Although this in an older video, I just wanted to say thanks. What a pleasant and educated chap, with excellent and clear explanations. Your videos are a strong asset to the global community. Well done sir. Subscribed!
I just discovered your Channel.... And electronics as well.... The way you explain the subject and the thorough process, in which you go about constructing your project is great! 100 year's from now, people will be watching.... Just thought I'd throw that out there.... Your are a great teacher.... I'm 50, but wish youtube was around 40 year's ago. 😀
I'm no electronics expert . .indeed I never studied electronics at school/college and am only just getting my head around the subject. However, you have explained this very well, and whilst I don't understand everything, I feel sufficiently confident, after watching this video, to build the reference box which will come in very useful for checking my multimeters
Very nice video. More than adequate for the hobbyist. For more serious applications, reference aging, drift, and temperature/humidity effects need to be considered. Also, you need to have friends with calibration equipment to really get those ppms :) Thanks for taking the time to post this video.
Do not really require this kind of accuracy for the repair work I do but it would improve my confidence rather than confirming one Fluke 77 against the other. I was going make my own but you have done all the hard design work so makes life easier, thanks.
I just came across this video, was thinking of getting a cheap ebay voltage reference board but this was a really well made and interesting video i think ill try and make my own now
Excellent build and explanation of circuit/construction. Thank you. I just need to get a Keithley 2000 Multimeter (or something similar) now to confirm satisfactory results :-(
Great explanation! What I do not understand is the transistor configuration. From my perspective the npn transistor is reverse connected. Emitter shall be towards ground. Please correct me if I am wrong.
I have negative experience using xl6009 xl6019 step-up modules in battery powered devices. This ic has minimum input operation voltage 5v. when input voltage getting lower, output voltage can be higher than sated before. For similar cases i using mt3608 based modules. It has 2v minimum input voltage
fully agree, no need for a dc to dc converter to bring the voltage up to 20v to have to reduce it to 15V via a regulator when you can directly use a MT3608.
Good tutorial on the voltage ref. I might make one. Very clean and well built. But there's one thing you should consider. What happen when you forget to switch off the box?
lavith.hcm Thanks. With regards your question about forgetting to switch off. You might like to check out my latest project on a 5 volt Reference where I include an auto switch-off feature, which could be added to this project. Check out link below: ruclips.net/video/sKFT9LouyTY/видео.html
Ref102C has a voltage input range from 11v to 36V, so why bother with a complicated DC to DC converter followed by a linear reg when you could just use a DC buck boost directly to 15V ??
Ah the ridiculousness of import duties - it costs less to import components on printed circuit modules than to buy the components individually! One question though, what is the power consumption of the reference and op amp? Is the complex rechargeable battery circuit really justified? How long would two 9v MN1604 ('PP3') alkaline batteries in series last for?
Nice and simple! What is the accuracy of the Keithley? Important to note that the drift will be dependant on the external trimming components when you use them.
In the circuit diagram at 9:30 am I missing something or is the pnp transistor upside down? It seems that the emitter of the pnp is connected through the led to the negative rail, and the collector through the 120 ohm resistor to the positive rail.
Thanks Michael for bring this to my attention. Looks like I made a silly error when I drew the circuit on my whiteboard. As you say the PNP transistor collector-emitter need to be reversed. When I have time I will draw an updated schematic and provide it as a download for other viewers to use. Thanks again, Louis
Regarding a single multi-turn pot: actually, the reason I raised this is not in regards to your circuit, but from a principle standpoint. Accuracy of ADC reference voltage a part, I'm designing a circuit where I'm comparing a voltage A with a threshold voltage B. Voltage B is variable between 0 and 5 volts, and I implemented what I think is a run of the mill pot. I assume that since there no other voltage divider resistors for voltage B, then it would be safe to assume that variations of temperature would not affect the ADC reading as the resistive material in the pot would vary in value by the same amount, the voltage. Remember this is predicated on a stable ARef voltage. Am I correct withe respect to the pot vs temp.?
+Skullcom Hobby Electronics Are there any other reference chips that could be used in this reference box since i am having a hard time finding REF102C. and can INA135PA replace INA104? Rather, seems finding all the different IC chips are hard or i don't know to look except Mouser / Farnell but shipping is expensive then.
Both the REF102c and the INA105 should be available from most electronic component suppliers. You could replace the REF102C with a LT1021C. There is no direct replacement for the INA105. I have never heard of the INA135PA. As well as Mouser / Farnell (Element 14) try also DigiKey and Radio Spares also try eBay. You do not say which country you are in so difficult to advise fully.
Before I retired I use to visit Stockholm a lot on business and use to take the ferry across to Turku in Finland. Had some great times there and loved the food.
The question is not to make it, adjust it and measure it once. How does it perform over time? How much does it drift? How many digits can you rely on, days, weeks, months later? Now, its about 6-7 years later, what is the value? It would make for an interesting follow up video...
If you were to use a run of the mill multi turn pot without resistors on either end, assuming that the temp drift would be linear throughout the range of the resistive material in the pot, then would that not eliminate any temp drift, and at the same time give you the facility to trim the voltage more accurately as intended?
Donovan Lym Ideally the resistance value of the trim pot should be as low as possible, for example a 1K Pot (25 turn type), but at the same time give you sufficient adjustment you could then used fixed value high tolerance resistors (0.1% or better) either side of the pot to make up the potential divider to the correct mid point required. Good quality high precision multi turn presets with ± 5% resistance tolerance and ± 10ppm/°C are very expensive.
