DIY BenchSupply [1/2]

#188 In this video I try something a bit more different than normal. Rather than talk about a specific technical subject, I want to take you trough my quest of building a power supply for home use. Its nothing particularly special, but it does not have to be, it just needs to work and be reliable.
Part 1: • DIY BenchSupply [1/2]
Part 2: • DIY BenchSupply [2/2]
Relevant links and datasheets:
www.ti.com/lit...
www.analog.com...
Special Thanks to all my supporters on Patreon!
If you liked this video be sure to check out my other videos and you can also subscribe to be up to date with all the new ones!
If you want to support the creation of more and better videos please consider checking out: / feszelectronics

[@Cynthia_Cantrell]

Too much noise in a switcher output is a common problem on boards that also have an RF output, or any clocking in the GHz+ range. The problem is that the clock multipliers in those chips typically have a bandwidth high enough to include the switching frequency of the power supply. That PS noise is then multiplied up by the same factor N that is used to multiply the low frequency reference up to the output frequency in the GHz range. Unfortunately this can add a HUGE amount of jitter onto the output clock, making it unusable for many applications. Sometimes a linear regulator can be added to clean up the switcher output, if you can find one that has enough loop bandwidth to cover the full range of the output noise. This isn't always a good solution however. If the switching noise is entirely triangular, a coil inductor between the PS output and the main filtering caps does a nice job of cleaning up the output. Choose the inductance to put the corner frequency a couple of decades below the PS switching frequency with the output caps. Be careful here though - most regulators can't handle the extra pole this inductor adds to the feedback loop - so take the voltage feedback point directly from the PS output BEFORE the inductors, not after them to avoid turning your PS into a big oscillator! On some switchers, you'll see large, sharp spikes at at each ramp change in the triangle output waveform. These are the output diodes turning off and ringing. Unfortunately, linear regulators and the coils above will just pass that high frequency noise (typically 20MHz & above) right through them. For this, you'll need to use large ferrite beads (i.e. large enough to handle the output current) right at the output of the switcher - BEFORE any coiled inductors. You'll want some small caps after these ferrite beads too - on the order of 10 to 100nF - small enough so that their SRF is still above the diode ringing frequency. On PCBs, the effectiveness of this RF filter is VERY dependent on layout, so keep all of the leads in this section as short as physically possible!

[@soothcoder]

You can always add another winding to the toroidal transformer. Not hard to do. Less likely to add noise to the output of the supply I think than the negative supply you have now.

[@FesZElectronics]

You are right, for this particular transformer that would have been a possibility - since the middle bit is accessible. Oh well, its a good idea to keep in mind for the next project :D

[@AdityaMehendale]

Since you are using a full-bridge-rectifier (as opposed to center-tapped or whatever) there is is no absolute "ground" point - you can choose to call any potential zero. If you only need a few V below zero to bias the output-stage, why not just use two or three power-diodes in series? 0.6V*3 = 1.8V which would be plenty to not get funky artefacts around zero. Its stability would be reasonable, and final output-voltage is determined by your feedback-network anyway (referenced to your "lifted" ground).

[@PainterVierax]

Nice educational project, I like the way you explain your train of thoughts and your modular approach. You choose a nice case to play with, though I'm not confident about TO-3 packages. Parts are hard to find, especially when shopping for genuine replacements with the same power characteristics. And even with a proper insulator on the radiator, its top cap remains exposed to hazard if not at the ground potential. So for those reasons I would rather modify the radiator to internally mount some TO-220 instead.

[@tasoszog]

"Not the greatest regulator by today's standards" agreed, but LM723 has been used in some super low-noise regulators of yesteryear with some immaculate noise performance! It's all related to the option of having the reference voltage being filtered out externally, an option not many IC regulators offer.

[@PhG1961]

Nice video! However, I'm still very satisfied with my DIY power supply from Elektor from the early eighties (1982). It's power output uses the famous 2N3055's which are (almost) indestructible.

[@Wtfinc]

I friggin love this. It has so much to teach

[@pieters286]

Recently aquired laser engraver to speed up pcb diy - found breadboards, veroboard prone to increase debugging effort. Looking fwd to the testing vid. Keep up the very entertaining and educational content!

