Simple DIY power supply + thermal drift compensation

Building a simple 2.2 - 5.5V adjustable linear bench power supply. This time, the current limit circuitry uses just discrete transistors. How to compensate the temperature drift of a bipolar silicon transistor B-E junction using a Schottky diode.
Please support my channel on Patreon:
/ diodegonewild
Instagram:
/ savage_danyk

[@robertatpierpontbeach]

Nice to see how you incorporated an SMD component to show the hint on how to make one fit on a through-hole board!

[@rtos]

Nice idea using a schottky diode to compensate for the Vbe drift with temperature.

[@WagTsX]

that full bridge rectifier at 13:05 made me think and check twice before just follow a datasheet, even from a reputable supplier

[@cisarvialpando7412]

Yes it's such a wrong design 🤦🏻‍♂️🤦🏻‍♂️🤦🏻‍♂️

[@RicoElectrico]

TI app notes are often sub-par quality. They clearly are not proofread and contain typos. Either they let their interns write them or outsource documentation to low-cost countries with poor engineering culture.

[@snakezdewiggle6084]

It's done on purpose ! You are only ment to be interested in the Part made by the company... ;)

[@ЁбрагимИпатенкоибнАдхарма]

Just revers-ingeneering some pcb with HC273, and TI pdf has swapped pins 8 & 9 that are D & Q of 4th trigger🤦‍♂

[@pierre7127]

Like often you put a lot of effort to provide a well build and tested solution. A really appreciate the detailed explanations about the functionality of the different components and why you chose them. Thank you very much for this video.

[@wolfrobben6914]

Nice video once again. I learned a lot.

[@DiodeGoneWild]

Thank you for your support and appreciation ;)

[@rolfts5762]

Thanks for sharing good&educational videos Dan. Really enjoyable to watch and listen (..the schematic-talks is extra interesting&enjoyable). Have nice week to all around you in Czechia, and everywhere else. /br, from norway

[@DiodeGoneWild]

Big thanks for your support!

[@d.k.9406]

His way to make this videos with all the work to make the shematics, charts and raphics... explain evey detail and the correliation between parts and their properties.... Make a perfect guide or walkthrough electronics... Not like others with non-functional nonsense but cool-looking 10M clicks and views ... leaving a lot of confused humans

[@salan3]

I am my worst enemy when it comes to building things. I want it to look perfect and so 90% of the time I don't actually get around to building it!! Seeing you 'just get the Dremel out and make a board' reminds me to stop being so fussy over unimportant things. Better to of made the thing and it not look like a kit then never getting around to it!

[@iamdarkyoshi]

Very cool project, love seeing discreet component designs :)

[@greengrayradio1394]

Very interesting, quite a simple circuit which keeps the current constant over a big temperature range! I will try this circuit, for sure!

[@essas.coisas]

what happened with the paper lol :) loved the table, and the box - very sci-fi :p

[@piconano]

At the risk of sounding like a Karen, I strongly suggest you show fibreglass more respect. Never cut or sand PCBs made of fibreglass inside the apartment. The microscopic shards of glass will get everywhere, including under your skin and in to your lungs. Do it outside and rinse hands with cold water thoroughly (Warm water opens up your skin's pours). Think of your super cute pets if you don't think of yourself. You still get a thumbs up from me because your content is awesome.

[@CliveChamberlain946]

Yup, I went to a junkyard for a used 12 Vdc centrifugal auto blower to take this debris into a vacuum cleaner bag located under the bench. Use a bendable aluminium dryer exhaust pipe to position the suction flow near the work being done.

[@liudas5377]

Ok Karen...

[@lmwlmw4468]

Nice bridge rectifier from Texas Instruments........

[@janno288]

Your videos never disappoint

[@KeritechElectronics]

Agreed!

[@d46512]

Dropping in a TL431 would be cheating and would not make for such an educational video 😂

[@DiodeGoneWild]

With a TL431, you'd have to drop 2.5 bloody volts just to sense the current. Way too lossy.