Replying to an old post... If your voltage measuring equipment is _"loading down"_ the voltage output, then you have absolutely ZERO need for this device. In order to take voltage measurements with even the slightest accuracy, your voltmeter should not be placing any significant load on the voltage source. Even the cheapest digital multimeter should have a high enough input impedance that thee will be almost no load whatsoever.
Thanks for your comment. As this unit is designed as a precision voltage reference source for calibration purposes only it does not need to have a load. If you were to use it with a load then you should take in to account that maximum output current for the REF102 chip is very low at around 10 mA up to this level the load regulation is 10ppm/mA max.
I'm finding that when the green LED is on, the red is illuminated simultaneously. I've checked and re-checked everything. When voltage is low the red is solely illuminated as expected. Is that what you see as well, or have I gotten something out of order? Thanks for the great videos!
Hi, If the green LED is on then the red LED should be off. Check your circuit around the LED's. The green LED is connected to a 2N3904 (NPN transistor) and the red LED is connected to a 2N3906 (PNP transistor) - check to make sure you did not use the same transistor type for each LED which would be wrong. Also check the wiring of the two 1K transistor base resistors. It is also possible the transistor feeding the red LED is leaky and may need replacing. Or you have accidentally connected an transistor the wrong way. Regards, Louis
As a follow-up to my last replay. One of my other viewers has pointed out that there is a small error on the schematic I drew on my whiteboard. Seems I made a silly mistake when drawing and did not notice at the time. The PNP transistor (2N3906) collector-emitter need to be reversed. This may have been causing your issue. Sorry about that. Regards, Louis
Hi . Thanks for this video. It's fantastic! How do you compare this reference to the other one from your videos (that one only had one output voltage)? Also isn't the red LED (or it's transistor) wired backwards in the schematic?
Thanks for your comments. All of the voltage reference projects produce a accurate voltage. The accuracy depends on the reference IC used. In this project I use a Texas REF102C which has an accuracy of ± 0.025%. In my other 5 volt reference project I use a Intersil ISL21009BFB850Z which has an accuracy of ± 0.01%, but this IC is sometimes hard to get. The Red LED is wired correctly. The cathodes of both the Red and Green LED's go to the negative supply point. The transistors are a PNP feeding the Red LED and a NPN transistor feeding the Green LED. both transistor bases are connected via 1K resistors to the Voltage Supervisor IC which switches between either High or Low output depending on the level of the battery voltage it is monitoring. So the Red LED comes on when battery low and the Green LED comes on when battery is OK.
Thanks for your answer. The circuit is OK and I understand how it works, but isn't the polarity of red diode is wrong on the schematic at mark 10:00 ? The transistor is a PNP with emitter pointed from the diode to base. But the diode is oriented from the emitter to GND. In other words "arrows" of Red LED and it's transistors emitter are pointing in opposite directions. I believe the emitter and collector of Red LED are swapped.
I had another look at the circuit in the video. The LED that I use in this project is a RED/GREEN three pin (Bi-Colour) LED with a common cathode. So the LED wiring is correct in the circuit. However, I see that I had the PNP (2N3906) transistor with the collector and emitter connections reversed as you quite rightly point out. Although this is technically incorrect the circuits still works due to a number of reasons and theory behind a transistor. The doping of the silicon of the emitter (P), base (N) and collector (P) junctions are all different, with emitter being the most heavily doped element. But although emitter is more doped than collector, the junction size of the collector is larger than the emitter so total number of charge carriers remains almost the same in both emitter and collector regions. So in this practical situation, of my circuit, it does not make a big difference if emitter and collector are exchanged.The other reason it still works in my circuit is due to the low supply voltage of only 3.7v, if the voltage was much higher then the transistor could possibly breakdown. Also if the transistor is reversed wired the transistor current gain is drastically reduced, but in this circuit that does not really matter. Thanks for bringing this point to my attention as it allowed me to cover this situation of what happens when a transistor is reverse connected.
You are using pins that sit in the project board holes and appear to anchor the pin in the board to provide a solderable mounting pin for the components. What are these pins called and where can I buy a bunch of them. Cannot find them in my catalogues - Newark etc.
If you search on eBay for "PCB Vero Board Terminal Pins Prototype" you should find them. Some links below: www.ebay.co.uk/itm/100-x-PCB-Vero-Board-Terminal-Pins-Prototype-/192083933896?hash=item2cb9183ac8:g:NH0AAOSwU9xUS78S also www.ebay.co.uk/itm/100-x-Vero-PCB-terminal-pins-prototype-connector-plug-/262088446288?hash=item3d05b05150:g:6F4AAOSwAYtWGiMb Also similar items from Farnells (Element 14) called VERO 18-0222 TERMINAL PIN, PCB, PK1000 (Order Code: 1209347)
Thank you very much. getting vero board here in the US is not easy. newark, which is a Farnell Co., does not stock it. I remember being able to get the stuff easily before I cam to the US.
Sorry to be so long replying;. the iniquities of aging I'm afraid. I did find the pins on E-Bay UK but no one wants to send them to the USA. So, my brother is getting me several hundred. Thank you very much for the information.
Perfect video and really nice project :) I've subscribed! Good luck and lot of health with new projects. Have you tried to use rotating knob in your Rigol PSU. Don't you find it hard to typing values via buttons?
+Paweł Wojciechowski Thanks Pawel. I am currently working on a new project I hope to upload soon. With regards my Rigol PSU I find typing in the values OK as I am use to it now. I find the rotating know is a bit slow in increasing the values.
What value are the Filter components? A choke & to caps in either side. Also why so many components? I've seen a much simpler units with 2.5V. Same for 10V as well. Thumbs UP anyway!
