Scullcom Hobby Electronics #32 - Design & Build 6½ Digit Millivolt Meter Part 1
Вставка
- Опубліковано 21 лип 2024
- In this project we will design & build a 6½ Digit Millivolt Meter with calibration. In this first part of this project we design and build the circuit and do an initial test. In part 2 we will cover the software and final test and build.
Below are links to the full schematic and PCB artwork.
www.scullcom.com/Millivolt_Met...
www.scullcom.com/Millivolt_Met... - Наука та технологія
This is amazing. The professor I never had. In 4 years of school and 15 years of industry, I never met someone who gets it like this and wants to teach. Can't wait for part 2.
+Michael Fox Thanks for your kind comments.
This really is one of the finest of electronics channels around here.
"Hear! Hear!"
+Marty H Thanks that's a nice compliment.
you are great !!!
really.
I've never come across this guy before, but very detailed, very precise, exactly the sort of stuff I like! Have just subscribed!
Yes I second; I have neither seen nor heard of a better hobby electronics project tutorial. I have make a couple Scullcom projects: 5 volt Precision Reference standard and Milliohm Meter using the Barbouri board. The best part of the Scullcom videos comes about halfway through when I say: "Ah, I understand how that works". Cheers, Mark
Lovely design, this will be a really fun project that makes a really great tool. I've already downloaded the plans and started shopping for parts, my son and I are looking forward to the build.
Thanks for all your great work!
+Aerik Forager Thanks. Good luck with the build. Part 2 of this project will be uploaded soon.
This is exactly what I need for my next project! Thank you very much! :)
Great project, thanks for sharing!
Looking forward to part 2!
+Andre Eisenbach Thanks. Working on Part 2 at the moment, finalising the software with possibility of making it easy for the hobbyist to adapt to their own needs. Looking at getting Part 2 finished soon so I can upload it to my channel.
Thank you. Your videos helped alleviate the tedium of my recent, month long, hospital stay.
+Keith Burton Thanks, Glad you enjoyed them. Trust your OK now.
Perfect timing, I just bought my first precision Vref board. Definitely going to try and build this - many thanks!
+JamstaUK Thanks. Good luck with your build.
Another great project Louis, learn something new with every new episode!
+Dino Papas Thank you Dino.
Just came across your channel. It is a job well done! Looking forward to part 2.
+Ventsislav Simonov Thank you, working on part 2 now.
A nice down to earth electronics channel... Can't wait for part two ;)
+Michael O'Toole Thanks. Part 2 is uploading at the moment.
That board pointer would make an ideal replacement for the broken aerial on my Sharp 8080. You make good electronics tutorial videos. Detailed but clear content without a big ego. Thank you sir.
+Inaflap Thanks but I am keeping my board pointer LOL
Nice video, thanks. Can't wait to see Part 2.
+Robert Calk Jr. Thanks Robert.
Excelent work :-) looking forward to see part 2
+Ståle Sætervik Thanks. Working on Part 2 now hope to upload soon.
Gday Mr Scullcom,came across your channel and being a learner im loving it,gathering some parts now from main suppliers here in Australia and going to have a crack at your projects.cant wait to see what you come up with next.I read another comment somewhere regarding a power supply,I think that would be an awesome project to do,especially with your simple way of explaining things.
Keep up the great work .
Thanks for your comments.
Thanks for the upload. As always a great video. Nice understandable explanation ( as usual ). You are a great inspiration for many people.
+Christian Christiansen Thanks for your kind comments, appreciated.
Very clear and considered video as usual, sir. Going to try and follow along with this one. Cheers!
+Benedict Nothing Thanks for your comments.
First rate video!!!
Can't wait for Part 2 :)
Your channel always has some of the best content on UA-cam. THANKS!!!!!
+Ray Reese Thanks for your comment.
I absolutely love your videos !!! Great work, never stop producing awesome content like this video. thank you!
+Mark Webcraft Thank you
+Scullcom Hobby Electronics Your very welcome.
