This whole series is like watching a music composer create a full-scale opera. What a happy find! I love this YT channel and can't wait to build this PSU - thank you so much for sharing your considerable talents, Dave!
Whoa!! Finally getting around to watching this series. You switched gears and didn't take anybody with you with the changes to the design from the previous video. Appreciate the thorough discussion provided on the design. Please don't stop.
Brilliant! I just came across this channel recently, it maybe 12 years since you made this splendid video but thank you for sharing this video. I was a bench technical officer & also specialize in microwave radio & solar power in telecommunications in my younger days some 41 years ago (I am 63 now & retired) & was attached with our local Telcom company Post & Telecommunications then (now Telikom PNG Limited) and I must say that is a good bench power supply. Thank you again, I wish you good health & May God Bless you. Greetings from Madang, Papua New Guinea!
Very thorough job! I'm glad that rift between you and Microchip seems to have fully healed. ;) I went the 'easy route' for current measurement, using (three) TI INA226 in my supply design. 16-bit ADC, I2C, on-board current, voltage, and power registers, etc. Other than that, it is SCARY, Dave, how closely our designs' functionality are to one another. Although I suppose there's only so many ways to make a wheel.... :)
@DaemonPanda Yeah, considered that, and actually had it on an early draft with another LCD using novel mounting, but just went with a LED on the front panel instead. Free micro pins is an issue.
I love videos like this where you show the whole design process. The power supply looks really good too. Perhaps a bit over-engineered, but it doesn't cost a whole lot more as far as I can tell. Keep up the good work, Dave. :)
Nice looking power supply. I designed and built one about 20 to 25 years ago when I was just learning electronics. I opened it up not too long ago and and was amazed that it never burned my house down! LOL. There's no fuse on the mains and no heat sink on the regulator. I still use it but need to built another now that I'm out of the "know enough to be dangerous" phase.
Absolutely awesome Dave. As a newbie most of this goes over my head but slowly and surely some of it is beginning to stick. I've equipped my workshop with most of your suggestions from an earlier vid but I can't wait to have a go at this. Hurry with the next chapter!!!! All the best to you and yours. Hope 2012 is great for you.
Just binge-watched this series and I never binge-watch anything. Dave has cleared up a ton of mysteries for me. I have a cheap SMPS 32V 5A PS with good operating specs but a terrible front panel design that I would like to replace with a more user friendly interface. Hopefully I've absorbed enough of Dave's design knowledge here to finally begin that project.
I just have to say that you have designed an amazing power supply and I cannot wait for further posts on it. Every aspect is appealing. I was considering getting a bench-top power supply as a component for my lab but to work off of this design would be amazing. Can't wait for future videos. On a side note, it was interesting that Digi-Key listed the ADC converter count but not the actual channel count. Understandably there was few 4-converter ADCs as most contain only a single converter and s/h.
@billysgeo The external ADC/ADC and uCurrent doesn't actually add a huge amount to the final BOM cost, you'd be surprised how much all the other stuff adds up to be.
I won't lie, the very initiative of doing this whole tutorial series is fantastic. Just an incentive to push you further in your next project's making : replace all this atmel, microchip ADC, DAC bushy BOM with a single SMT32F103 . There's already a 11-channels 12bits ADC in these little buggers, along with a 2-channel 12bits DAC, and you can buy pre-propulated boards on eB..y for less than $5 with all decoupling caps and mini USB-B socket mounted, with a neat 100mil SIL fanout. I can garantee you'll fall in love. I even wonder why we still use 8-bits MCU nowadays... Cortex M cores are cheaper, faster, consume less (since they can sleep most of the time) and are easier to debug. Don't get me wrong : I did rely on AT and PICs back in the early 2000, but the only thing that saved them from extinction is the popularity of the Arduino, which, in my experience, is more a fad than a revolution.
Jean-Louis Arnal STM might be great with all those built in features. But I feel you'll miss out on all the other design aspects shown in this video. There is a certain elegance in designing circuits that you just don't get with programming.
With the resistors, using identical parts in the feedback loops means they track both with tempco and with value. 0.1% resistors from the same batch and make at the same time are likely to match to better than 0.01%. Even 5% parts will match to 1% if used that way. Silicon resistors are always done that way, 30% tolerance and horrible tempco, but matching is perfect if you have similar dissipation across the chip.
@pisnahuj10 What's a "hobbyist" price. The case costs money, the LCD costs money, the heatsink costs money, the PCB and all the other parts cost money. It all ads up. This one does cost more than intended though...
I use a 6pin connector and an FTDI TTL-232R-5V USB-to-serial cable on all my designs now. It's a small investment to buy the cables, but it pays off because it's really easy to add UART access to a design plus you can standardise the "debug" port on everything and it doesn't add much to the BOM.
..with an external supply, some reverse polarity protection might be a good idea... Another issue is that the external mains PSU should have some weak coupling (e.g. 10-100K + 10-100nf in parallel) between its negative output and mains earth - you don't want a hard connection to avoid melting scope probe ground leads, but you also don't want the output floating at half mains, which is what you tend to get from non-earthed PSUs.