If you simply use a resistor network to divide the voltage to 5V then when you add a load on that 5V reference the load resistance will be in parallel with the bottom leg of the divider resistor and so effect the voltage out (lowering the 5V). Therefore, different load resistance will effect the accuracy of the 5V by different amounts. Regards, Louis
ah, good old ohm's law and output impedance. thank you for your detailed information. that helps a lot :) looking forward to see more of your projects.
Nice video! Just wondering how the accuracy is when a sufficient load is at the output of the device? Does that affect the accuracy of the reference etc. How can one deal with accuracy with changing load conditions such as changing audio levels etc. of an audio amplifier?
This device is not intended as a power supply to be put under load. The device is intended as an accurate reference voltage to check calibration of devices, such as a multi-meter.
Any voltage measurement device that you should use should never place a significant load across the voltage source. The perfect multimeter would have absolutely zero load, however most multimeters will place a tiny load across the test terminals, however these voltage references will easily handle that. No idea what you are rambling about audio levels... this device has nothing to do with loads or changing load conditions or audio signals or audio amplifiers.
Donovan Lym But the purpose of the project was to provide both 10 volt and 5 volt precision reference source. The Arduino only operates at 5v or 3.3v so you could not use it with a 10 volt reference. You could simply limit your reference to 5 volt if you wished and purchase that version of the reference IC. But I am not sure why you would want to use an Arduino. What do you plan to do with the 5 volt reference? Are you planning to use the Arduino to divide the 5 volts down to lower voltages? If so the accuracy would suffer.
Yes, forgot about the 10 volts. There is apparently a 5 volt version along the same TI series of ic's. What I'm trying to do is to use the Arduino to measure a voltage A between, say, 20 to 60 volts via a voltage divider. I forgot that I need to take into consideration the temp coefficients of the resistors. Anyway, assuming I take care of that, I wanted the measurement to be pretty accurate and therefore needed greater stability than my supply ic 7805 would provide as the ADC ref voltage and hence all my questions. So, questions again: 1. If I got the 5 volt version of the TI chip, would I be able to connect to the ARef pin without the diff amp? 2. As explained in my other posts , that measured voltage A is to be compared with another variable voltage B which would be obtained from a run of the mill pot between the 5 volt rail and ground. Since there is no other resistor, is it safe to assume that the voltage would stay constant with temp variation, providing the supply voltage from the 7805 remains relatively constant. 3. Assuming 2 above is yes, going further, would I be able to also connect the TI output to this pot? The pot is 10k linear trimmer. Hope it's all clear now as to what I'm trying to achieve and would appreciate any feedback.
Donovan Lym OK thanks I see what your trying to do. Answers to your questions: 1. Yes you can connect a 5 volt reference voltage directly to the Arduino ARef pin (you do not need a diff amp). You can connect any voltage directly as long as it is no higher than 5 volt maximum. REMEMBER - If you're using an external reference on the ARef pin, you must set the analog reference to EXTERNAL before calling analogRead(). Otherwise, you will short together the active reference voltage (internally generated) and the AREF pin, possibly damaging the microcontroller on your Arduino board. 2. If you simply are comparing 2 measured voltages with the same setup then I see no reason why the difference between the two would drift as any effect on temperature variations would effect both your A and B voltages equally. So the difference would not be affected. 3. Yes you could connect a 10K trimmer across the output of the 5 volt reference IC. The only thing you have to consider is that you do not exceed the maximum current load for the IC, which I think is around a Max of about 20mA (10K on the output will only draw 0.5mA) - check the datasheet of the IC you decide to use). OTHER POINTS: The Arduino has a 10-bit analog to digital converter. This means that it will map input voltages between 0 and 5 volts into integer values between 0 and 1023 (a total of 1024 steps). This means the resolution between readings of 5 volts is divided in to 1024 units (steps) or to 0.0049 volts (4.9 mV) per unit.
+Scullcom Hobby Electronics dark, I missed that. I bought a bag of 100uH,, oh well. I didn't need then anyway. Maybe that's why it didn't filter any more noise.
Yes the INA105 can be expensive but they are a very stable an accurate IC with ±0.01% Max gain error. You could use the lower cost INA133 but this has a ±0.05% Max gain error. Using the INA133 would reduce the accuracy of the 5 volt reference from ±0.01% to ±0.05% max error. As these are maximum error factors you may well find that the INA133 is still very accurate.
The best price I found and most accurate is 15 US dollars (includes shipping) This is the INA105AM w/8 pin round metal case (TO-99) Specs are available online. This is from China half the price and shipping is for pennies!!!!!
Hi Diego, If you mean the reference voltage output current then it can provide up to 10 mA output current. With no load on the output the total current of the 5V reference unit drain is about 30 mA (the biggest part of the current drain is due to the LED). Regards, Louis
Very nice job Louis. If you offered it as a kit I'm sure you'd have many clients. I have noticed however that the temp coefficient of potentiometers is never better than 100ppm. That's probably the worst tempco component in your design.
The drift may be introduced by the trimmer resistors and the trim pot and therefore would select resistors accordingly. Would you be Abeles to get the trim pot with similar low temp drift characteristics?
Donovan Lym Yes I agree. In some ways it may be better not to fit the trim pot and resistors at all as the accuracy is then guaranteed at ± 0.025% which would be adequate for most hobbyists. If you do use a trim pot then choose a good quality one. One of the best would be the BOURNS (manufacturer part No. 3250W-1-203) 25 turn trimmer with a tolerance of ± 5% and a Temperature Coefficient of ± 50ppm/°C but these are very expensive and can cost about £20 just for one.