[@RTX_Labs_Productions]

Cool one, I decided to build one by using a ATX PC Power Supply

[@zyghom]

I was waiting almost 19 minutes to see the cat - really? that is TOO LONG! please improve it! ;-)

[@Alexelectricalengineering]

Awesome like always 👍👍👍👍

[@tajtrlik1111]

Veľmi pekná a prepracovaná konštrukcia zdroja, tvoja schopnosť improvizovať ma neprestáva fascinovať a to myslím len v dobrom, lebo dokážeš pekne využiť to čo je doma k dispozícií a veľa z toho sú tvoje ďalšie výrobky. Aj ja vyrábam niektoré dosky, ktoré navrhujem ručne metódou deliacich čiar frézovaním, niektoré, ktoré navrhujem na PC v Eagle zase kreslím podľa vytlačenej predlohy fixkou a leptám chloridom železitým.

[@kemalkurt5257]

Very good project. Please keep up your work this schottky trick is something you cannot learn in the school thanks for the trick.

[@snakezdewiggle6084]

Thank you for ALL of Your Laborious work, scrap-paper oscilloscope traces and diagrams. One of best channels on yt. ;)

[@fabriziobrutti1205]

The bridge rectifier at 13:00 just killed my brain :_)

[@Comrade_YG]

The diodes are just going in circles

[@CliveChamberlain946]

Recommendation for my favorite inventor - don't breath PCB fibres grinding the copper! Go to auto junkyard and buy a used centrifugal blower to blow this debris into a vacuum cleaner bag located under the bench. Use a bendable aluminium dryer pipe to position the suction flow near the work being done.

[@beatrute2677]

man i love it when you do a power electronics vid. I always learn something new even the basic stuff.

[@cups3]

When I was still working I would always be building special test equipment for the shop. I tired to perfect a solid state DC load . I experimented on and off but never got the thermal compensation perfect ,it was better but I was never happy. The final was a thermistor epoxied to the heat sink which robbed base drive current from the driver transistors. Also I see you seem to have changed the Schottky resistor from 68K to 82 K ?

[@DiodeGoneWild]

Anything from 56k to 120k seems to work about the same. This range is moving the drift coefficient in the least steep part of a parabola I guess.

[@noneofyerbeeswax8194]

You can never get the thermal compensation perfect by using discrete transistors. You could use matched pairs (on the same die), but in my experience, nothing beats OpAmps when it comes to DC precision.

[@KeritechElectronics]

Nice project and some good science on the temperature compensation. The transformer surely gets hot - time for some ventilation holes :) I had a good laugh at what should be a FULL BRIDGE RECTIFIER but the datasheet got it wrong :)

[@Pirelli.]

Learned a lot again DGW, thank you! And now I know that a black painted hotsink dissipates heat better (3-8%)! I'm going to paint... ⚫😃

[@DiodeGoneWild]

thanks :) but remember, not every black paint is black for infrared.

[@bitelaserkhalif]

I made bench power supply using broken laptop charger and stepdown module. It cannot regulate current Only needed 0-12v anyway.

[@alexloktionoff6833]

Awesome scientific approach!

[@rolfts5762]

Thanks for video Dan/DGW.. (..today undersigned stomach hurted 🙂specially when pointing out the bridge-rect. (at 13min)) ,,//Have nice week. Thanks again, you make very good&educational content on the channel.

[@UpLateGeek]

What a coincidence, I was just looking at a new benchtop power supply. My current one is a "cheap" Chinese switchmode one, but it turns out you can get pretty reasonable linear power supplies for a decent price. The so-called "cheap" Chinese switchmode power supply costs about $90 Aussie dollary doos, but I can get a linear one from the silicon-based electronics supplier for $139. Building my own probably wouldn't be that much less, considering the transformer would be around $70 on its own. Anyway, this was an interesting little project. It's probably more worth it to make this kind of small power supply yourself, especially since you can simplify it for such a special purpose.