If i could make a suggestion for a part 2 upgrade, I would like to see you design some range switching capability, manual or auto switching. And maybe a way to choose the sample rate through the software to add some stability and accuracy.
+Mark Webcraft Hi Mark. It now looks like I will have 3 parts or more to this project as a result of all the interest. Part 2 which I am just finalising will cover the software and testing. Part 3 will cover improvements and upgrades which could also include range switching.
Very high quality content. Wonderful
Absolutely fascinating...cheers !
Yet another awesome video!
Thank you so much for your time and effort!!! :)
+Benjamin “Ozias” Esposti Thanks.
I like how your so neat with you schematics nice and clean not a bunch of bullshit you can't understand. Good job
+mario rojas Thank you
Yes, I too am always impressed with the neatness and professionalism of the drawings presented, and the concise and yet still complete explanation of the important facts.
Waiting for more ! perfect explanation 🙌
Thanks.
Awesome job!
Very nice to see you have another GREAT video for us. THANKS
+ElectronicWizzard Thanks.
Bravo! Excellent teaching!!
Welcome back, you've been missed!
+Majid Faraj Thanks. I have been busy just lately so had less time for my electronic projects.
Excellent, excellent work. You are amazing.
Thanks. I am just working on an update to this project with additional features. Hope to upload this video soon.
Nice one David......enjoyed that one. Still can't believe you etch your own boards, you have much more patience than me, I tip my hat to you!
+IanScottJohnston Thanks Ian. The longest part of making my own PCB is the design stage as you will know from your own projects. I think you must have got the name David from my Scullcom website, David is my youngest son who is the artist and writes the comics which I put online for him. By the way I enjoyed watching your last project and admire your tidy workbench!!! Regards, Louis
Very informative, as always. Thanks!
+jix177 Thank you.
Very nice project. Thanks for sharing.
+Bruno Villas Thanks.
Great work, thank you for this video.
+Милан Младеновић Thanks.
Awesome, again! Thank you for sharing.
+bajatecnología Thanks.
Very nice, thanks. I might have a go at building this over Christmas :).
+nickhill92 Thanks. Hope to get Part 2 ready for uploading soon so you should be OK for your Christmas project !!!
Brilliant. Thank you!
Many many thanks Louis...
nice one, I like it, you might want to consider putting a limit on the decimal places in the mv range so it does not try to show the Pico volts, this will not be stable no matter what and therefor only serves to confuse the user.
the other thing I would suggest as you seem to have a good update rate is decimation / over sampling, or simply reduce the sample rate of the ADC as they get better when you slow them down
very educational and nice presentation, I like it
+Peter Oakes (TheBreadboard) Thanks Peter for your comments. Good point about the pico volts on the milli volt readings. I will have a look at the sampling rate - I have already tried several options.
Nice. Very elegant.
+SuperBoobaloo Thank you.
Nice project! Ordered the IC's right away always good to have :p) I only have a problem with the zener as protection. The leak current through it can be significant. Rater use a clamp diode (no schottky) to the voltage rail. This has a lot lower leak current so does not effect the reading as much and works as good :)
+tengelgeer Thanks for your comments. I will look at alternatives for the zener protection. Looks like I will be doing a Part 3 of this project with upgrades !!!!
Probably the easiest to watch electronics channels on YT, user friendly voice and no histrionics :o) - any chance of a short video telling us your background in electronics and experience over the years? From young lad to where you are now with your knowledge and experience - your best successes and drastic failures might be entertaining too. Thanks for what you do on YT.
Regards, Stuart.
Thanks. If you are interest in my background there is an overview on my website, direct link is below:
www.scullcom.uk/about/
Can the voltage reference used in this device be used to calibrate another 6.5 digit multimeter?
Hi Luois, I wanted to ask, do you find a 5.1 zener diode for protection in a voltage devider with high impedance gives you some voltage errors around 1 to 3mv once the input hits the 3 volt mark? This is what I found in my circuits the diode still conduct some small current before the zener voltage even 2 volts below. Kind regards, joe
Outstanding! I'm a complete novice when it comes to electronics so it would be a great help if you could go though the schematic and explain why you did it that way, and maybe other techniques that could have been used. The novice is often given the impression that the final circuit is the only way the project could/should be done. Thanks again for a great video.
