I hope this old electronics gentleman, who explains everything with such accuracy, diligence, and thoroughness, will remain with us for a long time! I, for one, have learned a great deal from him.
I'm skipping a couple of past projects for the moment to get to this one, I'm having to build a much more compact workshop, where I'm building almost all home made equipment,because I enjoy the building, I've wanted one of these for sometime, thank you Louis you certainly have your finger on the pulse,I love these projects featuring TFT, screens, and the fact your using normal size parts, the nano and tft are quickly becoming the hobbyists and radio amateurs friend. 73 Paul M0BSW
Hi Paul, Looks like you are going to be busy. This is the first time I have done a project using a TFT touchscreen so it is also a learning process for me to. All the best, regards Louis
~Hello Louis, I like to keep the grey matter busy,, plus I'm learning something everyday,TFT & Oled are very interesting devices although Oled screens of a decent usable size, are still a bit expensive in my view.
I think your projects are the better ones to build, they seem to be very accurate and well designed. I had no idea all this stuff was available, I graduated tech school in 2000 and kind of dropped out of the scene for a while. No I wasn't in prison, worse, I lived with my ex. AKA wicked witch of the west
Thanks again for another awesome opportunity to learn electronics. Opposed to all the other channels trying to explain how the components work, your channel shows what you can with it. This triggers me to dive deeper in to the nitty gritty of electronics. Thanks again
When calibrating a source, I was taught about using the first digit and decimal in lower digit numbered meters to gain a digit at the end. For example, when setting a 10 volt standard, you would adjust it to 9.9999 on a 4 1/2 digit DMM, so the digits shift between 9.9999 and 10.000, which will get you closer to a reading of a 5 1/2 digit DMM at 9.99999. This would apply to a 5 1/2 to 6 1/2 digit DMM as well, and even a 7 1/2 digit. Essentially, you're fishing for that last unseen nine and splitting hairs.
WOW Louis, that was a long recess from Scullcom School of Electronics. But, family does come first. I think I can speak for all that we missed you so much. Yes we missed the projects and the videos---but the real truth is we missed YOU. It is so good to see you again. We love you buddy. Have a Merry Christmas and a Happy New Year. I am not sure how you come up with these ideas for projects but I like your way of thinking. Take care Louis. God bless you.
Hi Louis, very nice project and good to see you back. its obvious to see the work that has gone into the project..............thank you.................................................Berni
Thank goodness you have this info posted I have been racking my brain for days trying to build schematic for same project. Biggest difference in mine is substituting adc to 22bit mcp3551 which may be overkill but fun to think of resolution in 6 digit ranges.
Hi Tommy, Glad to hear you found it helpful. You mentioned the MCP3551 which is an ADC, but for this project you would need to use a DAC. Regards, Louis
Scullcom Hobby Electronics my mistake, I'm trying to build more of a self calibrating type bench multimeter/power supply to clear up any confusion. It's a rabbit hole for sure. It seems that your circuit would make it easy to make an add-on board to make this work especially since you have 2 voltage references to compare against and extra unused pins. I will try to get some simulation data when I get a chance to try this. Thanks.
Thanks Nick. I do spend a fare amount of time in preparing for my projects so it can take longer than doing reviews. I also spend time on editing my videos down to a maximum of about 45 minutes which also takes longer, especially deciding what to leave in and what to take out. Anyway just hope I get the balance right. Regards, Louis.
Excellent! Your projects just keep getting better. Thank-you for the time and effort you put into your videos, it must be heart-breaking to have to eventually wipe off those white-board diagrams....they are very much appreciated.
Nice project, board looks really great. One thing to consider is the ~30 ohm Drain-Source ON resistance of the DG303, it's a little high to use in the output circuit without upset to the calibration for differing load conditions without feedback compensation. One also has to watch the switch's leakage current as well. The switches are most likely ok in in the gain stage however I didn't check the gain sensitivity for the ON resistance vs the 25K internal op-amp resistors. Consider running the entire DAC/Op-amp chain for a fixed 10V output then use the analog switch to make a .5X ref voltage for the DAC for the low voltage range. Any switch error can be calibrated out with an adjustment resistor on the the reference and the reference .5 divider. Thanks for the video, I need to make one of these some day.
Thanks Graig for your comments. With regards the 30 ohm on resistance of the analog switch at the output. This should only be an issue once you start drawing say the higher or maximum current of this unit at about 30mA. But if the unit is used for checking the calibration of test equipment than I think it should not cause any real problems. That said, we could improve the accuracy at the higher currents by replacing the DG303B (U6) with a MAX4622 (but twice as costly) which has an ON resistance of only 3 ohm, but I would also have to alter the print layout on the PCB to match the changes in IC pin connections (maybe an option for another version of this project). Another option could be to use the existing analog output switch to control a small reed relay. Your second point is also interesting but the current DAC (MCP4922) I am using has a maximum supply voltage is only 5.5V. Regards, Louis
Very nice project, you could use the 2nd dac to trim the op amp divider voltage to give smaller than mV steps.... I think you could take it down to uV levels. I did a video of my repairs to a Fluke 343A DC calibrator on my channel earlier in the year, and am in the process of a video series repairing a Fluke 5200A AC calibrator, some might find it interesting.
Great work man! I respect the time and effort you put in minute details like detailed schematic drawn on white board. I loved it all. I am also working on a similar project and would like to show you once completed. Thanks a lot, keep it up.
Sharing your expert knowledge and encouraging the hobby engineer is wonderful. Without doubt the most educational and matter of fact electronics design engineer on here.
Very nice project. I like the touch screen with the audrino. I haven't got into using them. I have just been sticking to repairs. Might have to build some projects and utilize a touch screen interface. I liked your electronic load also. 👍
At the moment, I'm working hard on your DCLoad project and some modifications of mine. And now, you push the next out - I'm on hurry, on christmas time I don't leave the lab...;-) Nice and many thx!
thanks for sharing this great project! perhaps just a tiny simplification - if the touch controller is a xpt2046, no level shifting is needed at least for the touch connections. the xpt is rated for 2.7-5v according to data sheet...