@@scullcomhobbyelectronics1702 What about trimming the circuit, noting the resistances that were achieved, and replacing those resistors with fixed resistors that accomplish the same voltage division? All this doesn't matter if you don't have a more accurate reference to calibrate against. That may be the more difficult part for us hobbyists who don't have equipment calibrated. I assume you could build the box and have a third party calibrate it for you?
i want to build this, however ebay dont have any ref102c and the INA105 is about £13! my idea was to run off either a 12v psu or 9v batt. so getting rid of the lipo stuff, led, putting in a toggle switch... ahhh i got a headache now as i havent got the stuff to build one
+G.O.B.S I have one running now with a 9V on a Bread board for the last 14 hours. But I used the MT3608 DC-DC Step Up module, min at 2V and has less noise then the XL6009 5V min and the LM2577 3.5V min. It won't do 3 or 4 Amp, but that's not needed with this, it's smaller also. The Li-ion and LyPo cells had more noise,me think it's the protection built in. I have it tuned 10.0000 with my HP 34401A. I run the step up at 12V, skipped the 15V reg, and everything before it. I tired to filter the little ringing at 90mV Vpp but the inductor and caps did change it. I happen to not have a 1uF cap for pin 8, but tried a 0.47uF, it worked at first then once it charged went to 120mV Vpp. I'm running it for the next 24 hours and log it. I was going to do a volt cutoff but when I first starred I had a drained 9V, opps. 1 hour in and it kept diving off. Then I realized the 9v was under 2V and the step shit off but the ref kept the 10V out. REF takes 1mA, and both take 2.568mA. 9V should last.
+Scullcom Hobby Electronics thanks for the video, it was a fun project. Next solder to a PCB and box it. But I'm thinking about how to get an additional 2.5V and 7.5V in the box on the same board with the same accuracy. How low did you get the noise after it was soldered on the PCB? I'm thinking the little bit I have is mostly the bread board and cheap jumper wires.
I need to make one of those. I'd like to know, what were the little pins you were using that went through the holes on the perf board? I can't seem to figure out where to get them or what they are. Thanks.
They are some times referred to as STRIPBOARD TERMINAL PINS. If you do a search on eBay using that name you should find them. Also try searching for "Terminal Pin Solder Post for Stripboard Vero Board". They are usually sold in quantities of 100 for about £1.35. Hopefully that should find them for you.
The drawback is that you need a high-accuracy voltmeter to calibrate the voltage reference. In reality, it's not possible to get better than ±0.025% for most hobbyists. Grrr.
MrJetra - So you don't want to calibrate your 3-1/2 digit stuff for such a low cost? Why not? Cal tools for multimeters of any accuracy are more expensive than this build.
+betta67 you think its boring most probably cuz u don't appreciate engineering and designing a complete project from scratch. He explained everything in very great detail and its really hard to find a video that has a lot of knowledge u can gain from. About the ultra high precision voltage reference is used to calibrate meters... Now google the price of voltage calibration standard and then you will know how awesome this build really is. Bought modules are used for the easy stuff u don't reinvent the wheel every time.
Although this in an older video, I just wanted to say thanks. What a pleasant and educated chap, with excellent and clear explanations. Your videos are a strong asset to the global community. Well done sir. Subscribed!
Your video encouraged me to make my own precision voltage reference and other related projects.
Thanks for share your knowledge with us.
+agaelema Glad to hear it was of benefit.
+agaelema Glad to hear it was of benefit.
I just discovered your Channel.... And electronics as well.... The way you explain the subject and the thorough process, in which you go about constructing your project is great! 100 year's from now, people will be watching.... Just thought I'd throw that out there.... Your are a great teacher.... I'm 50, but wish youtube was around 40 year's ago. 😀
timothy svec Welcome to the world of electronics. Thanks for your kind comments.
I'm no electronics expert . .indeed I never studied electronics at school/college and am only just getting my head around the subject. However, you have explained this very well, and whilst I don't understand everything, I feel sufficiently confident, after watching this video, to build the reference box which will come in very useful for checking my multimeters
Very nice video. More than adequate for the hobbyist. For more serious applications, reference aging, drift, and temperature/humidity effects need to be considered. Also, you need to have friends with calibration equipment to really get those ppms :) Thanks for taking the time to post this video.
Do not really require this kind of accuracy for the repair work I do but it would improve my confidence rather than confirming one Fluke 77 against the other. I was going make my own but you have done all the hard design work so makes life easier, thanks.
Smashing job, nicely designed and really crisp clean construction :-)
Your use of an op-amp with built in precise resistors is brilliant, so clever :D
zx8401ztv Thanks for your kind comments.
I just came across this video, was thinking of getting a cheap ebay voltage reference board but this was a really well made and interesting video i think ill try and make my own now
Excellent build and explanation of circuit/construction. Thank you. I just need to get a Keithley 2000 Multimeter (or something similar) now to confirm satisfactory results :-(
Great project! Can save myself a Keithley meter now.
you blew my mind @35.34 wow 10.00000 V ?!?! really enjoyed the video !! learned a lot thanks for sharing.
+Asniex Asniex Thanks for you comment.
Thanks a lot. I will mount one for my lab.
have fun building it. Regards, Louis
Great explanation! What I do not understand is the transistor configuration. From my perspective the npn transistor is reverse connected. Emitter shall be towards ground. Please correct me if I am wrong.
Thank you for all your great videos.
Great and informative video, thank you for taking time to share!
So the led changes from red to green when the battery is full, but nothing prevents further charging?