+Null String Thanks for your comments. Very good point about deciding which circuit design techniques to use for a project. There are many factors I use when designing a circuit which include; which components such as IC's are currently available which meet my needs (this in itself brings a large number of options from different manufacturers), cost is also a factor as is ease of construction. I will take on board your suggestion.
Like the project, pretty nice.
I would suggest as first route of action to lower noise, is change the PCB layout:
Create a star ground and +5V at regulator, from which you feed on independent traces to the digital side (arduino,LCD, pull ups, switch, LED backlight) and the other pair of trace go to only analog side (ADC power supply, OPAmp, Ref, and ground for input) This way the switching noise the LCD creates does not go thru the traces the feed the analog part.
Also, make power and ground traces as thick as posible, to reduce inductance.
I myself would have preferred an RC low pass network between OPAmp and ADC, small R, not big cap, with an RC constant 20) I made a device with a 16bit ADC. The above improvement really helped me, even at 16bit. You are chasing 8bits more.
+Ivan Kocher Thanks Ivan for your informative comments. A star grounding would be the best solution. As I am planning a Part 3 I will take on board your good comments.
1nV resolution🙊 well done🙌
While waiting for some parts I've breadboarded the remaining circuit... I'm using a ATMega328P chip (instead of Arduino board)... so far so good... I have the KiCad schematic and layout files should anyone need them...
I see making a project in one box - millivoltmeter and the 100mA constant current source PCB from miliohm meter project.
(Scullcom Hobby Electronics #31 - Design & Build a Milliohm Meter)
Trough might be useful to make a current source that goes to higher voltage, probably 10V
I always wandered why the value of a reference voltage is for example 1.024 while if you use a 10 bit ad converter , the max value is 1.023 so there will always be a miscalculation.
Can this meter read microvolts to millivolts, in other words, where the voltage coming out of a system is micro/milli volt range, can this meter read it? Thanks
it's a true milliVoltmeter i'm agree with that (and it's possible to obtain 0.1mV resolution with a reference voltage Oven temperature stabilize )
i love your channel
Make Video
building a linear low noise voltage regulators that would be helpful to many people
+David Az Thanks David for your comments. I have many ideas for future videos. It just finding the time to do them!!
I have been looking for a DIY voltage meter project for some time, Millivolt meter is what I'm looking for. Can the Millivolt meter be used to measure differential voltage across components in a circuit? Or does it need to function like an oscilloscope? Thank You Austin
Hi, anyone could tell me why 0.1uF capacitors along with other capacitors?
Great video and project as usual.
I tried something similar with a ADR4540 ref, AD7787 ADC, and LT1639 op-amp and I couldn't get it anywhere near as stable as I would have liked (obvious perturbations at 4-5th digit). At the time I put it down to issues with the layout and decoupling but I couldn't really solve it so was abandoned.
It would be fantastic if in subsequent parts you could mention any specific considerations you took to minimise the noise. My experience suggests there is some nuance involved in getting down to 6.5 digits! Cheers.
+Leigh Robinson Thanks Leigh for your comments. I will be looking at and discuss minimising noise.
Hi. Good project. I want to try your schematic. But LTC2400 is not available in Stock. Could you suggest equivalent one ?
I'd be happy to watch ads for this kind of content
As a student in Music Technology this is going to help me a lot, since there's only one fluke available in the lab.
+Eduardo Caetano The Fluke meter I would image will be more accurate. But we will check the final project against other meters.
Thank you for your great work and especially for detailed explanations. Do you have a plan to add precision detector (AC-DC converter) so that this millivolt meter could be used for the measurement of sound signals level or other AC measurements?
Thanks. There are already 4 parts to this project. I may come back to it with further updates later. With regards measurement of sound signal levels I may cover this in a new project when I have more time. Regards, Louis
Good job sir.
+Fady Kazzazi Thank you.