[@henriquelausch6999]

It's amazing how that so called FUUULL BRIIDGE REECTIFIER will drive it's diodes wildly

[@davidmckendry4491]

FUUULL Bridge rectifier not a puny half bridge.We need the Mehdi the RECTIFIER.

[@wither8]

Get some styrofoam and you can make an insulated box that performs fairly well for temperature sweeps. The thermal mass of boards is little, so you can generally do alright by putting a PT100 from the lid of your container, at the center. There will be a thermal gradient just due to convection heat flow but you're not building a cal standard ;). Since you built your entire DIY temperature probe and everything, add a uc to simply record your temperature data as you sweep and then just read it out via a simple RS232. Then you won't have to make those tables by hand. It will also allow you to read-out time, thus you can see how things change as dtemperature/dt. It will also allow you to create a feedback loop to linearize your dtemp/dt, allowing you to negate thermal hysteresis, as well as controlling the heat source/sink (supplement the blow dryer with a set of cheap Chinese peltier devices and make an array of them) so you can model your desired rate of change. I've watched about 20 or 30 of your videos, maybe 10 of which contained tools you made (clever ring tester!). Write a simple LXI/SCPI/whatever interface (or even your own, though if you use the standard protocols, you can use Python, the scpi/lxi/etc libs, and leverage existing work, such as mporting libraries and being able to generate graphs within literally 5 lines-- see Marco Reps' videos) Fluke's reference 732 is a fascinating case study on how, via a simple analog circuit, Fluke was able to get incredible low-drift, long term stability using less than 6 inch styrofoam encasements and heated silicon padding. www.eevblog.com/forum/metrology/fluke-732a-with-popcorn-noise-sza/25/ and again, see Marco Reps' 732A tear down. The SZA, I think, was also used on their metrology grade transfer stations (the 7200As which, though power hungry, still perform far good enough to generate NIST cals for customres). In order to kill tempco issues, they ensure encapsulation so there is no airflow, and the SZA vref has this simple yet genius design Tempco on a BJT is described here: web.mit.edu/klund/www/Dphysics.pdf and people.engr.tamu.edu/s-sanchez/607%20Lect%204%20Bandgap-2009.pdf page 11. If you get matched pairs, you can do clever things to negate drift regardless of load. Great vids keep em up!

[@d.k.9406]

Danke! Thank you. Discrete is best. I love its simplicity .. but nevertheless its good functionality. Nice explained. TY for your work with all the measurements drawings/graphs. But love discrete circuits with insane complexity too. ....and then the need of your explainations how it works to understand it XD

[@matid8453]

maybe something power supply with power over 300W? I supposed be very interesting

[@borahkrsurajit]

Can you design a SMPS power supply with an output from 3V to 24V at 5 Amps. I need a variable supply with a wide voltage range for different charging and test requirements. All I can find are chinese ones at a reasonable price. Lab bench power supplies are very very expensive. I have experience with building circuits but dont know how to design one. Specially dealing with power electronics.

[@anggorotriatmojo1200]

Plutonium fireworks 🎉🎉🎉😂😂😂 I'm dead😂😂😂

[@liudas5377]

That cracked me up...

[@Betterstartnow]

Very good thermal testing, elegant design too (simple)!

[@nuritepegoz6042]

ALT YAZIYI AÇARSANIZ BENDE OKURUM VE TAKİP EDERİM LÜTFEN AÇINIZ

[@ashrafkhaleelibrahim3340]

Thanks for your informative videos.. Could you please make a video for a diy 12 volt 2 amp or more UPS ❤ thanks ❤

[@yashwantkammar4687]

I have ups but it last 45 min can you demonstrate or a project on ups power supply with sine or square wave as SMPS can run on both

[@munnavarbasha188]

How to defuse bulbs currents in 2.4 bulbs chemical reactions

[@sakurojason]

where do you get these boxes? I was planning to use one of those for one of my projects.

[@gaminglegend191]

You are the best teacher I have learned so much from yur channel I love yur diy test equipments how can i see some of the schematics

[@Scyth3934]

What did you get your degree in, if you went to college?