I have negative experience using xl6009 xl6019 step-up modules in battery powered devices. This ic has minimum input operation voltage 5v. when input voltage getting lower, output voltage can be higher than sated before. For similar cases i using mt3608 based modules. It has 2v minimum input voltage
fully agree, no need for a dc to dc converter to bring the voltage up to 20v to have to reduce it to 15V via a regulator when you can directly use a MT3608.
Excellent .how to make o to 10v variable source
Nice. All that's missing is one of your stick on fascias. (Nice idea that). Thanks.
Good tutorial on the voltage ref. I might make one. Very clean and well built. But there's one thing you should consider. What happen when you forget to switch off the box?
lavith.hcm Thanks. With regards your question about forgetting to switch off. You might like to check out my latest project on a 5 volt Reference where I include an auto switch-off feature, which could be added to this project. Check out link below:
ruclips.net/video/sKFT9LouyTY/видео.html
Great no-nonsense guide, thanks!
Thank you.
This channel is not active anymore, its a shame
Thank you so much for sharing this knowledge , the video is very informative to me .
+Zaid Hussain Thanks.
Ref102C has a voltage input range from 11v to 36V, so why bother with a complicated DC to DC converter followed by a linear reg when you could just use a DC buck boost directly to 15V ??
I would assume so the linear regulator can reduce ripple on the input to the REF102, which may pass a small amount of it on to it's output.
This is another gold. Thanks sir
Ah the ridiculousness of import duties - it costs less to import components on printed circuit modules than to buy the components individually! One question though, what is the power consumption of the reference and op amp? Is the complex rechargeable battery circuit really justified? How long would two 9v MN1604 ('PP3') alkaline batteries in series last for?
Hi, have you ever used LM 336 Z2.5 o Z5.0?
Why would an electrolytic capacitor slow the turn-on more than another dialectric of the same capacitance?
Great project Louis...
Thanks Marco.
Fantastic project.
What happens at higher voltages? I.E when the lipo fully charged at 4 or over volts.
Thank you for a great project to try.
+MonkeyMagic Elecifun Thanks. There were no issues at 4 volt. The TPS3809 voltage supervisor has a maximum supply voltage rating of 6 volts.
I will definitely build one of these, eventually. :)
Thank you, very interesting build.
Excellent tutorial video. Thank you for sharing.
Thanks Bob.
Thank You for Your great videos!
Hi, great project, is there any reason why I couldn't use another INA and divide by 2 again ?...cheers.
You can do that, but the error will probably increase due to the error contributed by each chip. You will also increase the noise output.
Nice and simple! What is the accuracy of the Keithley? Important to note that the drift will be dependant on the external trimming components when you use them.
You can improve your results by using low tempco resistors and recalibrating the reference after it is on for at least 2000 hours.
In the circuit diagram at 9:30 am I missing something or is the pnp transistor upside down? It seems that the emitter of the pnp is connected through the led to the negative rail, and the collector through the 120 ohm resistor to the positive rail.
Thanks Michael for bring this to my attention. Looks like I made a silly error when I drew the circuit on my whiteboard. As you say the PNP transistor collector-emitter need to be reversed. When I have time I will draw an updated schematic and provide it as a download for other viewers to use. Thanks again, Louis
Regarding a single multi-turn pot: actually, the reason I raised this is not in regards to your circuit, but from a principle standpoint. Accuracy of ADC reference voltage a part, I'm designing a circuit where I'm comparing a voltage A with a threshold voltage B. Voltage B is variable between 0 and 5 volts, and I implemented what I think is a run of the mill pot. I assume that since there no other voltage divider resistors for voltage B, then it would be safe to assume that variations of temperature would not affect the ADC reading as the resistive material in the pot would vary in value by the same amount, the voltage. Remember this is predicated on a stable ARef voltage. Am I correct withe respect to the pot vs temp.?
Donovan Lym Yes you should be OK.
100mH inductor, as you said, or 100uH ?
+Skullcom Hobby Electronics Are there any other reference chips that could be used in this reference box since i am having a hard time finding REF102C. and can INA135PA replace INA104?
Rather, seems finding all the different IC chips are hard or i don't know to look except Mouser / Farnell but shipping is expensive then.
Both the REF102c and the INA105 should be available from most electronic component suppliers. You could replace the REF102C with a LT1021C. There is no direct replacement for the INA105. I have never heard of the INA135PA. As well as Mouser / Farnell (Element 14) try also DigiKey and Radio Spares also try eBay. You do not say which country you are in so difficult to advise fully.
Greetings, I will keep an eye open. Sadly forgot to mention i am in Sweden.
Before I retired I use to visit Stockholm a lot on business and use to take the ferry across to Turku in Finland. Had some great times there and loved the food.
The question is not to make it, adjust it and measure it once. How does it perform over time?
How much does it drift?
How many digits can you rely on, days, weeks, months later?
Now, its about 6-7 years later, what is the value?
It would make for an interesting follow up video...
If you were to use a run of the mill multi turn pot without resistors on either end, assuming that the temp drift would be linear throughout the range of the resistive material in the pot, then would that not eliminate any temp drift, and at the same time give you the facility to trim the voltage more accurately as intended?
Donovan Lym Ideally the resistance value of the trim pot should be as low as possible, for example a 1K Pot (25 turn type), but at the same time give you sufficient adjustment you could then used fixed value high tolerance resistors (0.1% or better) either side of the pot to make up the potential divider to the correct mid point required. Good quality high precision multi turn presets with ± 5% resistance tolerance and ± 10ppm/°C are very expensive.
Great tutorial. Would it be an advantage to use another op amp to buffer the output so it no loaded down by your measuring equipment?