Welcome back with another outstanding presentation and explanation Louis. What is the bottleneck with display refresh? the SPI bus or the AT Mega 328? Perhaps one of the STM 32 F1 "Blue Pill" kind of sort of Arduino compatible boards might be faster. Dirt cheap on Ebay
Hi Rob, Thanks, The problem is the screen refresh rate for the TFT. I am currently using the Adafruit library which may not be the fastest. I believe someone may have modified this library for a faster screen refresh rate, so I will have a look for it. I have tried in the software to only write to the TFT if something changes, this helps. Hopefully this is something that we can improve as the project progresses. Regards, Louis
Another superb project by the master... One must be able to adjust the 4096 volt-ref with a meter like yours. Otherwise, it might not have the accuracy of 1mV, but it will surely have 1mV resolution. That'd reason alone to send it off for calibration, if you don't have the means to calibrate to 1mV. Figure-1 on page 3 of the DG303B shows how clever the switch is! I wonder how much noise these other ICs introduce into the static generated voltage by the voltage reference IC? Would one be able to connect a 1:1 push-pull audio amp to use as an output driver for a programmable power supply capable of 1Amp output?
Thanks for the comments. I will look at some easy ways for the hobbyist to calibrate this unit in Part 2. I will also be looking at overall performance and noise. We could use some of the techniques I use in this project with a programmable power supply, may be another project later. Regards, Louis
The circuit is flexible and accurate enough to have both in one! With the right output driver that won't load the ref part, but could deliver 0-12V at 1Amp. Maybe a JFET OP amp as a voltage follower + a current driver. Or even use the second DAC. One thing though, what will happen in a case of a sweep say from 2V to 10V? Since it switches between 2 ranges, will it not cut power (brake before make) for a few uS, where the interruption is small enough that the inline caps would pickup the slack?
You could add an LT3080 to drive larger loads, up to 1.1A and the voltage can be set by applying the right output voltage level directly to the control pin
Fantastic video once again. Your videos really help to make circuit design more accessible for beginners like me. I did notice that your tft refresh rate is very low. Have you considered trying the STM32F103C8T6 boards? The difference in speed is quite big, there is a comparison video between the stm32 and the arduino nano using a ILI9341 display, it really is night and day. Not to mention they can be bought for as little as $1.72 and have some other upgraded features as well.
Thanks Ronald for your comment. Thanks for bringing my attention to the STM32 board. I have not used this type of board yet but it looks very interesting and as you say the cost is very low. There must be a way of adding that type of board to the Arduino IDE software which would make life easy when programming. Regards, Louis
Thanks. The Arduino does not handle over 5V it only operates from 5V or 3.3V supply. We use the software code programmed in the Arduino to do the calculations and output a digital representation of that information on the SPI bus to feed both the TFT display and the 12 Bit DAC. It is the DAC that generates the actual voltage output in our case up to 4.096V and it is the other two OP Amps (their gain controlled by the Arduino and the CMOS anaolg switch IC) which use a higher supply voltage of ±15V to generate an output voltage up to 10V. Regards, Louis
One very easy thing you can do is link both DAC channels with, say, a 1000:1 divider on one of them giving it the ability to "nudge" the output a little way. Between them they then act as a coarse/fine adjustment pair, greatly increasing the set resolution and possibly allowing you a digit or two more on your setting, if you calibrate it well enough
Thanks Paul for the comment. I did have in mind something similar but I am also looking at some other uses for the second DAC. I will try and cover it in Part 2 of this project. Regards, Louis
Very nice video. In Part 2 it would be interesting to see what happens at the switching borders (4.000V / 4.001V and 8.000V / 8.001V ). It might turn out that 4.000V will measure more than 4.001V.
The MAX6(2/3)41 are actually a trade-off. The temperature coefficient isn't long term, reference drift is. the MAX63NN has roughly half the typical temperature coefficient - 0.5ppm, of the MAX62NN's - 1ppm typical. However the MAX62NN lists its long term drift spec at only 20ppm/1000 hours, beating the '63 at 30ppm/1000hours; I believe this is where their numbering scheme comes from but I could be wrong. So it's basically a trade off if you want the tightest short-term tempco, or the lowest possible long-term drift.
Nice to see new videos. I would power the dcdc from 18v to avoid polluting the 15v with some switching noise. Also do the 2x first and the 1x next for 4V to 8v. Also why stopping at 4V? Why don't you go to 4.095v before the x2?
Thanks for your comment. Yes I agree about feeding the negative DC to DC converter IC direct to the 18V input supply. I did hint at that in my video and will have a look at that in part 2. With regards the 2x first and 1x next for the 4V to 8V range, I did try that originally in my testing with the software but found for some reason it worked better the other way - I will have a further look at that. Regards, Louis
What do you use a DC voltage calibrator for? BTW like the fact you give the component list, PCB etch and everything right here, one stop, no need for a gerber file reader and all that stuff. I used to etch things like circuit boards and other things using ferric chloride at a company. EPA regs made it difficult to stay in business.
In the design spec you said you want 1mV step. But (if I understand correctly) in the 8-10V range, your resolution will reduced to 4mV because of the x4 gain form INA105. For example if I set the output to be 9.001V, arduino will tell the DAC to output 9.001/4 = 2.25025V will is not possible for a 12bit DAC. If I really want 1mV resolution, should I use 16/24bit DAC instead?
The resolution of the 12 Bit DAC can only go down to 1mV in the current design and so in the example you gave of 9.001V which is then divided by 4, the output from the DAC will be only good to 3 decimal places so it will output 2.250V, the INA105's will then multiply that by a gain of 4 giving an output of 9.000V. If you want more resolution from the DAC then a 16 Bit DAC would be a good option but the cost of a suitable 16 DAC would increase 5 times or more. A 16 Bit DAC would increase the resolution to a few uV. I will be looking at some update options in part 2 of this project to increase the resolution/accuracy. Regards, Louis
Hello Louis. Coming late to your nice project... I've a simple question: am I the only one to notice that the schematic around the switched OPamp doesn't fully reflect the bloc diagram? According to your long explanation about the choice of the multiplication factor 1x 2x 4x, these are cascaded by selecting x1 or x2 on both stages. But the schematic shows that the second switch actually select the output of the first OPA (x1 or x2) or the output of the fixed x2 gain of the second... Is there any advantage to do so instead of replicating the 1x, x2 gain selection of the first OPA to the second OPA?