[@catalinm756]

I did not know you are from romania.Where have you studied electronics engineering?

[@alaricsnellpym]

I'm taking a similar modular approach to an audio circuit I'm working on right now - but building each stage on a breadboard, before moving it to a PCB and then the next stage on the breadboard - joined by soldered wires in my case.

[@FesZElectronics]

Sounds like a fun project! Especially if you are designing it from scratch, there will always be problems, so the breadboard first, pcb second approach should ensure the pcb's are working with minor later modifications.

[@alaricsnellpym]

@@FesZElectronics yeah! My project is a microphone preamp and distribution system so my one mic drives two PCs and whatever transceivers I attach. And I'm putting adjustable mic gain, tone controls, and a global mute between, so there are plenty of well defined logical modules in the pipeline.

[@nsfeliz7825]

its a FOOOLL BRIDGE RECTIFIER😆😆

[@Mr.Leeroy]

For my Linear PSU project, I found it to be worth having auxiliary low-power PCB transformer for control circuitry. They are not terribly common at decent price (same as power transformer actually), but do exist and are very doable in single project quantities. Snatched some HAHN 3W 2x15VAC beauties at $2/piece.

[@romancharak3675]

Fascinating Topic.

[@iblesbosuok]

CA3140 can handle _Vicm_ down to its VEE. I apply it many times.

[@user-xy5bb3ld3t]

Keep up the fine work - it is much appreciated! I am a huge fan of the TO-3 package as well, perhaps it’s nostalgia for me.

[@FesZElectronics]

I don't know why, but the way its built - the circles and curves just make its design stand out; they don't make parts like they used to :))

[@Rudy7408]

Your shirt matches the project box color, intentional? Nice little project!

[@FesZElectronics]

I'm not a very artistic person, but I was told that keeping things color coordinated makes the visuals look nicer - so I try to apply this principle when possible; this time though, I think the color matching was by accident

[@Gengh13]

I have an idea for a future video, you can try to make a lower frequency active PFC on the secondary, you avoid high voltages and maybe improve the maximum power available from the transformer.

[@Betterstartnow]

I enjoy your videos, thank you. Question; do you know where I can locate an amulet that might keep someone from turning into a werewolf (asking for a friend).

[@ohmware2020]

subscribed! where is the part 2!?

[@asciithesaint]

Why just don't adjust the output with reference intead of feedback reference ?? Is there a problem that I don't know. I know for exapmle with buck even though ref voltage doesn't affect compensator calculations res and cap values change

[@asciithesaint]

accordimg to output voltage, but with linear just need a big cap for higher power.(?) I did this in a non efficient way driving though. Is there something idk?

[@asciithesaint]

kzbin.infoChXJfbEDnsw?feature=share

[@grozaadrian4134]

oh boy oh boy ! 😁

[@sefalibhakat143]

I was waiting..

[@Vitriolguy]

You want to soak away 28 V with a regulator to get -5 V rail? You better cool that thing, critical design choice.

[@FesZElectronics]

If the current is small, the large voltage difference is not a problem; since I only expect a few mA, there should be no thermal issues... maybe

[@Vitriolguy]

@@FesZElectronics I am really curious how this turns out. I really like your content. Keep that up.

[@DynoRC]

Bro be talking way too much

[@msana4420]

then go watch those other videos where you see hands and loud music in the background. smh

[@davidhyslop115]

@@msana4420 this fucking guy lol

[@peterfitzpatrick7032]

@@msana4420 Dino can't even string a sentence together properly... 🙄😂

[@N1gel]

I did thumb up as usual but have to question if you spent too long in Soudi? We read books from Left to Right and we draw circuit diagrams from Left to Right. Input at left. Output at right. High voltage at top. Lower v at bottom. Its simple, but dam annoying and unprofessional to be stupid enough to reverse the standard. A bit like Crockodile clips being wrongly named to Aligator cloips and 4x2 & 8x4 timbers being wrongly renamed as 2x4 & 4x8 by morons that havent yet learned how to pull the trousers we gave them, up much above their knees.