Replying to an old post... If your voltage measuring equipment is _"loading down"_ the voltage output, then you have absolutely ZERO need for this device. In order to take voltage measurements with even the slightest accuracy, your voltmeter should not be placing any significant load on the voltage source. Even the cheapest digital multimeter should have a high enough input impedance that thee will be almost no load whatsoever.
Should you have done the DVM measurement with a load (ma within the specified within what the part can drive ) ?
Thanks for your comment. As this unit is designed as a precision voltage reference source for calibration purposes only it does not need to have a load. If you were to use it with a load then you should take in to account that maximum output current for the REF102 chip is very low at around 10 mA up to this level the load regulation is 10ppm/mA max.
I'm finding that when the green LED is on, the red is illuminated simultaneously. I've checked and re-checked everything. When voltage is low the red is solely illuminated as expected. Is that what you see as well, or have I gotten something out of order? Thanks for the great videos!
Hi,
If the green LED is on then the red LED should be off. Check your circuit around the LED's.
The green LED is connected to a 2N3904 (NPN transistor) and the red LED is connected to a 2N3906 (PNP transistor) - check to make sure you did not use the same transistor type for each LED which would be wrong. Also check the wiring of the two 1K transistor base resistors.
It is also possible the transistor feeding the red LED is leaky and may need replacing. Or you have accidentally connected an transistor the wrong way.
Regards,
Louis
As a follow-up to my last replay. One of my other viewers has pointed out that there is a small error on the schematic I drew on my whiteboard. Seems I made a silly mistake when drawing and did not notice at the time. The PNP transistor (2N3906) collector-emitter need to be reversed. This may have been causing your issue. Sorry about that. Regards, Louis
Great stuff as always. Thanks.
Thanks for you comment.
Hi . Thanks for this video. It's fantastic! How do you compare this reference to the other one from your videos (that one only had one output voltage)? Also isn't the red LED (or it's transistor) wired backwards in the schematic?
Thanks for your comments. All of the voltage reference projects produce a accurate voltage. The accuracy depends on the reference IC used. In this project I use a Texas REF102C which has an accuracy of ± 0.025%. In my other 5 volt reference project I use a Intersil ISL21009BFB850Z which has an accuracy of ± 0.01%, but this IC is sometimes hard to get.
The Red LED is wired correctly. The cathodes of both the Red and Green LED's go to the negative supply point. The transistors are a PNP feeding the Red LED and a NPN transistor feeding the Green LED. both transistor bases are connected via 1K resistors to the Voltage Supervisor IC which switches between either High or Low output depending on the level of the battery voltage it is monitoring. So the Red LED comes on when battery low and the Green LED comes on when battery is OK.
Thanks for your answer.
The circuit is OK and I understand how it works, but isn't the polarity of red diode is wrong on the schematic at mark 10:00 ? The transistor is a PNP with emitter pointed from the diode to base. But the diode is oriented from the emitter to GND. In other words "arrows" of Red LED and it's transistors emitter are pointing in opposite directions.
I believe the emitter and collector of Red LED are swapped.
I had another look at the circuit in the video. The LED that I use in this project is a RED/GREEN three pin (Bi-Colour) LED with a common cathode. So the LED wiring is correct in the circuit. However, I see that I had the PNP (2N3906) transistor with the collector and emitter connections reversed as you quite rightly point out. Although this is technically incorrect the circuits still works due to a number of reasons and theory behind a transistor.
The doping of the silicon of the emitter (P), base (N) and collector (P) junctions are all different, with emitter being the most heavily doped element. But although emitter is more doped than collector, the junction size of the collector is larger than the emitter so total number of charge carriers remains almost the same in both emitter and collector regions. So in this practical situation, of my circuit, it does not make a big difference if emitter and collector are exchanged.The other reason it still works in my circuit is due to the low supply voltage of only 3.7v, if the voltage was much higher then the transistor could possibly breakdown. Also if the transistor is reversed wired the transistor current gain is drastically reduced, but in this circuit that does not really matter.
Thanks for bringing this point to my attention as it allowed me to cover this situation of what happens when a transistor is reverse connected.
I need every house... good job
You are using pins that sit in the project board holes and appear to anchor the pin in the board to provide a solderable mounting pin for the components. What are these pins called and where can I buy a bunch of them. Cannot find them in my catalogues - Newark etc.
If you search on eBay for "PCB Vero Board Terminal Pins Prototype" you should find them. Some links below:
www.ebay.co.uk/itm/100-x-PCB-Vero-Board-Terminal-Pins-Prototype-/192083933896?hash=item2cb9183ac8:g:NH0AAOSwU9xUS78S
also
www.ebay.co.uk/itm/100-x-Vero-PCB-terminal-pins-prototype-connector-plug-/262088446288?hash=item3d05b05150:g:6F4AAOSwAYtWGiMb
Also similar items from Farnells (Element 14) called VERO 18-0222 TERMINAL PIN, PCB, PK1000 (Order Code: 1209347)
Thank you very much. getting vero board here in the US is not easy. newark, which is a Farnell Co., does not stock it. I remember being able to get the stuff easily before I cam to the US.
You should be able to find it on eBay quite cheap.
Sorry to be so long replying;. the iniquities of aging I'm afraid. I did find the pins on E-Bay UK but no one wants to send them to the USA. So, my brother is getting me several hundred. Thank you very much for the information.
Glad to hear you got it sorted in the end.
Perfect video and really nice project :)
I've subscribed!
Good luck and lot of health with new projects.
Have you tried to use rotating knob in your Rigol PSU. Don't you find it hard to typing values via buttons?
+Paweł Wojciechowski Thanks Pawel. I am currently working on a new project I hope to upload soon. With regards my Rigol PSU I find typing in the values OK as I am use to it now. I find the rotating know is a bit slow in increasing the values.