Great video! Thank you very much! I would have gone with some things different but that would be more for effect :-P For example, I would have gone with a flyback for example for the -/+15V rails, can be implemented with something as easy to get as MC34063 maybe another more modern IC but still with trafo, it can take a wider input range, make it easy for battery conversion. Would have gone also probably for another implementation to get also negative voltages, cause why not. Could also have gone with a MOSFET reverse input protection and combine it with a soft power button. Still, the idea is pretty cool! Thanks for that!
Hi, Thanks for your comments. It is interesting you mentioned the MC34063 as I am using that IC on a new Function Generator project I am currently working on as my next You Tube video project. Regards, Louis
Hi Louis, Thank for the video and good work. Please keep it up as it is very educational with well explanation and right description too. However, I think you may eliminate using the gain switches by using the internal gain of 2 by the DAC itself, by using the 5V VREF and setting the INA105 with the gain of x2. It means, if the user inputs a maximum reference voltage of 10 to set, then it can be divided by 2 inside the Nano to drive DAC to generate 5V output on A and then configuring precision Diff OpAmp to gain of 2 (gounding the pin2) to get the output of 10V max. What do you suggest please? Regards, Pradip
Hi Pradip, Thanks for your comments. I did look at the x2 internal gain of the DAC ouput amplifier but that has a number of limitations. Since the maximum supply voltage for the DAC is only around 5 volts, that limits the maximum output voltage anyway to about that level. So it will only be of use if you were to use a lower voltage reference. Which would be an option. Another point is the internal OP amp of the DAC with x2 is not as accurate as an INA105 set to a gain of 2, so you may get more errors. Having said that, your idea is of interest if you did use a 2.048V reference as each step of the DAC will give exactly 0.5mV and so improve the overall resolution. Its worth giving it a try. Thanks, Louis
what a great project. lots of expensive components tho. Just thinking if you get a single more expensive component - DAC, say 18bit, and less precise other components, analog switches already introduce a problem anyway. Then use software to calibrate away the errors. Major drawback, you need access to something to calibrate against, $$
Why does your Fluke meter continue to give readings after you remove the probe from the circuit? Is this a peak hold feature? It's weird because the value keeps changing after you remove your probes.
This is due to the fact that the Fluke 8842A has an input impedance of around 10 GigOhm on voltage ranges below 20 volt. When it auto-ranges above 20 volts then the input impedance of the meter drops to about 10 MegOhm (similar to a standard multimeter) and then you don't see this effect. The residual voltage you see remaining on the display is what is held by circuit capacitance. You can zero the reading if you wish by simply shorting the test leads, if the test leads are simply left floating then the reading will take a long time to drop to zero and that the changing reading you see on the display is the effect of that. This is one of the effects of having a multimeter with a very high input impedance. But I am sure you will understand the benefits of having a very high impedance at lower voltage readings. If you look at 40:03 of my video you will see at that point I short the meter leads together. Regards, Louis
Awesome project! I am curious to understand if instead of using the first Op Amp as a X1 and the second as an X2 in the case of the 4V output range provides a more stable output than if you have the gain switched the other way around, X2 for the first Op Amp and the second as X1. I say this because any noise introduced by the first, however little, is amplified by the second Op Amp. Verification only requires a minor software change. I wouldn't provide this comment if we weren't interested in a precision and stable output. Meanwhile, I'm curious as to how to employ the second DAC... :)) Great work, Cheers!
Thanks for your comments. With regards the 2x first and 1x next for the 4V to 8V range, I did try that originally in my testing with the software but found for some reason it worked better the other way - I will have a further look at that. In part 2 of this project I will be looking at its performance including noise and take on board what you say. I have a number of ideas for the second DAC but have not finalised that yet, hence the reason why I have provided the extra socket connection options on the schematic and PCB. Regards, Louis
wow, what would be cool is to place the 6.5 digit multimeter and the calibrator both in the same instrument enclosure 🤔 like the Ronan X86 with 2 input & 2 output leads 🤩thanks 💒
Thanks. Sorry about the follow-up to the function generator on XR 2201. I think I got distracted and moved on to other projects. I will have to dig it out again and have a look. Anyway I may look at a further project on Pulse Generator which may be of some help later. Regards, Louis
Amazing project! But I did not unterstand why the circuit has a precision of 0.01V up to 10V. If you must divide by 4 (to get 10V) before using the DAC and the DAC has a resolution of 1mV and later amplify by 4 - would the result not have plus/minus 4mV? You are losing 2 Bits or not?
Thanks that's a very good question. I will try and explain. If for example we select a reference voltage of say 8.444V then before we feed that information to the 12 Bit DAC we divide it by 4 in software so the DAC sees a request to provide 2.111V. The output from the DAC is accurate to 1mV (due to the milli volt step resolution). The output from the DAC is now an analog voltage of 2.111V. The first precision OP Amp takes that analog voltage of 2.111V and amplifies it by 2 giving 4.222V it then passes to the second OP Amp which again amplifies if precisely by 2 to gave an output voltage of 8.444V (remember the Precision OP AMP's work on analog voltages while the DAC works on digital 1mV steps). Any error on the output voltage will depend on the accuracy of the OP Amps. That's the reason I am using the INA105 which has a maximum gain error of only 0.01% when operated at a gain of x2 and only 0.001%gain error for a gain of x1. So in my example here the output of the first OP AMP may be out by a maximum of 0.01% meaning the output reading may be as high as 4.2221V (only out by 0.1mV) this is now fed to the second OP Amp which may have the same error of 0.01% so in this case the output will now be around 8.445V. Only 1mV error and that is the maximum, in reality it would be better than that. The main cause of any error in the output voltage is more likely to do with the DAC itself as it has the following intrinsic errors: Integral Non-Linearity (INL) ±2 LSB (2 steps) error which is the maximum deviation between an actual code transition point and its corresponding ideal transition point, after offset and gain errors have been removed. Also there is the Differential Non-Linearity (DNL) of ±0.2 LSB error which is the measure of variations in code widths from the ideal code width. A DNL error of zero indicates that every code is exactly 1LSB wide. There is also an Offset Error of ±0.2% of Full Scale Reading, which is the deviation from zero voltage output when the digital input code is zero. These 3 type of errors in the DAC could result in an overall error of around a maximum of say ±5mV from the output voltage of the DAC. This is something we may be able to address in software calibration later. Have a look at the MCP4922 datasheet pages 19 and 20 for a fuller explanation. In part 2 of this project I will have a look at the accuracy of our DC Voltage Calibrator with a view to resolving some of the issues. Sorry for the long answer but your question was a good one. Hope this helps. Regards, Louis
Louis, correct me if I'm wrong, but I think there's a problem... This works with values that end in even numbers (as in your example with 8.444V) but it doesn't work with any value which is both larger than 4.000V and which ends in an odd number. For any value meeting these conditions, there will always be a 1mV error. Consider two examples: 1. You set 4.001V. The Arduino divides by 2, giving 2.0005. The 12-bit DAC can't resolve the last digit (0.5mV), and we can't know whether the output will be 2.000 or 2.001. Either way, the DAC's 0.5mV error is doubled by the op-amp, becoming a 1mV error at the output. 2. You set 8.443V The Arduino divides by 4, giving 2.11075. The 12-bit DAC can't resolve the last two digits (0.75mV), but the output will round up to 2.111. This 0.25mV error is quadrupled by the op-amps, becoming a 1mV error at the output. You need a 14-bit DAC with 0.00025V resolution to overcome these division errors.