@SigEpBlue @zega74 Yes, Altium. It's the package I've been using for 20+ years. I have not learned another package sufficiently yet. It's still an open design if Altium is used. What if I did it in Eagle like 90+ % of other "open" hardware designs? You'd have to actually pay $800 or so for the full package of Eagle just to do this size board.
@EEVblog Programming a bootloader and then flashing the firmware via RS232 might be an option. So first of all you programm the bootloader and firmware via ISP. Then configure the reset pin as an i/o pin (programming via isp will not work anymore). For further firmware- updates, you use RS232 and the bootloader. I will recommend to programm the avr via isp on a separate programming jig. So you have the option to use the ISP dedicated Pins PB3, PB4, PB5 as individual I/O pins.
i love how this series ended up its awesome Dave! i was checking parts here in my country wouldn't cost over 12 euros without transformer "i love it" "beauty" :)
Dave if I could make a suggestion, fix the LED/RST issue by putting a jumper in there. Kind of hackish, but also very experimenter friendly. And cheap. And easy.
@starlino The problem with moving to that chip is that the pin requirements become harder. He's trying to reduce the number of pins currently, while it would reduce the part count, it introduces new problems, mainly replacing 5 pins to drive the SPI to 4 12-bit ADC pins and 2 12-bit DAC pins. I know the current schematic is showing that he's using 8 pins, but he states he wants to use a ceramic osc. There is work around though such as using a I2C port expander (comment coming soon).
Internal converters only will give 12 bits if you put the processor to sleep during the conversion. You have to buffer all output leads as well or you will get ground bounce from led current.Add that and the ext converter is looking easier from both cost and software complexity. i worked with 12 bit converters designed in the 1960's. Took 3 cards to implement the analogue side, and used lots of Intersil parts. Did do a stable 12 bits though.
Dave, Thanks for sharing this design. Seems like those negative comments do not understand the concept of learning. Even though it has become more complex, it is something that those guys can learn from.I just checked on that calculator. HP museum; Vintage calculator sold for $130. amazing calculator. too bad they don't sell it anymore. there are all kinds of Programmable Scientific calculators so I suppose one can program them to do the same functions but still this one is special. they don't make it anymore.
Hi Dave, you have connected RESET pin to LED, so you have to disconnect it by fusebit. That means that you cannot use standard ISP for programming. You need to use High Voltage Parallel Programming (HVPP) or (HVSP) for serial.
@EEVblog You can use buttons PB1, PB2, PB3, PB4 with R-2R resistor ladder and internal atmega ADC to decode pressed button. This will give you 3 extra pin.
ATxmega AVR controllers have 12bit ADC and 12bit two channel DACs. Smallest have 34 GPIO and cost under 3 USD. They were available when you made this video.
I apologise if someone has already suggested it (a quick scan showed me another person as thought about it), but you can reduce your pin count a little by using a I2C Port Expander such as the MCP23008 ($0.78 ea. in 100 qty) to drive the buttons and a few of the other non timing sensitive inputs/outputs. Or to reduce the need for a more expensive LCD module (but means more time is spent writing software), a 16-bit port Expander such as the MCP23017 ($0.91 ea. in 100 qty) could be used.
I'm only using trought hole components, but I'm into SMD, and getting the right equipment for that sort of work at the time,hoping SMD is easy to use and fit, In high frequency applications, I agree..SMD is as I concider the best choice, but for starters like me, trought hole is a hole lot easier :)
AWESOME as allways Dave! Can't wait to see the other videos! And by the way that heatsink thing is really a trap for young players!!, and I fell into that a couple of times. Hopefully after scratching my head for a while, I realized my mistake and put some insulator washers and that silica thing to keep it from shortening...you know how that goes...."fuckup and learn"...of course, then you change it to "Fortune and glory, kid. fortune and glory" :)
You said the shunt are expensive but Farnell or Digikey has some nice ones. Ohmite make cheap Kelvin shunts 10 mOhm to 50 mOhm, 0.5 %, 50 ppm/C, 0.5 W - 1 W several types. Search for LVK12, LVK20 or LVK24. You get lower drop, no problem, just select MAX4080T with higher gain and it's all good. Or put pair of 50 mOhm shunts in series :) Next, Microchip also does some cheap precision dual opamps, MCP60x2. For the low offset one, if someone really needs PDIP, there is TC7650 with 5 uV max offset.
@starlino I said nothing about open source, I was talking about the arduino compatibility specifically. It is by far the most common programming environment for microcontrollers for beginners, and thus I think it is a good idea to use it.