[@infofoolygun]

🤩🙋🏻‍♀️🤩

[@Alchemetica]

Most informative

[@CABohol]

The bridge rectifier is going to circle. 😄

[@faridsafazadeh1137]

Although I know unfortunately you never answer to any question or comment but one more time I try: could you please put link about the supplies you use. For example I would like to buy an electronic box just as you used...

[@DiodeGoneWild]

The box is KM-175GTRC , I've bought it from TME dot eu

[@snipersquad100]

Nice educational video. 👍🐱

[@gkolko]

Thanks👍

[@asmbinelectro3899]

Thank you.. You are genius.. Can this power adjust currant between 0Amp to 3Amp with lm338t

[@jms019]

Wonder why you didn't use a transistor of the same kind for compensation.

[@DiodeGoneWild]

For this, you need a similar thermal coefficient, but a different voltage drop.

[@lorenzorentniop717]

Love the construction!

[@hernancoronel]

Daniel, new video! Thanks!

[@dimitarnikolov3527]

I wonder what would happen if you flip the schottky diode and remove the pullup resistor. The effect should be similar but would it perform better or worse 🤔

[@celsoneves2368]

Very good

[@d.k.9406]

Danke! Thanks! Love discrete components.

[@DiodeGoneWild]

Thank you for your support ;).

[@marillima6350]

Thanks

[@VIPINSAINI_20]

I was wondering, when you will give/build a 50V supply

[@Freeman666Gordon]

What's about single power transistor voltage and current limiting?

[@thecriss88]

I/d love to see a project for super accurate thermometer based on NTC sensor and some resistor divider (also compensated for thermal drift)

[@gitnote]

OMG! Are you Nandor???

[@marcalnavarrofrias2248]

06:00 +23s = ❤

[@gkcnalaka]

❤

[@leopiipponen7693]

Very interesting experience :)

[@Edmorbus]

Thanks for sharing

[@liudas5377]

Does the dog eat the cat's food?

[@SAHILKHAN-lu8oq]

12:55 That's a serious mistake from a Reputed supplier!! that's completely unexpected, how this document even became public!!

[@davidmckendry4491]

I love it when your cat makes an appearance in your videos so cute.

[@EasyOne]

pcb !

[@omniyambot9876]

Make a switching one alreadyyy. Like boost/buck converters with analog components. Your expertise in power is unparalleled. My circuits are always failures.

[@berndeckenfels]

Shouldyou put the compensation diode on the heatsink?

[@DiodeGoneWild]

why on the heatsink? It's meant to be the same temperature as the small BC547C transistor. That's why it's right next to it.

[@berndeckenfels]

@@DiodeGoneWild ah ok i thought the trans was also on the heatsink

[@t1d100]

Super-excellent!

[@sonatapuidokaite5388]

bloody scientist!

[@465kuba]

Thick soldering ;)

[@DiodeGoneWild]

It has to survive an apocatlypse. Sadly, the power plant that powers it won't...

[@mehmetaliozdemir_]

Why don't good videos have subtitles?

[@8Ugri8]

How do you calculate current sensing resistor?

[@DiodeGoneWild]

Here the voltage threshold is about 0.49V, and I needed to limit the current to about 0.25A, so 0.49V / 0.25A = 1.96 ohms.

[@8Ugri8]

@@DiodeGoneWild i am confused. Ube should be 0,6 -0,7v how you gonna open transistor with 0,49v?

[@DiodeGoneWild]

The schottky voltage drop is added before it goes into the base.

[@8Ugri8]

@@DiodeGoneWild ok, thank you! I am trying to get my head over it🤔

[@Kris_M]

Potted pots? Why's that?

[@DiodeGoneWild]

no dust getting in ;).

[@liudas5377]

For the plutonium protection...

[@anonymoususer6448]

"the most scientific solution of the year" Love it ! :-) kzbin.info/www/bejne/gpjOpaiHrpqoqbM now I remember why Visicalc was invented (for the not-so-old people: visicalc -> lotus123 -> Excel)