GBP 10,-- is a very good price for that Caddock part. You can get 1/10 dividers from Vishay, or even single S102 resistors which will be more expensive than the Caddock part. Also single resistors won't provide as good tempco tracking. I think you did a good choice on that. Please do not use these 1% standard resistors, they do not provide you with a good enough temperature coefficient for that nice 24 bit ADC. I think the OP777 is rather a bad choice and that would work better with an OPA(2)333 or LT2051 or maybe that MAX that EEVBlog uses in the µCurrent. I am a bit concerned about the Z-Diode since they tend to draw a rather high non linear reverse leakage current and I think that some BAV199 to ground and +5V would do a better job.
+Chip Guy Vids Many thanks for your comments. Yes the Caddock resistor network is the best option but I also wanted the hobbyist to have a cheaper option as some times cost is an issue, of course this is at the expense of accuracy, I agree. Your point on the OP Amp is a good one. I am looking at a number of other options for this such as the one you mentioned LT2051 as well as the LT1028 and a number of others from different manufacturers. Also I will have a look at some Chopper Stabilised OP Amps. I take your point about the zener, but wanted some simple input protection - I will look at some other options. Looks like I will be doing a Part 3 to this project!
+Scullcom Hobby Electronics Nice, looking forward to part 2 and 3. Why I was commenting so elaborate: I did pretty much the same circuit using an LTC2411-1 ADC. The circuit I made is able to measure both, positive and negative voltage but the disadvantage is that my ground reference is actually the 1.25V VREF itself. The circuit runs on 3.3V. It didn't make any difference to me since the entire rest of my circuit was floating anyway. Mine uses the OPA2333 dual opamp where the 2nd OPA is used as a precision current exciation source for PT100 or other sensors.
Where is the parts layout ? The schematic is posted. Great project and professional delivery!!
+Phil Spargo You will find a link to the parts layout in Part 2 of this project. The link is also below:
www.scullcom.com/Millivolt_Meter_layout.pdf
i kind of like TI until you start using it for hobby projects, Mmm 🤔 we have Analog Devices PCB stuff running for 40 years 🥳nice 🥰thanks this is a real hobby project 🤩🌴🐬
this would be an excellent nuts and volts project... 🍒🥂🥧☕
Another great project. Also same thought as 'Tign Meg' about software smoothing (averaging)?
Cheers,
Mark
*******************
+Mark Beeunas Thanks Mark. Will add some averaging options in software.
Very nice - idea for software, work out which digits are stable and only display them.
+David Pilling Thanks for your comments. Reducing the number of decimal places on the Millivolt range to say 4 places would work fine. This can be done in software. I will discuss this in Part 2 of this project.
The voltage divider causes inaccuracy for low voltages measurement below a threshold (say 2.048V) so it would be nice to have a relay to switch out the voltage divider when the measured voltage less than the threshold.
THANKS
You could add a relay but I have found mine to work fine. I am currently working on an update to this project with improved calibration method.
long time no see， hello there ！mister， fantastic stuff
+Bell A Thank you. Been busy just lately with other work!
Thanks
Nice project. What is the point of those two pull-ups on SPI data and clock lines between MCU and ADC? Do we really need them?
+michalvanco Thanks for your comment. You are correct strictly speaking we do not need any pull-up resistors on the SPI data and clock lines. Some people seem to use pull-up resistors to ensure that random data isn't clocked into
the ADC while the main processor is booting up and its pins may be in an
uncertain state. Pull-up resistors are only normally only needed for I²C protocol. However, there may be a point in adding a pull-up resistor on the CS (Chip Select) pin to ensure that no corrupt data gets to the ADC at start up during initial switch on.
Yet another very well put together video. I admire your patience, it must take an extraordinary amount of prep time to do these.
On the Project. Do you intend to or have you any software smoothing (averaging)?
+Tign Meg Thanks for your comments. I am working on the software for part 2 and have tested software both with and without averaging. I will discuss this in part 2 of this project.
The Max6341 seemed the best to me.