Sorry I had to check what gilding the lilly meant - according to urban dictionary:Gilding the Lilly- To adorn unnecessarily something already beautiful. The expression is a condensation of Shakespeare's metaphor in King John: "To gild refined gold, to paint the lily ... is wasteful and ridiculous excess." i.e. So basically what you mean there Dave - you "tarted up" your desisgn in places. that expression is a corker - it putting it straight in me lexicon :)
@spuzzdawg I've designed and hosted soldering workshops for novices. I found there were far less soldering faults on SMD boards than on through hole boards. Main source of faults are usually dry joints, which are almost non-existent with SMD. Add in the fact that you don't have to flip the board all time, and faults caused by parts moving are eliminated. My students used $15 Jaycar irons to solder SOIC and 1206 parts and they finished the task in half the time it took for the through hole board.
Hey Dave, first off im loving the project, and i can't wait to see how it turns out. I was thinking about the MOSFET on resistance, that needed to be under 1.25*10^-3 ohms, how about adding two MOSFETS in serial thus lowering the overall resistance, or atleast making it an option on the board, so it could be soldered in if needed. Well thats my 2 cents, all the best from denmark.
I know this is 12 years old, but still a cool project. Thinking about building this, after I review the components for newer / better versions. I'm assuming the kits are no longer available?
Newhaven display: - 3 extra pins on the MCU HD44780-compatible display: - 1 extra pins on the MCU because you can share lines with the encoders, switches, and/or comm lines of the analog converters
@starlino I look a look at that very interesting device. Was considering that because the DAC is designed for audio purposes, wondering if it could pose a problem. The datasheet states that each channel needs a stream of data else they will load with a single default value... could be a headache. Also states some other internal circuitry, but didn't look into it.
Could you explain why the DAC-D, CK and CS pins are 'fixed'? Is this a PCB type decision (like pin-out kind of things)? Because, in that situation you'd have to use software SPI anyway, as MISO is an input for the AVR, and DAC-D needs to be an output there. Also the search on digikey with 4 converters is explained that digikey doesn't allow you to search for no. of inputs, but number of converters actually in the chip. Those with 4 converters in there can take 4 samples at once, simultaneously
@zega74 It remains an open design if it can be easily recreated in another medium. Dave's been gracious enough to share his entire schematic with us, and who knows how many hours of work he's put into all of this (engineering, taping, building, coding, etc.). Still, I can take his design and recreate it in EAGLE, DEX, TINA, Multisim, DesignSpark, DipTrace, or any of the other odd-number of 'free' ECAD packages. Nothing is lost; IMO, open-source ECAD packages just aren't done cooking yet.
@EEVblog I was just kidding, give out gerbers at the end and you can lock it! I am also pretty novice in KICAD so i understand your reluctance, and also having access to such a nice peace of SW like Altium designer, i would not switch SW! Just make it at least 12 or 15 volts, 10 and some is such a odd value... Oh and adding function buttons (memories for preset programmable voltage values) are the best, try to implement that, I never liked fiddling on power supply pot. to fine tune Vo.
I like that you did get rid of that current led opamp from the previous design, that one was a bad call. But if you are worried about burden voltage on that mosfet shunt, why are you not referencing negative side of your output voltage divider for measurement on drain of that mosfet micro current shunt? I hope your micro will do some out. voltage correction based on these measurements anyway... Is the SCH/PCB SW KICAD?
Doh, this series started so great as basic lab power supply, now this isn't anything like it anymore... You really should have kept this high end stuff in separated video series.
AnatoFIN Actually, Dave stated from the very beginning that this was going to be a MCU-controlled power supply with constant current limit. He had to start with the basic analog part of the PSU, perhaps that's why you liked the first few videos. And except for the uCurrent part of the design all of the circuitry you see here is pretty much necessary for a decent MCU controlled power supply.
What do you think about creating an adjustable switching bench power supply? Should be able to at least use the entire primary side of a PC power supply, since it boosts voltage to over 300 volts. You could either make a new secondary side, or just upgrade the secondary side with higher voltage rated caps and whatever other components need upgraded. The power monitor chip is probably just powered off the 5 volt stand-by transformer. Should be as simple as hooking up your secondary side's output to a variable voltage divider so the the voltage monitor chip always see's 12 volts; and, when you dived the voltage more the actual output goes up. i guess to get less than 12, you could use a simple linear regulator.
Hey Dave, after tearing down the Tekpower and Tenma "Handheld" Power Supplies it looks similar to what you created. I wonder if the Chinese designers came across your videos.
Hello Dave, why don't you use a zener diode instead of schottky for overvoltage protection? If I understand right, once you've reached the overvoltage level schottky will start conducting in the reverse direction irreversibly... sorry for my english
I know this is a very old video, and also my comment is likely to have been asked several times over, but having OCD, I need to ask: In this video, there are multiple references to 0-2.048V or 0-10.24V, even a 2.048V voltage reference. But 2048 steps starting at 0V gives you 2.047 and likewise 10.23. I am aware the error is rather small, but it is still an error. Right. Why isn't the voltage reference 2.047V?
@EEVblog quite possible, sounds like some digital artifact. The parts where you are talking about Digikey and the LCD are artifact free. Its nothing major, and you can be heard clearly, but it is a bit distracting. As for the video itself: I like it! Can't wait to see the next video!