Thanks FlyingScotsman. Yes you are correct the 12 Bit DAC does have some limitations when I start dividing the input in software and the value is an odd number. In Part 2 I will be looking at the accuracy of this project. I did consider using a 16 Bit DAC at one stage but they are much more expensive and come in a SMD format, which is what I was trying to avoid on this project. We may look at ways of improving the accuracy later. In the first instance I also wanted to try and match the Time Electronics DC Voltage Calibrator I showed which has an accuracy of around 0.02%. Thanks for your input which is appreciated as it adds to the overall understanding. Regards, Louis
Thanks John for your comment. I believe the INA105 OP Amp I use already has Output Short Circuit protection built-in, like many OP Amps these days. According to the datasheet it states that the "Output Short Circuit to Common is Continuous". Regards, Louis
I hope this old electronics gentleman, who explains everything with such accuracy, diligence, and thoroughness, will remain with us for a long time! I, for one, have learned a great deal from him.
We all miss you! Hope you'll get more time for your hobby! God luck and wish you lots of health and no more problems!
It great to have you again doing videos... I think you gave us some time to finish the past projects :)
Thanks. Its nice to get back to electronics and a new project. I need to get back in to the swing of things again. Regards, Louis
I'm skipping a couple of past projects for the moment to get to this one, I'm having to build a much more compact workshop, where I'm building almost all home made equipment,because I enjoy the building, I've wanted one of these for sometime, thank you Louis you certainly have your finger on the pulse,I love these projects featuring TFT, screens, and the fact your using normal size parts, the nano and tft are quickly becoming the hobbyists and radio amateurs friend.
73 Paul M0BSW
Hi Paul, Looks like you are going to be busy. This is the first time I have done a project using a TFT touchscreen so it is also a learning process for me to. All the best, regards Louis
~Hello Louis, I like to keep the grey matter busy,, plus I'm learning something everyday,TFT & Oled are very interesting devices although Oled screens of a decent usable size, are still a bit expensive in my view.
I think your projects are the better ones to build, they seem to be very accurate and well designed. I had no idea all this stuff was available, I graduated tech school in 2000 and kind of dropped out of the scene for a while. No I wasn't in prison, worse, I lived with my ex. AKA wicked witch of the west
Love the calibrator project, I can see this one being built by lots of folks. Well done.
Thanks Ian. Its not quite up to the standard of your unit but it may prove adequate for the hobbyist. All the best Louis
Thanks again for another awesome opportunity to learn electronics. Opposed to all the other channels trying to explain how the components work, your channel shows what you can with it. This triggers me to dive deeper in to the nitty gritty of electronics. Thanks again
Thanks for your comments. Glad to hear you find it helpful. Regards, Louis
When calibrating a source, I was taught about using the first digit and decimal in lower digit numbered meters to gain a digit at the end. For example, when setting a 10 volt standard, you would adjust it to 9.9999 on a 4 1/2 digit DMM, so the digits shift between 9.9999 and 10.000, which will get you closer to a reading of a 5 1/2 digit DMM at 9.99999. This would apply to a 5 1/2 to 6 1/2 digit DMM as well, and even a 7 1/2 digit. Essentially, you're fishing for that last unseen nine and splitting hairs.
WOW Louis, that was a long recess from Scullcom School of Electronics. But, family does come first. I think I can speak for all that we missed you so much. Yes we missed the projects and the videos---but the real truth is we missed YOU. It is so good to see you again. We love you buddy. Have a Merry Christmas and a Happy New Year. I am not sure how you come up with these ideas for projects but I like your way of thinking. Take care Louis. God bless you.
Thanks James. Very nice of you to say. Have a great Christmas :-) Regards, Louis
Glad to see you back again. Great little unit, worth the wait for this one.
Thanks nice to be back. Regards Louis
Sir, your voice was just made for explanation of projects. Thank you for explaining things, in a way that I totally understand.
Thanks Timothy for you nice comment. Regards, Louis
Welcome back! Great to see you're back in the saddle, and new projects coming along. This just made my day!
Hi William, Glad to hear it :-). Nice to be back. Regards, Louis
glad to here you call it an electronics workshop, I don't get why so many people on youtube call it a Lab
An exceptionally well presented project. Bundles of information and zero guess work required. Brilliant!
It's great to see you again. I will follow this other project. Thanks a lot to share your acknowledge
Thanks for your comment. Regards, Louis
Good to see you back. Thanks for the great projects !
Thanks for you kind comment. Regards, Louis
Good to see back in the lab Louis. This will be an outstanding project to build from start to finish. Thank you.
Thanks Mike. Hope to get Part 2 filmed soon. Regards, Louis
Woohoo! A Scullcom video in my subscriptions? Christmas has come early. Glad to see you haven't abandoned making videos.
Thanks. Have a good Christmas. Regards, Louis
Amazing engineer, amazing indeed👌
Thank you :)
Great return project! Looking forward to Part 2, including how to calibrate the calibrator and exploring future options.