What value are the Filter components? A choke & to caps in either side.
Also why so many components? I've seen a much simpler units with 2.5V.
Same for 10V as well.
Thumbs UP anyway!
The choke is 100 millihenry and the capacitors are 10nF and 100nF (with the 100nF on the input side of the 15 volt regulator.
Nice work
I would like to ask a very stupid question, but really wants to know the answer.
Why not use resistor to divide for 5V?
If you simply use a resistor network to divide the voltage to 5V then when you add a load on that 5V reference the load resistance will be in parallel with the bottom leg of the divider resistor and so effect the voltage out (lowering the 5V). Therefore, different load resistance will effect the accuracy of the 5V by different amounts. Regards, Louis
ah, good old ohm's law and output impedance. thank you for your detailed information. that helps a lot :) looking forward to see more of your projects.
Nice video! Just wondering how the accuracy is when a sufficient load is at the output of the device? Does that affect the accuracy of the reference etc. How can one deal with accuracy with changing load conditions such as changing audio levels etc. of an audio amplifier?
This device is not intended as a power supply to be put under load. The device is intended as an accurate reference voltage to check calibration of devices, such as a multi-meter.
Any voltage measurement device that you should use should never place a significant load across the voltage source. The perfect multimeter would have absolutely zero load, however most multimeters will place a tiny load across the test terminals, however these voltage references will easily handle that.
No idea what you are rambling about audio levels... this device has nothing to do with loads or changing load conditions or audio signals or audio amplifiers.
If you wee to get the 5 volt version of the TI ic, could you eliminate the precision diff amp and feed directly into the Arduino AREF pin?
Donovan Lym But the purpose of the project was to provide both 10 volt and 5 volt precision reference source. The Arduino only operates at 5v or 3.3v so you could not use it with a 10 volt reference. You could simply limit your reference to 5 volt if you wished and purchase that version of the reference IC. But I am not sure why you would want to use an Arduino. What do you plan to do with the 5 volt reference? Are you planning to use the Arduino to divide the 5 volts down to lower voltages? If so the accuracy would suffer.
Yes, forgot about the 10 volts. There is apparently a 5 volt version along the same TI series of ic's. What I'm trying to do is to use the Arduino to measure a voltage A between, say, 20 to 60 volts via a voltage divider. I forgot that I need to take into consideration the temp coefficients of the resistors. Anyway, assuming I take care of that, I wanted the measurement to be pretty accurate and therefore needed greater stability than my supply ic 7805 would provide as the ADC ref voltage and hence all my questions.
So, questions again:
1. If I got the 5 volt version of the TI chip, would I be able to connect to the ARef pin without the diff amp?
2. As explained in my other posts , that measured voltage A is to be compared with another variable voltage B which would be obtained from a run of the mill pot between the 5 volt rail and ground. Since there is no other resistor, is it safe to assume that the voltage would stay constant with temp variation, providing the supply voltage from the 7805 remains relatively constant.
3. Assuming 2 above is yes, going further, would I be able to also connect the TI output to this pot? The pot is 10k linear trimmer.
Hope it's all clear now as to what I'm trying to achieve and would appreciate any feedback.
Donovan Lym OK thanks I see what your trying to do. Answers to your questions:
1. Yes you can connect a 5 volt reference voltage directly to the Arduino ARef pin (you do not need a diff amp). You can connect any voltage directly as long as it is no higher than 5 volt maximum. REMEMBER - If you're using an external reference on the ARef pin, you must set the analog reference to EXTERNAL before calling analogRead(). Otherwise, you will short together the active reference voltage (internally generated) and the AREF pin, possibly damaging the microcontroller on your Arduino board.
2. If you simply are comparing 2 measured voltages with the same setup then I see no reason why the difference between the two would drift as any effect on temperature variations would effect both your A and B voltages equally. So the difference would not be affected.
3. Yes you could connect a 10K trimmer across the output of the 5 volt reference IC. The only thing you have to consider is that you do not exceed the maximum current load for the IC, which I think is around a Max of about 20mA (10K on the output will only draw 0.5mA) - check the datasheet of the IC you decide to use).
OTHER POINTS:
The Arduino has a 10-bit analog to digital converter. This means that it will map input voltages between 0 and 5 volts into integer values between 0 and 1023 (a total of 1024 steps). This means the resolution between readings of 5 volts is divided in to 1024 units (steps) or to 0.0049 volts (4.9 mV) per unit.
Scullcom Hobby Electronics Thanks for your responses. Great videos. Keep them coming. May I ask what your background is?
Donovan Lym Thanks. I have worked in the electronics field for many years and way back in the 1970's I did teach engineers for about 7 years.
Amazing project! Thanks!
Thanks.
i thought lipo battery will overdischarge if below 3.6v.
but still your video is great..
Super! thanks by video! very nice explanation
Thanks.
Do you use a 100 μH (you say this at 8:47) or a 100 mH (you say this at 12:45) inductor?
***** Sorry about that, It was a small 100 mH inductor.
+Scullcom Hobby Electronics dark, I missed that. I bought a bag of 100uH,, oh well. I didn't need then anyway. Maybe that's why it didn't filter any more noise.
INA105 is a 33 dollar piece is there a low cost replacement
Yes the INA105 can be expensive but they are a very stable an accurate IC with ±0.01% Max gain error. You could use the lower cost INA133 but this has a ±0.05% Max gain error. Using the INA133 would reduce the accuracy of the 5 volt reference from ±0.01% to ±0.05% max error. As these are maximum error factors you may well find that the INA133 is still very accurate.