Hi Dave. I LOVE your Videos here. I think I know what Digikey's "mistake" is. You have got 4 Channels but inside you have only ONE Converter. So it converts the first input, then the second and so on. If you have 4 Converters they would convert all 4 Inputs at the same time. Its the same with the PICs. The 8bit PICs have ONE ADC but many Channels which you have to choose in your Registers.
Amaizing to see a proffesional engineer doing his job and seeing the end result. Dave thanks so much for this mini series, I've learned much from them. But howbout' sharing the schematic with teh internet ? :)
Re: the ADC and voltage reference, if you have 12 bits of resolution, 0 current = b00000000000 and full scae current would be b111111111111, that means you really only have 4095 divisions across your scale, so all the math is off and one step actually equals 500.122... mV. Or am I wrong?
Here's the thing, I am a beginner (currently using batteries to power my breadboard) which means this PSU is simply way beyond my skill lvl to build. What I need (I think) is a roughly 1A, 0-6 or 12V, regulated, dual rail PSU, with perhaps one or two other useful features BUT not too complex to build SO at least I can get started. Ideas anyone? I'd like to come back to this one when I'm more experienced.
@starlino Won't that mess up the Arduino compatibility? I think many people appreciate being able to use that programing environment to mod the kit. Other than that, nice suggestion!
As far as I can tell, the 2.048V refence doesn't *quite* get you spot on when you divide by 2048. Your 12-bit adc/dac runs from 0 to 4095, so each bit corresponds to 500.1̅2̅2̅1̅0̅0̅uA. Ideally you would want a 2.0475V reference, or you would want your software to multiply/divide by 2047.5. Mind you, your voltage reference is only accurate to 0.25%, so it could be anywhere between 2.04288 and 2.05312, but on average your adc will be a little high and your dac will be a little low.
For posterity (because I've been reading a *lot* of wrong info, and don't want to be part of the problem) it appears I'm wrong. After digging into some datasheets, every ADC/DAC I could find mapped all-ones (eg 255 on an 8-bit converter, or 4095 on a 12-bit converter) to one bit shy of Vref, so in Dave's case, 4095 does indeed map to 2.0475 volts. This means your typical converter can't *quite* go all the way to Vref, but it certainly makes the math easier.
is there any chance to use RGB backlight on that LCD screen to indicate if everything is OK(green backlight) or PSU is overloaded(red or blinking red backlight)? that would be awesome indication.
No Digi-Key FAIL at all: there are separate columns for the number of ADC *converters* vs the number of *channels*. Most ADCs with multiple channels still only have one converter, and the same goes for microcontrollers.
Dave, Since this looks like it will be a kit, would it be possible to have the final PCB newbie friendly? The schematic looks quite busy, and it would be magnificent to possibly have the PCB silkscreened with boxes around each functional area on the PCB (i.e., voltage reference area, microcontroller and associated parts, etc.). The motivation being that there are plenty of PSU designs out there, along with kits, but we're all taking baby steps with you through this design. :)
Wow this was just what I was looking for, but I hope to have a 0-20v supply. I might have missed it, but did you have the Arduino Sketch listed yet? Ive been working on a OLED volt meter end so far and figured that 0-20v would give 20mV res., but im not getting stable reading.
This whole series is like watching a music composer create a full-scale opera. What a happy find! I love this YT channel and can't wait to build this PSU - thank you so much for sharing your considerable talents, Dave!
The theory is really useful. Im glad dave went all the way with his design. So much to learn here.
Yes!
Whoa!! Finally getting around to watching this series. You switched gears and didn't take anybody with you with the changes to the design from the previous video. Appreciate the thorough discussion provided on the design. Please don't stop.
Brilliant! I just came across this channel recently, it maybe 12 years since you made this splendid video but thank you for sharing this video. I was a bench technical officer & also specialize in microwave radio & solar power in telecommunications in my younger days some 41 years ago (I am 63 now & retired) & was attached with our local Telcom company Post & Telecommunications then (now Telikom PNG Limited) and I must say that is a good bench power supply. Thank you again, I wish you good health & May God Bless you. Greetings from Madang, Papua New Guinea!
Wow, it became very complicated but I like it! It's very interesting to see how an expert would design these things.
yes of cause!!!
"Decided not to use the more expensive 10-turn potentiometers"
Head = exploded.
Very thorough job! I'm glad that rift between you and Microchip seems to have fully healed. ;)
I went the 'easy route' for current measurement, using (three) TI INA226 in my supply design. 16-bit ADC, I2C, on-board current, voltage, and power registers, etc.
Other than that, it is SCARY, Dave, how closely our designs' functionality are to one another. Although I suppose there's only so many ways to make a wheel.... :)
Went from handy dandy to holy moly in less than one video.
he is building a high precision instrument. amazing. Good job. a good engineer sharing some of his capability and knowledge.Thanks a lot.
Absolutely magnificent Dave, you have outdone yourself!
@DaemonPanda Yeah, considered that, and actually had it on an early draft with another LCD using novel mounting, but just went with a LED on the front panel instead. Free micro pins is an issue.