Снимаю шляпу перед Вами.
Why not use a higher cap value like 10uf on the output of the voltage reference since its output is just a fixed DC voltage and you want to clamp down as much as possible on any noise?
+Kevy Kevy, thanks for your comment. You could use a 10μF tantalum instead of the 3.3μF I used, its up to you, both are OK. The Analog Devices datasheet for the ADR4550 states; "A 1 μF to 10 μF electrolytic or ceramic capacitor can be connected to the input to improve transient response in applications where the supply voltage may fluctuate. An additional 0.1 μF ceramic capacitor should be connected in parallel to reduce supply noise".
How may of these digits are usable and trustworthy?
It looks like you have made a 4 digit meter...
Great stuff. You should modify an alarm clock to be like the one from the movie 1408.. could really scare some folks
+Guy B Thanks. Interesting comment !!!!
Scullcom Hobby Electronics are u gonna do it or not
+Guy B I have not seen the film but believe it is a psychological horror movie which, sorry I am not in to. So I have no plans to build the alarm clock used in the film.
Ooow you have been busy :-)
i really like the independant modular adc, professional equipment use the same idea :-D.
Also the method of display can be changed as well, so you havnt restricted the design to only one type of microcontroller :-D
Slowing the rate of readings down may make it easyer to look at, if the noise cant be removed i would lower the resolution when running in Mv mode.
I thought the typical 9meg resistor and 1meg grounded resistor would be quite typical, but ill have to wait untill the end, suspense :-D :-D
+zx8401ztv Thanks for your comments. I have looked at the sampling rate and been testing several options to see which is best. I am still playing with the input divider network to try and keep the cost down for the hobbyist. The Caddock (type 1776-C681) precision divider network I showed in the video are expensive, but work well. The cheapest one I could find was from Mouser Electronics at £10.66. The schematic that you can download from the link in my comments uses the Caddock resistor network but I am looking at some other options. The next step is to finish the software for Part 2.
+Scullcom Hobby Electronics
Hmm expensive.
The noise, if you grounded the input, then bridged a 100 ohm resistor across R1 for a second, does the noise vanish?
Does the value of R1 need lowering?
+zx8401ztv Thanks for your suggestion. Once I get the software finished, so I can upload Part 2 of this project, I will have a closer look at lowering any noise issues. I will look in to the value of R1 and see what the effect is. Looks like I will be doing a Part 3 to this project covering improvements and upgrades !!
Hi, Thanks for the Video
Thank you.
It is very useful and I always love to home-brew. Thanks once again
Excellent approach, but for milli accuracy wouldn't 16bit resolution be enough down to 62.5 uV and some ICs have on board voltage reference (0.05% accuracy) like MCP3426 as well as very affordable price £2.50?
THANKS
I have used both a 12bit and 16bit ADC in the past but find the 24bit option gives better resolution overall.
I agree
Lose the 5V1 zener at the input, the leakage will do bad things to your accuracy as it is both voltage and temperature sensitive. Protection use a reverse biased transistor CB junction to the zener, fed from Vcc with a 100R resistor. Another reverse biased CB junction to ground and you will protect the opamp input without any major leakage, and the BJT's ( chose for low Vcb leakage from datasheet, they will be better than any regular diode and loads better than a soft zener junction) are cheap. Decouple the reference supply as well, using either another 5V regulator to feed it alone, or a RC decoupling on the power line.
+SeanBZA Thanks for your comments. I will look at some alternative options for input protection.
Probably I'm missing something, but the OP777 is not a rail-to-rail OpAmp. I don't see a negative supply voltage, so I don't understand how you accurately measure voltage near 0V.
+Hendrik Lipka The OP777 can also be used on a single supply voltage from +3 volt to +30 volt. If you check the Datasheet it covers this form of use.
I knew that already, but my question was about the output voltage swing (where the electrical characteristics show a swing of about 100mv down to the negative rail). And the specification for single supply voltage is with Vcm=2.5V, so its not R2R there too.