Scullcom Hobby Electronics I need your email. I must talk with you about a project.
The best price I found and most accurate is 15 US dollars (includes shipping) This is the INA105AM w/8 pin round metal case (TO-99) Specs are available online. This is from China half the price and shipping is for pennies!!!!!
it's would be wizer to poer with 3-4 AAA bataris - with no pulsations!! and cheeper!!!
Great video. Thanks
Thanks.
Hi, the 5V reference which current drain?
Hi Diego,
If you mean the reference voltage output current then it can provide up to 10 mA output current. With no load on the output the total current of the 5V reference unit drain is about 30 mA (the biggest part of the current drain is due to the LED).
Regards,
Louis
It's not for sourcing current. It's for use in high impeadance front ends on measuring devices.
Very nice job Louis. If you offered it as a kit I'm sure you'd have many clients. I have noticed however that the temp coefficient of potentiometers is never better than 100ppm. That's probably the worst tempco component in your design.
Nice design sir...
A good improvement would be to add a small fet so it cuts off the battery thus preventing deep discharge damage to the LiPo
Thanks for the comment and suggestion.
The BMS has the protection already. That's what he explained at the start when he talked about a battery mounted BMS.
Scullcom Hobby Electronics I need your email so I can talk with you about a project.
You can send me a message in Private Messages in You Tube.
The drift may be introduced by the trimmer resistors and the trim pot and therefore would select resistors accordingly. Would you be Abeles to get the trim pot with similar low temp drift characteristics?
Donovan Lym Yes I agree. In some ways it may be better not to fit the trim pot and resistors at all as the accuracy is then guaranteed at ± 0.025% which would be adequate for most hobbyists. If you do use a trim pot then choose a good quality one. One of the best would be the BOURNS (manufacturer part No. 3250W-1-203) 25 turn trimmer with a tolerance of ± 5% and a Temperature Coefficient of ± 50ppm/°C but these are very expensive and can cost about £20 just for one.
@@scullcomhobbyelectronics1702 What about trimming the circuit, noting the resistances that were achieved, and replacing those resistors with fixed resistors that accomplish the same voltage division? All this doesn't matter if you don't have a more accurate reference to calibrate against. That may be the more difficult part for us hobbyists who don't have equipment calibrated. I assume you could build the box and have a third party calibrate it for you?
EXCELLENT
i want to build this, however ebay dont have any ref102c and the INA105 is about £13!
my idea was to run off either a 12v psu or 9v batt.
so getting rid of the lipo stuff, led, putting in a toggle switch... ahhh i got a headache now as i havent got the stuff to build one
+G.O.B.S You can get the IC's online from either Farnell element14 or Mouser Electronics.
cool thanks :)
+G.O.B.S I have one running now with a 9V on a Bread board for the last 14 hours. But I used the MT3608 DC-DC Step Up module, min at 2V and has less noise then the XL6009 5V min and the LM2577 3.5V min. It won't do 3 or 4 Amp, but that's not needed with this, it's smaller also. The Li-ion and LyPo cells had more noise,me think it's the protection built in. I have it tuned 10.0000 with my HP 34401A. I run the step up at 12V, skipped the 15V reg, and everything before it. I tired to filter the little ringing at 90mV Vpp but the inductor and caps did change it. I happen to not have a 1uF cap for pin 8, but tried a 0.47uF, it worked at first then once it charged went to 120mV Vpp. I'm running it for the next 24 hours and log it. I was going to do a volt cutoff but when I first starred I had a drained 9V, opps. 1 hour in and it kept diving off. Then I realized the 9v was under 2V and the step shit off but the ref kept the 10V out. REF takes 1mA, and both take 2.568mA. 9V should last.
+Scullcom Hobby Electronics thanks for the video, it was a fun project. Next solder to a PCB and box it. But I'm thinking about how to get an additional 2.5V and 7.5V in the box on the same board with the same accuracy.
How low did you get the noise after it was soldered on the PCB? I'm thinking the little bit I have is mostly the bread board and cheap jumper wires.
+Scott J Thanks for your comments. Glad to hear it worked OK. Once built and soldered the noise levels were very very low.
Can i use the ina101?
No, the INA101 is a different type of OP Amp which also needs an external resistor to set the gain. Best to stick with the INA105 as in the video.
I need to make one of those. I'd like to know, what were the little pins you were using that went through the holes on the perf board? I can't seem to figure out where to get them or what they are. Thanks.
They are some times referred to as STRIPBOARD TERMINAL PINS. If you do a search on eBay using that name you should find them. Also try searching for "Terminal Pin Solder Post for Stripboard Vero Board". They are usually sold in quantities of 100 for about £1.35. Hopefully that should find them for you.
Good.
The drawback is that you need a high-accuracy voltmeter to calibrate the voltage reference. In reality, it's not possible to get better than ±0.025% for most hobbyists. Grrr.
MrJetra - So you don't want to calibrate your 3-1/2 digit stuff for such a low cost? Why not? Cal tools for multimeters of any accuracy are more expensive than this build.
Why would you need a mobile/portable very (ultra) high precision voltage reference? It's by far an overshoot! Completely useless...
+betta67 and also way too long and boring... beside the fact that it uses bought modules...
+betta67 you think its boring most probably cuz u don't appreciate engineering and designing a complete project from scratch. He explained everything in very great detail and its really hard to find a video that has a lot of knowledge u can gain from. About the ultra high precision voltage reference is used to calibrate meters... Now google the price of voltage calibration standard and then you will know how awesome this build really is. Bought modules are used for the easy stuff u don't reinvent the wheel every time.