Thanks. I will be working on Part 2 in the next days. Regards, Louis
Hi Louis, very nice project and good to see you back. its obvious to see the work that has gone into the project..............thank you.................................................Berni
Hi Berni, Thought I would try something different. First project with a touchscreen. We learn as we progress :) Regards, Louis
this gentleman is a genius
Nice to see you back, great project and thank you.
Amazing, this is exactly what I used looking for to expand my electronics knoweledge, a thoroughly explained design process and process
Interesting watch. Quick to part 2 as during the video I was curious what the overall accuracy and stability would be...
Good to have you back with new vids, Louis. Looking forward to the next part.
Thanks Peter. Regards, Louis
Louis- We love your videos and hope you are well. Looking forward to some future updates and wish you the very best!
Very nice project. Thanks you very much for sharing all the details with us. Looking forward for part 2.
Thanks for your comment. Regards, Louis
Glad you're back with another great little project. Something to get my teeth into over the Christmas holidays :-)
Thanks for the comment. Hope to get Part 2 of this project done before Christmas. Regards, Louis
Welcome back! Nice project!
Thanks Tony. Regards Louis
Thank goodness you have this info posted I have been racking my brain for days trying to build schematic for same project. Biggest difference in mine is substituting adc to 22bit mcp3551 which may be overkill but fun to think of resolution in 6 digit ranges.
Hi Tommy, Glad to hear you found it helpful. You mentioned the MCP3551 which is an ADC, but for this project you would need to use a DAC. Regards, Louis
Scullcom Hobby Electronics my mistake, I'm trying to build more of a self calibrating type bench multimeter/power supply to clear up any confusion. It's a rabbit hole for sure. It seems that your circuit would make it easy to make an add-on board to make this work especially since you have 2 voltage references to compare against and extra unused pins. I will try to get some simulation data when I get a chance to try this. Thanks.
OK I understand now. All the best with your project. Regards, Louis
Nice to see yet another great project. I'll take quality over quantity with your videos every time, Louis.
Thanks Nick. I do spend a fare amount of time in preparing for my projects so it can take longer than doing reviews. I also spend time on editing my videos down to a maximum of about 45 minutes which also takes longer, especially deciding what to leave in and what to take out. Anyway just hope I get the balance right. Regards, Louis.
Excellent! Your projects just keep getting better. Thank-you for the time and effort you put into your videos, it must be heart-breaking to have to eventually wipe off those white-board diagrams....they are very much appreciated.
Thanks Tom, Yes it is a shame when I have to wipe off some of my drawings. But I hope they help with the explanations :) Regards, Louis
Nice project, board looks really great. One thing to consider is the ~30 ohm Drain-Source ON resistance of the DG303, it's a little high to use in the output circuit without upset to the calibration for differing load conditions without feedback compensation. One also has to watch the switch's leakage current as well. The switches are most likely ok in in the gain stage however I didn't check the gain sensitivity for the ON resistance vs the 25K internal op-amp resistors.
Consider running the entire DAC/Op-amp chain for a fixed 10V output then use the analog switch to make a .5X ref voltage for the DAC for the low voltage range. Any switch error can be calibrated out with an adjustment resistor on the the reference and the reference .5 divider. Thanks for the video, I need to make one of these some day.
Thanks Graig for your comments. With regards the 30 ohm on resistance of the analog switch at the output. This should only be an issue once you start drawing say the higher or maximum current of this unit at about 30mA. But if the unit is used for checking the calibration of test equipment than I think it should not cause any real problems. That said, we could improve the accuracy at the higher currents by replacing the DG303B (U6) with a MAX4622 (but twice as costly) which has an ON resistance of only 3 ohm, but I would also have to alter the print layout on the PCB to match the changes in IC pin connections (maybe an option for another version of this project). Another option could be to use the existing analog output switch to control a small reed relay.
Your second point is also interesting but the current DAC (MCP4922) I am using has a maximum supply voltage is only 5.5V. Regards, Louis
Most excellent presentation of a great project, especially the schematics - jawdroppingly awesome. Thanks for sharing your knowledge.
Glad you enjoyed it. Regards, Louis
Very nice project, you could use the 2nd dac to trim the op amp divider voltage to give smaller than mV steps.... I think you could take it down to uV levels. I did a video of my repairs to a Fluke 343A DC calibrator on my channel earlier in the year, and am in the process of a video series repairing a Fluke 5200A AC calibrator, some might find it interesting.
Thanks for your comments. I am some upgrades planned for later. Thanks for the tip on your Fluke 343A DC calibrator. Regards, Louis
Great work man! I respect the time and effort you put in minute details like detailed schematic drawn on white board.
I loved it all. I am also working on a similar project and would like to show you once completed. Thanks a lot, keep it up.
Thanks for your comments, much appreciated. Regards, Louis
Very glad you’re back. Another great learning project! Good explaination also. Thumbs up!
Thank you :-)
Great work and very good description . Bravo à vous monsieur !
Thank Rene, much appreciated. Regards, Louis
Thank you verry much for all the effort put in your projects and videos. Learn a lot every time.
Just one idea, in terms of signal to noise ratio - isn' it better to do the higher gain in the first stage (2x 1x instead of 1x 2x) ?
Thanks this may be an option which could be done in software. I will have a look at noise in Part 2 of this project. Regards, Louis
I sure hope to see more great videos soon Louis. Hope you are well.
The return of the master
Thanks :-)
Glad to see you back! thanks for another great project.
Thanks Simon.
As always extremely educational and expertly explained, wonderful stuff. Keep them coming.
Thanks.
Sharing your expert knowledge and encouraging the hobby engineer is wonderful. Without doubt the most educational and matter of fact electronics design engineer on here.
Thanks for your kind comment, much appreciated. Regards, Louis
it's great to have you again doing videos. i was very excited when i saw the notification from youtube
Thanks. Managed to get back in my electronics work shed. Regards, Louis
Can’t wait for part 2 of this. Keep up the good work 👍🏼
Thanks.
Very nice project. I like the touch screen with the audrino. I haven't got into using them. I have just been sticking to repairs. Might have to build some projects and utilize a touch screen interface. I liked your electronic load also. 👍
Nice to have you back. And as usual a great video.