But later on in the data sheet there is an explanation that, with light loads, in a buffer configuration, the output can go down as low as 1mV (page 11). Maybe this explains why you see several millivolts readings with shorted inputs...
+Hendrik Lipka Thanks. You have answered your own question. The datasheet covers this and as the project was designed specifically as a millivolt meter it should be OK. I am looking at other precision OP Amps with better performance and will cover that in Part 3 of this project.
very nice; 6.5 digits stability is going to be hard to achieve without an LM399 and low tempco resistors
Thanks for your comment. Adding a LM399 is possible but you would also need some additional circuitry as well to set the initial reference voltage.
Binge watching your videos today :)
If your reference is 4.096V, how can you measure to 5V? Is it not 4.096V maximum? Maybe I'm missing something.
From the ADC datasheet: The converter accepts any external reference voltage from0.1V to VCC. With its extended input conversion range of-12.5% VREF to 112.5% VREF, the LTC2400 smoothly resolves the offset and over range problems of preceding sensors or signal conditioning circuits.
A voltage reference for an ADC is merely a yard stick for which it compares samples to. The farther you get at either extreme, the larger the error.
Same supply for analog and digital circuits...?
+horse1066 Thanks for your comment. Using the same supply should not be a problem with this project. It is very low current, therefore it is unlikely to have issues due to digital switching spikes on the power lines. In any case I have ensured there is plenty of decoupling for the I.C.'s. in this project.
Thanks. Better later than never hehe.
About that resistor divider network being costly, I don't think that changing it with low tempco resistors that age well will make it less expensive. Divider at the front end is the part that we sadly can't cheese easily with low cost components.
+kroplaaaa Thanks for your comment to which I agree.
Lets think where the errors comsome comments one from: op777 100uV offset, makes 1mV. The resitors may not be precision, and may have temperatur drift. I recommend using thin film resistors. Also the zener has leakage, i recommend removing it, normally have internal protection.
+over2there Thanks for your comments. There are a number of options for the input OP Amp buffer. You could replace it was a LT1028 which has a very low input offset voltage (but also more expensive). The OPA277 from Texas is another option. You could even use a Chopper Stabalised OP Amp which correct for input offset voltage. I do show on the video a Caddock precision resistor divider network as an option for the input. The other resistors in the circuit are not critical and do not effect the result. Calibration in software will also help with any fixed errors.
got some tips on the way, thanks
+Ronen Shemesh Thanks.
ok well , yes there is a real engineering with a big balls
+Deni Komici Thanks, interesting comment !!!!
wow. I'm so sad that no new content has been posted in the last 2 years...
I'm sorry, but you're not getting 24 bits out of the OP777.
Not even close.
This is a lovely project and extremely well presented, and I have previously enjoyed other videos on this channel. It is possible I have misunderstood or not seen the whole story so far, but I'm afraid I don't think this is fair to call this a 6 1/2 digit project. In your component selection section you show voltage reference errors of around 3ppm in several parameters, your adc is quoted to have a 4ppm error (though it may be some calibration here can help) and the noise error in your buffer or resistor network is not discussed.
If a value has a 4ppm error then a true reading of 999,995 could be in reality 999,991 to 999,999, that is to say only the first 5 digits have any useful information, the final digit is random. It seems to me you may get a 6 digit reading but you only have 5 digit accuracy. If the multiple part noise errors combined push above 10ppm you start dropping from 5 digits to 4 digits. Your own voltage reference tests show 4 digits of stability in your readings at the 5V reference test and less at the 50mV test. Of course you can calibate out absolute error but you cannot calibrate out noise. Your Arduino can print a floating point number to as many digits as you choose to use in the floating point arithmetic in software but if the lower digits are random you may as well not display them. Your own tests suggest this is a 4.5 digit meter at best.
I'm so sorry if I am being a wet blanket as I admire your skill and teaching, and if you or someone else here shows my mistake I will be pleased. With kind regards.
+Chris Baker I was having the same questioning, but not daring to express it to such nice man.