Dave, I love this design! I would buy a kit when you have it finalized. Looking forward to the final videos. Keep up the great work and videos!
I love videos like this where you show the whole design process. The power supply looks really good too. Perhaps a bit over-engineered, but it doesn't cost a whole lot more as far as I can tell. Keep up the good work, Dave. :)
Man, I get the feeling there needs to be about 4 more videos between parts 4 and 5 in order to properly explore all this new stuff.
Nice looking power supply. I designed and built one about 20 to 25 years ago when I was just learning electronics. I opened it up not too long ago and and was amazed that it never burned my house down! LOL. There's no fuse on the mains and no heat sink on the regulator.
I still use it but need to built another now that I'm out of the "know enough to be dangerous" phase.
Absolutely awesome Dave. As a newbie most of this goes over my head but slowly and surely some of it is beginning to stick. I've equipped my workshop with most of your suggestions from an earlier vid but I can't wait to have a go at this. Hurry with the next chapter!!!! All the best to you and yours. Hope 2012 is great for you.
This looks and feels professional.
You got yourself another subscriber.
Just binge-watched this series and I never binge-watch anything. Dave has cleared up a ton of mysteries for me. I have a cheap SMPS 32V 5A PS with good operating specs but a terrible front panel design that I would like to replace with a more user friendly interface. Hopefully I've absorbed enough of Dave's design knowledge here to finally begin that project.
I just have to say that you have designed an amazing power supply and I cannot wait for further posts on it. Every aspect is appealing. I was considering getting a bench-top power supply as a component for my lab but to work off of this design would be amazing. Can't wait for future videos.
On a side note, it was interesting that Digi-Key listed the ADC converter count but not the actual channel count. Understandably there was few 4-converter ADCs as most contain only a single converter and s/h.
@billysgeo The external ADC/ADC and uCurrent doesn't actually add a huge amount to the final BOM cost, you'd be surprised how much all the other stuff adds up to be.
@MrP00f If so, sorry. I used a noise removal algorithm on this one to get rid of the fan conditioning noise. Artifacts of that?
I won't lie, the very initiative of doing this whole tutorial series is fantastic.
Just an incentive to push you further in your next project's making : replace all this atmel, microchip ADC, DAC bushy BOM with a single SMT32F103 . There's already a 11-channels 12bits ADC in these little buggers, along with a 2-channel 12bits DAC, and you can buy pre-propulated boards on eB..y for less than $5 with all decoupling caps and mini USB-B socket mounted, with a neat 100mil SIL fanout.
I can garantee you'll fall in love.
I even wonder why we still use 8-bits MCU nowadays... Cortex M cores are cheaper, faster, consume less (since they can sleep most of the time) and are easier to debug. Don't get me wrong : I did rely on AT and PICs back in the early 2000, but the only thing that saved them from extinction is the popularity of the Arduino, which, in my experience, is more a fad than a revolution.
Jean-Louis Arnal STM might be great with all those built in features. But I feel you'll miss out on all the other design aspects shown in this video. There is a certain elegance in designing circuits that you just don't get with programming.
@smithyduk I originally wanted a buzzer, but forgot to include it in the final design - Doh!
Very nice, thanks for making the informative series Dave! I think I'll buy a kit when they're finished!
With the resistors, using identical parts in the feedback loops means they track both with tempco and with value. 0.1% resistors from the same batch and make at the same time are likely to match to better than 0.01%. Even 5% parts will match to 1% if used that way. Silicon resistors are always done that way, 30% tolerance and horrible tempco, but matching is perfect if you have similar dissipation across the chip.
@pisnahuj10 What's a "hobbyist" price. The case costs money, the LCD costs money, the heatsink costs money, the PCB and all the other parts cost money. It all ads up. This one does cost more than intended though...
I use a 6pin connector and an FTDI TTL-232R-5V USB-to-serial cable on all my designs now. It's a small investment to buy the cables, but it pays off because it's really easy to add UART access to a design plus you can standardise the "debug" port on everything and it doesn't add much to the BOM.
..with an external supply, some reverse polarity protection might be a good idea...
Another issue is that the external mains PSU should have some weak coupling (e.g. 10-100K + 10-100nf in parallel) between its negative output and mains earth - you don't want a hard connection to avoid melting scope probe ground leads, but you also don't want the output floating at half mains, which is what you tend to get from non-earthed PSUs.
@SigEpBlue @zega74 Yes, Altium. It's the package I've been using for 20+ years. I have not learned another package sufficiently yet. It's still an open design if Altium is used. What if I did it in Eagle like 90+ % of other "open" hardware designs? You'd have to actually pay $800 or so for the full package of Eagle just to do this size board.
