EEVblog #224 - Lab Power Supply Design - Part 3

[williefleete]

shoot, thats pretty low noise, the one i rolled myself using spare parts, an LM324 and external series pass transistors does about 50 odd mV of noise, i didnt have my own Oscope but i used a PC based scope, it seemed to measure 50 - 100 mV above/below 0 (AC coupled), pretty horrible compared to the 3080 but not bad considering i cobbled it together

[janslos]

@aseglkj Yep, I have looked at your channel, nothing much to learn from there. There are 2 kinds of people, the ones who like to share what they know, even if it is a bit odd or not totaly as it should be, and the others who only critisize the first ones but have nothing to contribute themselves. I personaly like the first ones, but hey, that is my opinion.

[aladaris]

and... Bang!

;)

[envisionelec]

I love the LT3080 - great component with some surprising features. I use it a lot.

[zega74]

It is nice (personally i would not use LM's but that's me)

Whole OP-amp for current limit LED?!

Just hook up LED in series with base of current limiting BJT, lower series base resistor, and add one resistor between base and emiter of current limiting BJT and voila.

I've found i like 100microamps trough 5mm clear hi eff red led just fine, more then 3 mA is too bright for me, so choosing right resistors you can get it to work nice!

[aseglkj]

What a horrible design. You could've achieved and in fact exceeded your requirements by using a dual rail-to-rail opamp and a NPN darlington, without using some single-source, expensive, hard to find super IC, at 1/4 the cost. In fact, you could've used the same freaking TLC2272 and replaced the useless LT3080 with a TIP120. Also, the thermal resistance of the pass element would be much lower.

Also, please stop tapping expensive instruments with your pens. It's really annoying.

[EEVblog]

@aseglkj Stability across a wide range of loads is a major consideration with a roll-your-own solution. See a previous video I have done on exactly this solution.

It is not a horrible design, I deliberately traded off extra cost for development simplicity and guaranteed stability.

[envisionelec]

@aseglkj Could you be a slightly-less intense dickhead? Why not design your own competing supply and sell it head-to-head (if you're so smart...). What's that saying about spending a large fortune to make a small one? Familiarize yourself.

[aseglkj]

@envisionelec actually, douchefag, I did design my own bench supply 10 years ago when I was still in university. Still works like a charm. 0 to 30V and 0-3 amps, with a bunch of opamps and two 2N3055s. Parts cost is less than 10 bucks, not counting transformer, heatsink, case and meters. it's a standard exercise for anyone who learns about opamps and many hobbyists do it. but you wouldn't know, since you don't know jack and are talking out of your ass.

[Shroommduke]

@aseglkj i'd love to see the schematic, maybe you could do a video illistrating explaining and showing it in action.

I also encourage you to learn some social skills... good luck

[D4nnyTh4B0y]

@aseglkj Wow, you are an ignorant person...

[williefleete]

i've had a little play around with an old school bench supply, used the original Xformer but gutted the late 60's era regulating circuits. i have managed to get reasonable results using a series pass transistor and the "jellybean" quad opamps. not current limited but i do have a constant current mode with preset current limits and voltages using the radio buttons and rotary switches the old supply had, 0 to near 15 volts with reasonable regulation and less than 2 amps

[SeanBZA]

The bounce at the beginning is due to the output current sink turning on. Nothing due to the LT part, just the dropout and fast turn on of that old current sink.

[tsaavik]

It's interesting how the scope registers the tap to the screen at 38:50.

[randomgarfield]

@1710WL get it right! it's 1.21 jigawatts. ;) hehehe

[TheWelly888]

Thanks for the three parter! I built my own lab bench psu several years ago from a Electronics and Wireless World magazine article but I could not properly understand how it worked. Now you have given me a better idea!

Strangely enough I have recently been given a free bench psu by my work place as long as I can source a replacement thyristor.

[LCRAMBOR1]

PDF with list of parts, please..:-P I'd like to build it but my electronics background is limited... Thanks

[Dysl3xicDog]

Dave do you have a BOM and a official schematic to work from?

[EEVblog]

@Dysl3xicDog Not yet, patience!

[chupascz]

i love these tutorials especially beginner advices. very educational.

[SajjadBro]

Hey dave.

I used that same circuit with LM234 op-amp. I am thinking of taking the output of differential amplifier into ADC and displaying the current on LCD. But there is a triangular wave at the output of differential amplifier when it was in constant current mode. Is it the slew rate which matters in this case? + when i replaced 1Ohm resistor with 0.1Ohm then there was too much noise across that resistor!!!

[Jprojects1]

Bang!

[ak47only]

I have a new drinking game: you have to drink every time Dave says BANG!



[randomgarfield]

@ak47only way ahead of you there.