It seems to me that digit number is a short hand for relative accuracy, and that it becomes unacceptable to let trailing digit float to whatever value they want. I think by calibration you can get all the digit to a precise value, but you need a transfert standard that is of higher accuracy than the equipment under calibration, and thermal design becomes an issue when such a high relative precision is sought.
+Chris Baker Thanks Chris for your detailed comments all of which are valid. In an ideal world the 24 bit ADC has a total of 16,777,216 possibilities and as we are using a 4.096 volt reference then each bit represents about 0.2 µV (0.00000024V). In reality you have to take in to account all the errors and any noise so the true result will be less accurate. Some of the errors can be reduced by calibration and others by taking steps to reduce as much of the noise as possible. I have found in the past that most 24 bit ADC give good results to 10 bit accuracy and in this case each bit would represent 4 µV. I will compare the results of this project against my calibrated Keithley 2000 bench meter to see what the final result is. Your comments are always welcome as it makes one think of things they may have missed or improvements that can be made. Thank you.
+Chris Baker Digit number in multimeters does not refer to accuracy it refers to the number of digits that the display can show. Supposedly those digits are noise-free but not necessarily accurate. The most accurate meter I know of (3457A + opt 2) is 8 1/2 digits but has an accuracy of 4ppm over 1 year which means it's only 5 digit accurate. Your standard industrial grade 6 1/2 multimeters are +- 50ppm -ish accurate which means 4 digits of accuracy.
nice
Thanks.
If you can accurately measure 5.1415321 mV that's not 6 1/2 digits, that's a full 7 digiter you got there. I am waiting to see what device you will be using to both calibrate that and check its calibration.
+SteelBlueVision Thanks for your comments. The 24 bit ADC does allow some great resolution to 6 decimal places or more. Calibration is always a difficult part and depends on many things.
+Scullcom Hobby Electronics Which also begs the question: Why did you choose to use a 4.096 V voltage reference that does not support trim? How will you be able to trim it to 4.096000 V, since the accuracy of the ADC depends on the accuracy of the voltage reference and the voltage reference is +/- 0.02% (i.e., 200 ppm). This is not nearly accurate enough for the level of resolution you are shooting for unless you are hand picking among hundreds of regulators to find a "perfect" one. Sure long term drift is very important, but if your initial voltage is off by +/- 0.02% (or even +/- 0.01%), drift will be the least of your problems in achieving an accurate output from the ADC.
Please don't take any of this negatively, your UA-cam channel is my favorite among the myriad electronics channels out there.
+SteelBlueVision Thanks for your additional comments. I selected a fixed voltage reference IC in order to make it easier for the hobbyist. Although the initial accuracy is ±0.02% once the reference voltage is stable its long term drift is only 25 ppm. Both the temperature and line regulation are very stable. If you did use a voltage reference with trim facility, although you could set the initial accuracy the fact that you have added additional resistors and a preset resistor they may in themselves cause problems as a preset resistor normally have a ±10% tolerance, which could give an error up to 100,000 ppm (even if you bought a costly ±5% tolerance preset at a cost of about £25 it sill would be 50,000 ppm) and 0.1% resistors are 1000 ppm. The other factor I considered in this project is that the hobbyist does have access to costly professional calibration equipment. These are some of the consideration I took when designing this project with the hobbyist in mind. Thanks for your additional feedback which I do not take negatively as it reminds me to explain some of my thought process which I may have forgot to mention in my video. Thanks.
Scullcom Hobby Electronics Thank you for replying.
I certainly agree with the vast majority of the concerns you addressed. However, I would like to address several items you mentioned:
I primarily use resistors that have a
+SteelBlueVision Thanks for the additional comments. There are many options to upgrade this project to improve performance and accuracy. My initial aim was for the hobbyist on low budgets but I will probably do a part 3 to this project covering possible upgrade options.
"LCD Display" --> Liquid Crystal Display Display...
I winced every time you said that.
Nothing wrong with that. It's a kind of display and the prefix says what kind. You can have LED display, LCD display, flip dot display and so on.
@@jan.tichavsky Nothing wrong with that? I don't think you get it...