Thanks Christian. Regards, Louis
great to have you back
Thanks Roland, nice to be back. Regards, Louis
At the moment, I'm working hard on your DCLoad project and some modifications of mine.
And now, you push the next out - I'm on hurry, on christmas time I don't leave the lab...;-)
Nice and many thx!
Thanks Wolfgang. Looks like you are going to be busy over Christmas :-) Regards, Louis
Ohhh yeah!
Amazing design Louis...
Excellent project
Thanks.
thanks for sharing this great project! perhaps just a tiny simplification - if the touch controller is a xpt2046, no level shifting is needed at least for the touch connections. the xpt is rated for 2.7-5v according to data sheet...
Thanks. Yes the TFT uses does use the XPT2046 touch screen controller. Thanks for the information. Regards, Louis
Excellent project and presentation.
Thanks.
Welcome back with another outstanding presentation and explanation Louis. What is the bottleneck with display refresh? the SPI bus or the AT Mega 328? Perhaps one of the STM 32 F1 "Blue Pill" kind of sort of Arduino compatible boards might be faster. Dirt cheap on Ebay
Hi Rob, Thanks, The problem is the screen refresh rate for the TFT. I am currently using the Adafruit library which may not be the fastest. I believe someone may have modified this library for a faster screen refresh rate, so I will have a look for it. I have tried in the software to only write to the TFT if something changes, this helps. Hopefully this is something that we can improve as the project progresses.
Regards,
Louis
Nice project, great to see you back.
Thanks.
Great video , well done explained . Thank you for your effort in making this project.
Thanks, Your welcome. Regards, Louis
I love your videos. Learn a lot every time. Thank you!
Thanks Kostas.
Hi Louis, Thank you once again for a great project. Bruce.
Thanks Bruce.
Another superb project by the master...
One must be able to adjust the 4096 volt-ref with a meter like yours. Otherwise, it might not have the accuracy of 1mV, but it will surely have 1mV resolution. That'd reason alone to send it off for calibration, if you don't have the means to calibrate to 1mV.
Figure-1 on page 3 of the DG303B shows how clever the switch is!
I wonder how much noise these other ICs introduce into the static generated voltage by the voltage reference IC?
Would one be able to connect a 1:1 push-pull audio amp to use as an output driver for a programmable power supply capable of 1Amp output?
Thanks for the comments. I will look at some easy ways for the hobbyist to calibrate this unit in Part 2. I will also be looking at overall performance and noise.
We could use some of the techniques I use in this project with a programmable power supply, may be another project later. Regards, Louis
The circuit is flexible and accurate enough to have both in one! With the right output driver that won't load the ref part, but could deliver 0-12V at 1Amp. Maybe a JFET OP amp as a voltage follower + a current driver. Or even use the second DAC.
One thing though, what will happen in a case of a sweep say from 2V to 10V?
Since it switches between 2 ranges, will it not cut power (brake before make) for a few uS, where the interruption is small enough that the inline caps would pickup the slack?
Interesting idea. The break before make should not cause any issues as the time involved is only 50 nano seconds.
You could add an LT3080 to drive larger loads, up to 1.1A and the voltage can be set by applying the right output voltage level directly to the control pin
Awesome project. You have great projects and such good details!
Thanks. Regards, Louis
Fresh content , nice video .... Thanks for that pasion you are the man
Thanks.
Very nice project.
Fantastic video once again. Your videos really help to make circuit design more accessible for beginners like me. I did notice that your tft refresh rate is very low. Have you considered trying the STM32F103C8T6 boards? The difference in speed is quite big, there is a comparison video between the stm32 and the arduino nano using a ILI9341 display, it really is night and day. Not to mention they can be bought for as little as $1.72 and have some other upgraded features as well.
Thanks Ronald for your comment. Thanks for bringing my attention to the STM32 board. I have not used this type of board yet but it looks very interesting and as you say the cost is very low. There must be a way of adding that type of board to the Arduino IDE software which would make life easy when programming. Regards, Louis
Its nice to see you're back!
Nice to be back doing electronics. Cheers, Louis
This is so awesome ❤ Thank you so much for making this educational video...
Thank you. Regards, Louis
Thank you sir for your time and the great video.
Thanks Alexandre.
always happy to have you back with amazing projects !!
sorry about the stupid question , but how can an arduino nano handle over 5volts?
Thanks. The Arduino does not handle over 5V it only operates from 5V or 3.3V supply. We use the software code programmed in the Arduino to do the calculations and output a digital representation of that information on the SPI bus to feed both the TFT display and the 12 Bit DAC. It is the DAC that generates the actual voltage output in our case up to 4.096V and it is the other two OP Amps (their gain controlled by the Arduino and the CMOS anaolg switch IC) which use a higher supply voltage of ±15V to generate an output voltage up to 10V. Regards, Louis
Great video! You did an amazing job on everything, thank you.
Glad you enjoyed. Regards, Louis
One very easy thing you can do is link both DAC channels with, say, a 1000:1 divider on one of them giving it the ability to "nudge" the output a little way. Between them they then act as a coarse/fine adjustment pair, greatly increasing the set resolution and possibly allowing you a digit or two more on your setting, if you calibrate it well enough
Thanks Paul for the comment. I did have in mind something similar but I am also looking at some other uses for the second DAC. I will try and cover it in Part 2 of this project. Regards, Louis
can't wait for the second part ! great video ! i learned a lot , thank you !
Thanks Anjay. Hope to start work on part 2 soon. Regards, Louis
Great project, my friend!
As always your channel is 10.
I always follow your videos!
I already gave my like!
Michel Rio de Janeiro - Brazil
Thanks Michel, Glad you enjoy. Regards, Louis
Very nice video. In Part 2 it would be interesting to see what happens at the switching borders (4.000V / 4.001V and 8.000V / 8.001V ). It might turn out that 4.000V will measure more than 4.001V.
Thanks Alex for your comment. I will have a closer look at those crossover points and try and remember to show it in Part 3. Regards, Louis
The MAX6(2/3)41 are actually a trade-off. The temperature coefficient isn't long term, reference drift is. the MAX63NN has roughly half the typical temperature coefficient - 0.5ppm, of the MAX62NN's - 1ppm typical. However the MAX62NN lists its long term drift spec at only 20ppm/1000 hours, beating the '63 at 30ppm/1000hours; I believe this is where their numbering scheme comes from but I could be wrong. So it's basically a trade off if you want the tightest short-term tempco, or the lowest possible long-term drift.