@EEVblog Programming a bootloader and then flashing the firmware via RS232 might be an option. So first of all you programm the bootloader and firmware via ISP. Then configure the reset pin as an i/o pin (programming via isp will not work anymore). For further firmware- updates, you use RS232 and the bootloader. I will recommend to programm the avr via isp on a separate programming jig. So you have the option to use the ISP dedicated Pins PB3, PB4, PB5 as individual I/O pins.
i love how this series ended up its awesome Dave! i was checking parts here in my country wouldn't cost over 12 euros without transformer "i love it" "beauty" :)
Dave if I could make a suggestion, fix the LED/RST issue by putting a jumper in there. Kind of hackish, but also very experimenter friendly. And cheap. And easy.
@nlhans1990 They aren't fixed, that's a leftover on the schematic I didn't clean up. Yes, software SPI.
@Vlakpage The Microchip opamps only go to 6V max.
Dave uses opera? Awesome :D. Best browser out there.
I like the shirt Dave. It looks like you're getting comfortable. Have fun, bud!
@starlino The problem with moving to that chip is that the pin requirements become harder. He's trying to reduce the number of pins currently, while it would reduce the part count, it introduces new problems, mainly replacing 5 pins to drive the SPI to 4 12-bit ADC pins and 2 12-bit DAC pins. I know the current schematic is showing that he's using 8 pins, but he states he wants to use a ceramic osc. There is work around though such as using a I2C port expander (comment coming soon).
@DaemonPanda No, I use Chrome mostly. Opera had less clutter for the video.
Use the RGB back light as current overload indicator by switching from green or blue (user selectable) to red.
Internal converters only will give 12 bits if you put the processor to sleep during the conversion. You have to buffer all output leads as well or you will get ground bounce from led current.Add that and the ext converter is looking easier from both cost and software complexity. i worked with 12 bit converters designed in the 1960's. Took 3 cards to implement the analogue side, and used lots of Intersil parts. Did do a stable 12 bits though.
Dave, Thanks for sharing this design. Seems like those negative comments do not understand the concept of learning. Even though it has become more complex, it is something that those guys can learn from.I just checked on that calculator. HP museum; Vintage calculator sold for $130. amazing calculator. too bad they don't sell it anymore. there are all kinds of Programmable Scientific calculators so I suppose one can program them to do the same functions but still this one is special. they don't make it anymore.
Hi Dave, you have connected RESET pin to LED, so you have to disconnect it by fusebit. That means that you cannot use standard ISP for programming. You need to use High Voltage Parallel Programming (HVPP) or (HVSP) for serial.
@EEVblog
You can use buttons PB1, PB2, PB3, PB4 with R-2R resistor ladder and internal atmega ADC to decode pressed button. This will give you 3 extra pin.
ATxmega AVR controllers have 12bit ADC and 12bit two channel DACs. Smallest have 34 GPIO and cost under 3 USD. They were available when you made this video.
I apologise if someone has already suggested it (a quick scan showed me another person as thought about it), but you can reduce your pin count a little by using a I2C Port Expander such as the MCP23008 ($0.78 ea. in 100 qty) to drive the buttons and a few of the other non timing sensitive inputs/outputs. Or to reduce the need for a more expensive LCD module (but means more time is spent writing software), a 16-bit port Expander such as the MCP23017 ($0.91 ea. in 100 qty) could be used.
I'm only using trought hole components, but I'm into SMD, and getting the right equipment for that sort of work at the time,hoping SMD is easy to use and fit, In high frequency applications, I agree..SMD is as I concider the best choice, but for starters like me, trought hole is a hole lot easier :)
Fantastic Dave!!! Waiting time laps of PCB design :)
AWESOME as allways Dave! Can't wait to see the other videos! And by the way that heatsink thing is really a trap for young players!!, and I fell into that a couple of times. Hopefully after scratching my head for a while, I realized my mistake and put some insulator washers and that silica thing to keep it from shortening...you know how that goes...."fuckup and learn"...of course, then you change it to "Fortune and glory, kid. fortune and glory" :)
@starlino Didn't catch that one, I only found a 40pin DIP 24F chip I think. Will have to investigate, thanks.
You said the shunt are expensive but Farnell or Digikey has some nice ones. Ohmite make cheap Kelvin shunts 10 mOhm to 50 mOhm, 0.5 %, 50 ppm/C, 0.5 W - 1 W several types. Search for LVK12, LVK20 or LVK24.
You get lower drop, no problem, just select MAX4080T with higher gain and it's all good. Or put pair of 50 mOhm shunts in series :)
Next, Microchip also does some cheap precision dual opamps, MCP60x2. For the low offset one, if someone really needs PDIP, there is TC7650 with 5 uV max offset.
@snik2pl Yeah, I only just found that out. That limitation kinda sucks...
Awesome again Dave!!!
@starlino I said nothing about open source, I was talking about the arduino compatibility specifically. It is by far the most common programming environment for microcontrollers for beginners, and thus I think it is a good idea to use it.
Sorry I had to check what gilding the lilly meant - according to urban dictionary:Gilding the Lilly- To adorn unnecessarily something already beautiful. The expression is a condensation of Shakespeare's metaphor in King John: "To gild refined gold, to paint the lily ... is wasteful and ridiculous excess." i.e. So basically what you mean there Dave - you "tarted up" your desisgn in places. that expression is a corker - it putting it straight in me lexicon :)
Will we see the finished power supply soon? It's looking good just from the schematic, maybe a little over-the-top tho.
