited for somebody to solve this problem. Of course I miss a x10 probe now, maybe with a little switch x1/x10. And of course there is a need for input protection. Sorry for misunderstanding: I did not want to create another multimeter or an oscilloscope for high voltage. My proposal is, to improve and simplify one of the most used functions in daily life (voltage in numbers and shape). This function is already implemented now, but - at the first glance - hidden very well under a lot of lines, colors and options.
My proposal is inspired by the small format of the NANO: when working on your table you can use a larger instrument with larger bandwith etc. But working outside with a pocket size instrument often means working under pressure and in poor conditions, and you will be glad to see just the most important without need of make settings.
I think, the topic is "how to improve the NANO" and that's just a proposal. It was born by the fact that the NANO would not lie on my table in order to explain it to an employee if it wasn't a little bit complicated.