Nicely drawn schematic!
Thanks.
Welcome back! Great work!
Thanks Mauro, Nice to be back in my electronics workshop working on a new project. Regards, Louis
Nice to see new videos. I would power the dcdc from 18v to avoid polluting the 15v with some switching noise. Also do the 2x first and the 1x next for 4V to 8v. Also why stopping at 4V? Why don't you go to 4.095v before the x2?
Thanks for your comment. Yes I agree about feeding the negative DC to DC converter IC direct to the 18V input supply. I did hint at that in my video and will have a look at that in part 2. With regards the 2x first and 1x next for the 4V to 8V range, I did try that originally in my testing with the software but found for some reason it worked better the other way - I will have a further look at that. Regards, Louis
thank you for another great educative video
Thanks Jaroslav much appreciated. Regards, Louis
Wonderful presentation.Thanks.
Thanks John.
What do you use a DC voltage calibrator for? BTW like the fact you give the component list, PCB etch and everything right here, one stop, no need for a gerber file reader and all that stuff. I used to etch things like circuit boards and other things using ferric chloride at a company. EPA regs made it difficult to stay in business.
A dual tracking regulator might be necessary to preserve the dc accuracy of the opamps!
Well done !!! Congratulations and thanks !!!
Thanks Manel.
In the design spec you said you want 1mV step. But (if I understand correctly) in the 8-10V range, your resolution will reduced to 4mV because of the x4 gain form INA105. For example if I set the output to be 9.001V, arduino will tell the DAC to output 9.001/4 = 2.25025V will is not possible for a 12bit DAC. If I really want 1mV resolution, should I use 16/24bit DAC instead?
The resolution of the 12 Bit DAC can only go down to 1mV in the current design and so in the example you gave of 9.001V which is then divided by 4, the output from the DAC will be only good to 3 decimal places so it will output 2.250V, the INA105's will then multiply that by a gain of 4 giving an output of 9.000V. If you want more resolution from the DAC then a 16 Bit DAC would be a good option but the cost of a suitable 16 DAC would increase 5 times or more. A 16 Bit DAC would increase the resolution to a few uV.
I will be looking at some update options in part 2 of this project to increase the resolution/accuracy. Regards, Louis
Hello Louis. Coming late to your nice project... I've a simple question: am I the only one to notice that the schematic around the switched OPamp doesn't fully reflect the bloc diagram? According to your long explanation about the choice of the multiplication factor 1x 2x 4x, these are cascaded by selecting x1 or x2 on both stages. But the schematic shows that the second switch actually select the output of the first OPA (x1 or x2) or the output of the fixed x2 gain of the second... Is there any advantage to do so instead of replicating the 1x, x2 gain selection of the first OPA to the second OPA?
24:44 I think you did a lot of work to draw this diagramm. Great Job!
Thanks. Yes it did take me some time to draw. Regards, Louis
Nice project.. I'm really impressed with the whiteboard schematics, It must have taken a bit of effort? :-)
Thanks Mikael. Regards, Louis
Given you can only multiply by 2 or 4, how do you get, say, 9.7V?
Thanks for free knowledge.
Thanks Miloš.
Great video! Thank you very much! I would have gone with some things different but that would be more for effect :-P
For example, I would have gone with a flyback for example for the -/+15V rails, can be implemented with something as easy to get as MC34063 maybe another more modern IC but still with trafo, it can take a wider input range, make it easy for battery conversion. Would have gone also probably for another implementation to get also negative voltages, cause why not. Could also have gone with a MOSFET reverse input protection and combine it with a soft power button. Still, the idea is pretty cool! Thanks for that!
Hi,
Thanks for your comments. It is interesting you mentioned the MC34063 as I am using that IC on a new Function Generator project I am currently working on as my next You Tube video project.
Regards,
Louis
Hi Louis,
Thank for the video and good work. Please keep it up as it is very educational with well explanation and right description too.
However, I think you may eliminate using the gain switches by using the internal gain of 2 by the DAC itself, by using the 5V VREF and setting the INA105 with the gain of x2. It means, if the user inputs a maximum reference voltage of 10 to set, then it can be divided by 2 inside the Nano to drive DAC to generate 5V output on A and then configuring precision Diff OpAmp to gain of 2 (gounding the pin2) to get the output of 10V max. What do you suggest please?
Regards,
Pradip
Hi Pradip, Thanks for your comments. I did look at the x2 internal gain of the DAC ouput amplifier but that has a number of limitations. Since the maximum supply voltage for the DAC is only around 5 volts, that limits the maximum output voltage anyway to about that level. So it will only be of use if you were to use a lower voltage reference. Which would be an option.
Another point is the internal OP amp of the DAC with x2 is not as accurate as an INA105 set to a gain of 2, so you may get more errors.
Having said that, your idea is of interest if you did use a 2.048V reference as each step of the DAC will give exactly 0.5mV and so improve the overall resolution. Its worth giving it a try. Thanks, Louis
Surely the 7912 is a negative 12Vdc regulator not a negative 15Vdc one?
what a great project. lots of expensive components tho. Just thinking if you get a single more expensive component - DAC, say 18bit, and less precise other components, analog switches already introduce a problem anyway. Then use software to calibrate away the errors. Major drawback, you need access to something to calibrate against, $$
Thanks Fred. Yes agree some of the parts are a little costly. In Part 3, I may make some changes to help with calibration. Regards, Louis
Why does your Fluke meter continue to give readings after you remove the probe from the circuit? Is this a peak hold feature? It's weird because the value keeps changing after you remove your probes.
This is due to the fact that the Fluke 8842A has an input impedance of around 10 GigOhm on voltage ranges below 20 volt. When it auto-ranges above 20 volts then the input impedance of the meter drops to about 10 MegOhm (similar to a standard multimeter) and then you don't see this effect. The residual voltage you see remaining on the display is what is held by circuit capacitance. You can zero the reading if you wish by simply shorting the test leads, if the test leads are simply left floating then the reading will take a long time to drop to zero and that the changing reading you see on the display is the effect of that. This is one of the effects of having a multimeter with a very high input impedance. But I am sure you will understand the benefits of having a very high impedance at lower voltage readings.