I like this design :D
@spuzzdawg I've designed and hosted soldering workshops for novices. I found there were far less soldering faults on SMD boards than on through hole boards. Main source of faults are usually dry joints, which are almost non-existent with SMD. Add in the fact that you don't have to flip the board all time, and faults caused by parts moving are eliminated. My students used $15 Jaycar irons to solder SOIC and 1206 parts and they finished the task in half the time it took for the through hole board.
Thanks Dave
Hey Dave, first off im loving the project, and i can't wait to see how it turns out.
I was thinking about the MOSFET on resistance, that needed to be under 1.25*10^-3 ohms, how about adding two MOSFETS in serial thus lowering the overall resistance, or atleast making it an option on the board, so it could be soldered in if needed.
Well thats my 2 cents, all the best from denmark.
Number of channels and number converters are different parameters. Converter are the actually hardware engines
I know this is 12 years old, but still a cool project.
Thinking about building this, after I review the components for newer / better versions.
I'm assuming the kits are no longer available?
Newhaven display:
- 3 extra pins on the MCU
HD44780-compatible display:
- 1 extra pins on the MCU because you can share lines with the encoders, switches, and/or comm lines of the analog converters
@DaemonPanda Or you could use it to indicate when the current limiter is activated, etc, which could be pretty neat.
Is there a written version of this whole series ? This is a goldmine but I personally learn stuff better than by watching videos.
you are amazing dave! when are you going to start your own company!
@Zadster
Not the ft232 - but an unpopulated SSOP-28 with an unpopulated USB connector
and the CN3
@starlino I look a look at that very interesting device. Was considering that because the DAC is designed for audio purposes, wondering if it could pose a problem. The datasheet states that each channel needs a stream of data else they will load with a single default value... could be a headache. Also states some other internal circuitry, but didn't look into it.
@ultimatefang007 He spent like 5 mins talking about that exact question. You should watch the entire video :)
Could you explain why the DAC-D, CK and CS pins are 'fixed'? Is this a PCB type decision (like pin-out kind of things)? Because, in that situation you'd have to use software SPI anyway, as MISO is an input for the AVR, and DAC-D needs to be an output there.
Also the search on digikey with 4 converters is explained that digikey doesn't allow you to search for no. of inputs, but number of converters actually in the chip. Those with 4 converters in there can take 4 samples at once, simultaneously
@zega74 It remains an open design if it can be easily recreated in another medium. Dave's been gracious enough to share his entire schematic with us, and who knows how many hours of work he's put into all of this (engineering, taping, building, coding, etc.).
Still, I can take his design and recreate it in EAGLE, DEX, TINA, Multisim, DesignSpark, DipTrace, or any of the other odd-number of 'free' ECAD packages. Nothing is lost; IMO, open-source ECAD packages just aren't done cooking yet.
Why would you need a 1% timebase accuracy for the UART rather than 5% (a drift of less than half a bit per transmission)?
What if you use the LCD backlight for status? Display goes red when you go into current limit, green when you're not, yellow when you're at 90% load.
@EEVblog
I was just kidding, give out gerbers at the end and you can lock it!
I am also pretty novice in KICAD so i understand your reluctance, and also having access to such a nice peace of SW like Altium designer, i would not switch SW!
Just make it at least 12 or 15 volts, 10 and some is such a odd value...
Oh and adding function buttons (memories for preset programmable voltage values) are the best, try to implement that, I never liked fiddling on power supply pot. to fine tune Vo.
I like that you did get rid of that current led opamp from the previous design, that one was a bad call.
But if you are worried about burden voltage on that mosfet shunt, why are you not referencing negative side of your output voltage divider for measurement on drain of that mosfet micro current shunt?
I hope your micro will do some out. voltage correction based on these measurements anyway...
Is the SCH/PCB SW KICAD?
Doh, this series started so great as basic lab power supply, now this isn't anything like it anymore...
You really should have kept this high end stuff in separated video series.
Basic supply was more or less taken care of in the first video, after that, it became an awesome PSU
I agree. This is too fancy for 1A power supply.
AnatoFIN Actually, Dave stated from the very beginning that this was going to be a MCU-controlled power supply with constant current limit. He had to start with the basic analog part of the PSU, perhaps that's why you liked the first few videos. And except for the uCurrent part of the design all of the circuitry you see here is pretty much necessary for a decent MCU controlled power supply.
Great series, I'm planning to make my own version using an STM32 which has 12bit DAC/ADC with some cheap seven segment display modules
What do you think about creating an adjustable switching bench power supply? Should be able to at least use the entire primary side of a PC power supply, since it boosts voltage to over 300 volts. You could either make a new secondary side, or just upgrade the secondary side with higher voltage rated caps and whatever other components need upgraded. The power monitor chip is probably just powered off the 5 volt stand-by transformer. Should be as simple as hooking up your secondary side's output to a variable voltage divider so the the voltage monitor chip always see's 12 volts; and, when you dived the voltage more the actual output goes up. i guess to get less than 12, you could use a simple linear regulator.