[tuberlook1]

I have to say I thoroughly enjoyed this series, my other fav was the current source episode so it was nice to see that back in action, an excellent piece of kit. Anyway just a bit of text to say thanks Dave and great work.

[techgood]

Amazing Video, Merry Christmas Dave!!! (in Advance)

[bswain9999]

The original spec was for 6V but now we seem to be maxing at 5.5V... :/

[EEVblog]

@bswain9999 Depreciation, the global economy is tanking, didn't you hear. Gotta save all the volts we can :->

[SeanBZA]

@EEVblog Limited by the reference. Up the unregulated rail to 12V and change the opamp gain to 3 and you will get 6V out, just change the one resistor from 22k to 27K will probably do it.

[Shroommduke]

@EEVblog "saving volts" a metaphor to the global ecoonomic paradigm and the absurd cries for nations to save the ethereal money that they create and distroy at will!

your blowing my mind man... volts, money, wealth amps power. M*W=P P/M=V ...

OK, I'm going off the deep end...

[LeonValkenborg]

Hey Dave,

Im Loving this "Lab Power Supply" series more then any other.

I have considered building it myself, and I came to the conclusion that the amperage and voltage that you have based this series on is a too low.

If it was ~20A and ~35V, I would have built it by now.

You are working with such a low voltage and amperage design that I just cant bother with taking the design to that stage.

I wrote a WHOLE LOT! of massively positive things, but apparently I have a limit to how much i can write.

[randomgarfield]

@LeonValkenborg Dave already stated he has no interest in high voltage and high current PSUs. Seriously though how often do you require more than 12V and an amp when you're breadboarding something. However his design does scale (although expensive). If you want 20A, just use 18 LT3080s. :) They'll also go up to 30V without issue too, but you'd be really silly to scale it that large.

[laharl2k]

@randomgarfield for higher current its better to use a switch mode design as you wont need a truck sized heatsink, further more you can control both voltage and current in the same stage if you use a micro or a dedicated ic.

[randomgarfield]

@laharl2k I wasn't being serious about scaling this up. Although you can make very large linear supplies that don't require huge heatsinks if low voltage output isn't a requirement.

[EEVblog]

@randomgarfield @LeonValkenborg Yes, there are countless high current/high voltage linear and switchmode PSU's on the market, and they are cheap as chips. I'm doing something different, which will become clearer with time...

[randomgarfield]

@EEVblog does it have a flux capacitor by chance? :)

[DanFrederiksen]

@EEVblog nah switchmode supplies are not that common and often well over 100$ and often the same in shipping.

I think you could have a steady stream of new very interesting episodes exploring various design branches, switchmode, digital, simple buck and cost optimization through fancy resonant switchers like Vicor supplies etc.

seems to me that a 30V 10A supply from mains could be the size of a pack of cigarettes and weigh very little.

and could end up as a popular commercial product for you...

[ZitroOgu]

Hi Dave, can't you connect the LT3080 Vcontrol pin before the 1R shunt to lower the dropout voltage at higher currents?

[EEVblog]

@ZitroOgu You can, yes, but only if you have the package that has the pin separate. Keeping lowest common denominator here. Forgot to mention that.

[TheUrdyggradsil]

This is an excellent series of videos! Thanks for sharing them...

It's really useful to see the whole design and test cycle. The whole project has demystified the 'how to make a PSU' to the point where I think I understand it clearly - rather than just using a 7805... so, now I'm off to order an LT3080, comaprators and 1R resistors :)

[lyrez62]

Amazing video, maybe one of your best ones , ever ! Thanks !

[rotlerin]

I only wish that there was someone with your knowledge and personality in the electronics repair indusry. I am familiar with Jestine from Malasia and Dave from Oregon, but apart from that no one seems to come near your standard. For a newbie like me 70% of this great video soared over my head.

[SigEpBlue]

Nice work yet again, Dave. I'm reading data sheets in your voice now. 8)

[DanFrederiksen]

well done videos. I hope you'll do subsequent videos with digital control and switchmode instead.

a good power supply doesn't need a 10 turn pot if it's digital.

[4833504F]

with the added uc this design would be great hardware for a diy single cell lipo charger :)

[EEVblog]

@4833504F You can already buy those for $5 :->

[4833504F]

@EEVblog lab power supplies as well ;)

i'd like to combine power supply and charger for all kinds of accumulators

[sifterjoe]

Dave, great tutorial. I was wondering if you would consider doing a short video showing how the PWM would replace the pots for the limit control voltages. Obviously PWM from the duty adjust on a signal generator.

But then I was hoping you could show someone how to do the same thing with a simple opamp 4 bit DAC (r*2^n opamp adder) that is controlled by DIP switches or a shift register. Just to show what you can do for controls if someone wasn't using a PWM for the reference voltage.

[compaq1275]

Thanks Dave! I've been wanting to make a power supply for some time now, and I think I have a design to work with now!