If you look at 40:03 of my video you will see at that point I short the meter leads together.
Regards, Louis
Awesome project! I am curious to understand if instead of using the first Op Amp as a X1 and the second as an X2 in the case of the 4V output range provides a more stable output than if you have the gain switched the other way around, X2 for the first Op Amp and the second as X1. I say this because any noise introduced by the first, however little, is amplified by the second Op Amp. Verification only requires a minor software change. I wouldn't provide this comment if we weren't interested in a precision and stable output. Meanwhile, I'm curious as to how to employ the second DAC... :)) Great work, Cheers!
Thanks for your comments. With regards the 2x first and 1x next for the 4V to 8V range, I did try that originally in my testing with the software but found for some reason it worked better the other way - I will have a further look at that. In part 2 of this project I will be looking at its performance including noise and take on board what you say. I have a number of ideas for the second DAC but have not finalised that yet, hence the reason why I have provided the extra socket connection options on the schematic and PCB. Regards, Louis
Louis, it may be due to an impedance mismatch into the first stage. Thanks!
Thanks, have to take a further look at it.
wow, what would be cool is to place the 6.5 digit multimeter and the calibrator both in the same instrument enclosure 🤔 like the Ronan X86 with 2 input & 2 output leads 🤩thanks 💒
Great master again perfect job
Thanks ;-)
Well done as always. But I am still waiting for next (last) part of function generator on XR 2201. Especially pulse mode. When can I expect it?
Thanks. Sorry about the follow-up to the function generator on XR 2201. I think I got distracted and moved on to other projects. I will have to dig it out again and have a look. Anyway I may look at a further project on Pulse Generator which may be of some help later. Regards, Louis
Amazing project! But I did not unterstand why the circuit has a precision of 0.01V up to 10V. If you must divide by 4 (to get 10V) before using the DAC and the DAC has a resolution of 1mV and later amplify by 4 - would the result not have plus/minus 4mV? You are losing 2 Bits or not?
Thanks that's a very good question. I will try and explain.
If for example we select a reference voltage of say 8.444V then before we feed that information to the 12 Bit DAC we divide it by 4 in software so the DAC sees a request to provide 2.111V. The output from the DAC is accurate to 1mV (due to the milli volt step resolution). The output from the DAC is now an analog voltage of 2.111V. The first precision OP Amp takes that analog voltage of 2.111V and amplifies it by 2 giving 4.222V it then passes to the second OP Amp which again amplifies if precisely by 2 to gave an output voltage of 8.444V (remember the Precision OP AMP's work on analog voltages while the DAC works on digital 1mV steps). Any error on the output voltage will depend on the accuracy of the OP Amps. That's the reason I am using the INA105 which has a maximum gain error of only 0.01% when operated at a gain of x2 and only 0.001%gain error for a gain of x1. So in my example here the output of the first OP AMP may be out by a maximum of 0.01% meaning the output reading may be as high as 4.2221V (only out by 0.1mV) this is now fed to the second OP Amp which may have the same error of 0.01% so in this case the output will now be around 8.445V. Only 1mV error and that is the maximum, in reality it would be better than that.
The main cause of any error in the output voltage is more likely to do with the DAC itself as it has the following intrinsic errors:
Integral Non-Linearity (INL) ±2 LSB (2 steps) error which is the maximum deviation between an actual code transition point and its corresponding ideal transition point, after offset and gain errors have been removed.
Also there is the Differential Non-Linearity (DNL) of ±0.2 LSB error which is the measure of variations in code widths from the ideal code width. A DNL error of zero indicates that every code is exactly 1LSB wide.
There is also an Offset Error of ±0.2% of Full Scale Reading, which is the deviation from zero voltage output when the digital input code is zero.
These 3 type of errors in the DAC could result in an overall error of around a maximum of say ±5mV from the output voltage of the DAC.
This is something we may be able to address in software calibration later.
Have a look at the MCP4922 datasheet pages 19 and 20 for a fuller explanation.
In part 2 of this project I will have a look at the accuracy of our DC Voltage Calibrator with a view to resolving some of the issues.
Sorry for the long answer but your question was a good one. Hope this helps.
Regards, Louis
That answer was absolut fantastic! Hope you are making new videos 7 times a week now! ;-)
Thanks. If only I had the time to do it - I wish :-) All the best, Louis.
Louis, correct me if I'm wrong, but I think there's a problem...
This works with values that end in even numbers (as in your example with 8.444V) but it doesn't work with any value which is both larger than 4.000V and which ends in an odd number. For any value meeting these conditions, there will always be a 1mV error. Consider two examples:
1. You set 4.001V. The Arduino divides by 2, giving 2.0005. The 12-bit DAC can't resolve the last digit (0.5mV), and we can't know whether the output will be 2.000 or 2.001. Either way, the DAC's 0.5mV error is doubled by the op-amp, becoming a 1mV error at the output.
2. You set 8.443V The Arduino divides by 4, giving 2.11075. The 12-bit DAC can't resolve the last two digits (0.75mV), but the output will round up to 2.111. This 0.25mV error is quadrupled by the op-amps, becoming a 1mV error at the output.
You need a 14-bit DAC with 0.00025V resolution to overcome these division errors.
Thanks FlyingScotsman. Yes you are correct the 12 Bit DAC does have some limitations when I start dividing the input in software and the value is an odd number. In Part 2 I will be looking at the accuracy of this project. I did consider using a 16 Bit DAC at one stage but they are much more expensive and come in a SMD format, which is what I was trying to avoid on this project. We may look at ways of improving the accuracy later. In the first instance I also wanted to try and match the Time Electronics DC Voltage Calibrator I showed which has an accuracy of around 0.02%.
Thanks for your input which is appreciated as it adds to the overall understanding.
Regards, Louis
you need to put a series resistor on the ouput to limit the current in case of a s/cregards john
Thanks John for your comment. I believe the INA105 OP Amp I use already has Output Short Circuit protection built-in, like many OP Amps these days. According to the datasheet it states that the "Output Short Circuit to Common is Continuous". Regards, Louis
Really good Explanation with those Drawings! please keep your good work up;)
Thanks.