Hey Dave, after tearing down the Tekpower and Tenma "Handheld" Power Supplies it looks similar to what you created. I wonder if the Chinese designers came across your videos.
Hello Dave, why don't you use a zener diode instead of schottky for overvoltage protection? If I understand right, once you've reached the overvoltage level schottky will start conducting in the reverse direction irreversibly... sorry for my english
I know this is a very old video, and also my comment is likely to have been asked several times over, but having OCD, I need to ask:
In this video, there are multiple references to 0-2.048V or 0-10.24V, even a 2.048V voltage reference. But 2048 steps starting at 0V gives you 2.047 and likewise 10.23. I am aware the error is rather small, but it is still an error. Right. Why isn't the voltage reference 2.047V?
I was really hoping for a bench supply, like a dual output, etc.
You should make your own EEVBlog Title block for Altium
@EEVblog quite possible, sounds like some digital artifact. The parts where you are talking about Digikey and the LCD are artifact free. Its nothing major, and you can be heard clearly, but it is a bit distracting.
As for the video itself: I like it! Can't wait to see the next video!
Hi Dave. I LOVE your Videos here. I think I know what Digikey's "mistake" is. You have got 4 Channels but inside you have only ONE Converter. So it converts the first input, then the second and so on. If you have 4 Converters they would convert all 4 Inputs at the same time. Its the same with the PICs. The 8bit PICs have ONE ADC but many Channels which you have to choose in your Registers.
Amaizing to see a proffesional engineer doing his job and seeing the end result. Dave thanks so much for this mini series, I've learned much from them. But howbout' sharing the schematic with teh internet ? :)
Re: the ADC and voltage reference, if you have 12 bits of resolution, 0 current = b00000000000 and full scae current would be b111111111111, that means you really only have 4095 divisions across your scale, so all the math is off and one step actually equals 500.122... mV. Or am I wrong?
Here's the thing, I am a beginner (currently using batteries to power my breadboard) which means this PSU is simply way beyond my skill lvl to build. What I need (I think) is a roughly 1A, 0-6 or 12V, regulated, dual rail PSU, with perhaps one or two other useful features BUT not too complex to build SO at least I can get started. Ideas anyone? I'd like to come back to this one when I'm more experienced.
@ultimatefang007 12bit resolution instead of 10bit.
@g1fsh There is one more thing in the design people haven't seen yet...
@starlino Won't that mess up the Arduino compatibility? I think many people appreciate being able to use that programing environment to mod the kit. Other than that, nice suggestion!
As far as I can tell, the 2.048V refence doesn't *quite* get you spot on when you divide by 2048. Your 12-bit adc/dac runs from 0 to 4095, so each bit corresponds to 500.1̅2̅2̅1̅0̅0̅uA. Ideally you would want a 2.0475V reference, or you would want your software to multiply/divide by 2047.5.
Mind you, your voltage reference is only accurate to 0.25%, so it could be anywhere between 2.04288 and 2.05312, but on average your adc will be a little high and your dac will be a little low.
For posterity (because I've been reading a *lot* of wrong info, and don't want to be part of the problem) it appears I'm wrong. After digging into some datasheets, every ADC/DAC I could find mapped all-ones (eg 255 on an 8-bit converter, or 4095 on a 12-bit converter) to one bit shy of Vref, so in Dave's case, 4095 does indeed map to 2.0475 volts. This means your typical converter can't *quite* go all the way to Vref, but it certainly makes the math easier.
22:50 You dont have issues driving an ADC from an op-amp without a resistor to block capacitive loading?
is there any chance to use RGB backlight on that LCD screen to indicate if everything is OK(green backlight) or PSU is overloaded(red or blinking red backlight)? that would be awesome indication.
No Digi-Key FAIL at all: there are separate columns for the number of ADC *converters* vs the number of *channels*. Most ADCs with multiple channels still only have one converter, and the same goes for microcontrollers.
Why did you decide against using the hardware SPI for both ADC and DAC?
Have u put the code for this project up anywhere I'm building a psu and having trouble getting the mcp4922 working
Optical vs mechanical rotary encoder. What was the design decision there?
thunbs up just for using base2 numbers on your voltage reference ... (double that if you are also using accembly for the microprocessor code ;) )
Dave,
Since this looks like it will be a kit, would it be possible to have the final PCB newbie friendly? The schematic looks quite busy, and it would be magnificent to possibly have the PCB silkscreened with boxes around each functional area on the PCB (i.e., voltage reference area, microcontroller and associated parts, etc.).
The motivation being that there are plenty of PSU designs out there, along with kits, but we're all taking baby steps with you through this design. :)
Wow this was just what I was looking for, but I hope to have a 0-20v supply. I might have missed it, but did you have the Arduino Sketch listed yet? Ive been working on a OLED volt meter end so far and figured that 0-20v would give 20mV res., but im not getting stable reading.