[abpccpba]

Well you have found your place in the web. Your banter has near the same tempo as a baseball announcer and color man in one.

Well done.

[Th3Su8]

This series of videos has been well planned out and very informative. Thank you for posting these. I really do appreciate the effort you put into your videos. When you get around to doing the video for the final project I would hope that you also include some information on "the traps for young players" when designing the layout of the PCB. I will be looking forward to more of these educational videos from you. Thanks

[EEVblog]

@Th3Su8 My videos are almost never "planned out", these ones included. They kinda "just happen" :->

[williefleete]

alright i might have a go with this supply, i wont have the regulator or the opamps, (i have got 324's though) i might want to find a site that gives free samples or similar for some of the parts and i can probably scrounge some of the passives

[srhenderson186]

Wow, Dave. It's really special to see this latest series of vids. Comparing where you started to where your at now, I can discern no less determination, charisma and honesty than that first vid... but without the distraction of that day job, well, what your doing here is really meaningful and helpful. As a father, I'm certain that Sagan will be proud of what your helping people everywhere learn to do (both in terms of EE, but more also generally: how to help other people and make ends meet too)

[KozmykJ]

I've really enjoyed following this process.

Thanks.

[menushadesanayaka255]

Thanks Dave!!...

Hoping to see more videos like this in the future ......

[Wilfredkazoks]

Well done Dave, you are really getting the presentation to a high standard. Reminding us what the traces, readings and scope settings are really makes it easy to follow along.

I just wish you would set the exposure to manual. It is distracting (to me at least) to see the paper come in overbright and then turn grey, followed by the background lightening up when you remove the page. Since the light level in the lab is fixed there is no need to have the camera setting the exposure automatically.

[EEVblog]

@Wilfredkazoks The light is not the same from all angles, so it constantly has to be changed, and my camera has a problem where it automatically jumps back out of manual exposure mode to auto after returning from playback mode. I do try and put it on manual exposure when possible, I forgot this time with the paper, sorry.

[heroineworshipper]

Amazing such a complicated design could work on the 1st try. A linear power supply is useful for audio.

[randomgarfield]

The EEVBlog drinking game. Dave says "Bang", take a swig. You'll be drunk in the first 10 minutes and in hospital before the end of the video. :)

[fingerboy18]

Wow, that DaveCad-XL is impressive!

[superdau]

Why did you use a second opamp for the limit LED (except for the fact that it comes free with one of the packages already in the design ;) ). Couldn't you just hook another transistor to the output of 1b to drive a LED against V+? The output of 1b seemed to jump to at least 1V as soon as current limit mode kicks in. If it's the part count I would not have hooked up the second opamp in parallel, but have it compare the output of 1b to lets say 0,5V.

What did I miss?

[EEVblog]

@superdau You didn't miss anything, the fact that it was available led me to use it, and it was helpful to show show the issue with identical comparator inputs. You can certainly use another transistor, and that's a more accurate way to do it.

[horfittunge]

Wow, this must seriously be the best tutorial I've ever watched. Super great job Dave!

[EEVblog]

@horfittunge Thanks. Interesting in that it's not even supposed to be a tutorial, it's just me breadboarding a design and doing a few tests and yapping on while doing it :->

[horfittunge]

@EEVblog In either case, I love it!

Keep it up.

[EEVblog]

@SuperEvilbug Yes, this will hopefully become a kit. The final design I have in mind will be more than what you see here, and rather novel I think.

[electrodacus]

So 37:10 is how you found high resolution mode? :)

[EEVblog]

@electrodacus Yeah. I thought it was a good example of how it can hide things, so quickly shot a separate video.

[Microman171]

What disadvantages are there to using a discrete voltage regulator in place of the LT3080? You inspired me to have a play, so I built one from an LM358 and an N-Channel MOSFET.

[EEVblog]

@Microman171 The LT3080 is designed to be stable. If you roll your own, you have worry about stability over a wide range of loads. May or may not be a problem for you, but something to watch out and test for.

[xng14]

Hi Dave!

The design videos are great! Definitely keep them coming!

What's next? pcb and case design? panel meters?

[EEVblog]

@xng14 Yep, I have a nice, and I think kinda novel build coming up, will be a while though...

[Afrotechmods]

Beauty!

[nomione1]

@Afrotechmods : Bangg ! 

[turbochargedbrick]

Do you often draw schematics on paper? :)

[scompo]

@turbochargedbrick it's davecad!

[turbochargedbrick]

@scompo Paper is too large and not yellow!

[Mtaalas]

It's wonderful to see that you're doing so much work every day. This really is an example of "how it should be done". This will be the quality I'll start to want from everyone who even desires to make any tutorials about subject x :D

Without people like you Dave, YouTube wouldn't have any quality videos. Keep these coming :)

[marcelojr]

Great video!!!