I found it difficult to understand the usage of opAmp but your schematic made me break the barrier. Although am not yet saturated with opamp.. your project help so much
Wow what a great project. Must be very satisfying to have designed and built it yourself. I love the mixture of old bits and new technology to construct your project of very competent construction. Wonderful well done, enjoy.
Thank you, I'm glad you like it! It is indeed a good feeling to build something yourself, but there is always something you wish to improve. Now I'm working on a second version of a linear power supply which hopefully will be more professional and easier to recreate.
I made one using one uA741, one LM317, and one Tip 35 and it can supply a max voltage of 24 volts and around 3 amps. My aim is to make a good homemade regulated power supply as cheap as possible.
I love these DIY PS. Nice done 👍. A long time ago... 30 years or so... I also made a DIY power supply. I used a motor-controlled variac as the input voltage, to minimise the dissipated power. The voltage across the control transistors is constant (automatically controlled via the motor) so also the dissipated power of the transistors is constant regardless of the output voltage. So in addition, I also had a 4A 0-280VAC, 40A 0-25VAC, 20A 0-50VAC and stabilised 40A 0-25VDC or 20A 0-50 VDC. Made my own current measuring resistor (hand wound resistor) for the sake of the 40A. DC was with adjustable current limit. Now it's in the basement and is nothing more than an anchor... (Oh how heavy that thing is 😂)
Thank you! That sounds like a great solution (ignoring the immense weight I can only imagine it had xD)! I wander how it reacted when the output voltage changed rapidly, especially when it goes up. I guess that the motor-controlled variac didn't adjust very fast (or did it?). I see that this problem would be ocurring especially in the constant current mode, when the load rapidly decreases. Or maybe it's not a big deal and the power supply handled it well?
Good job thats a nice design! Can we get more details from the current limit circuit because im trying to build something similar and im having trouble
Hi! Great video and amazing project, very interesting. I would like to ask you for an explanation: I struggle to understand the function of the P-channel mosfet and how it decreases/increases the output voltage
Unfortunately I haven't performed precise measurements of these parameters and there is no active power factor correction in this power supply, tho it might be a good idea to implement a passive PFC.
So the 5 V regultors are for voltage reference for opamps right? What about stability of the outpout with a big load? How much your outpout voltage drop let's say at 2 amps current for example?
Yes, the 5V regulators are currently used as voltage references for the op-amps. However, this is not the most optimal design choice. Using a dedicated voltage reference like the TL431 could provide better precision and stability. After replacing the P-MOSFET with a BJT, the output has shown stability even under heavy loads. More details on this change can be found in the README file on the GitHub repository. While I haven't precisely measured the output voltage drop at a 2A load, I estimate it to be less than 10mV. As demonstrated in the video at the 5:39 mark, there's no noticeable drop in output voltage even at a 3A load.
10 mV is very good as with common l317 and simple design it can drift (at least in my case) about 180 mV or something in both ways which is quite annoying while charging LiPo batteries 🙄 I have to monitor the current all the time...
Hi Dominik, your project is exactly what i need " Linear power supply" that can deliver my requirement of 24V 10A. Are there any updates / new developments, between your clip date till now ?.
The only thing I was trying, was to increase the stability of this design. In simulation I figured out, that replacing the Q1 MOSFET with an PNP BJT improved the stability, though I haven't tested it in real life. For someone that has some experience with electronics I recommend to try this modification and check if the stability improved. [EDIT: I tested it and it worked wonderfully!] In addition to that, I'd like to mention two easy performace improving mods - using some precision voltage references (e.g. TL431) instead of two 7805s. And I don't recommend using a switching converter for the fans, because it introduces some high frequency noise to the output.
It could easily withstand 60V but you would have to modify it slightly. On the linear regulator board there is a simple linear voltage regulator built on one BJT, you would have to make sure it doesn't overheat and that it can take over 60v on the input. And of course some feedback resistors would have to be changed, and transistors should have appropriate voltage rating.
I am no control loop expert, I struggle with this in my own designs. But, from what I do know suggests your loop gain is ultra high with the BC639(which op-amps are not designed for, if you want to keep decent phase margin to stop oscilations). I assume the 1nF comp cap on the LM358 half being necessary because of this? ...or am I missing something? Slow voltage response? You have a 10uF output cap so I assume not, but how? BTW I love someone for ONCE designing a supply with a GIGANTIC output cap, which craps all over CC control. Would LOVE TO SEE transient response scope captures for your design :)
Yes, this 1nF capacitor is to stop oscillations. Do you have any suggestions to help with stability of this circuit? I'd be glad to receive some advice, I had a lot of trouble with this circuit oscillating. I might do some transient response measurements in the future ;)
@@dominikworkshop6007 OK, that answers it. I expect the stability problem is because of the ultra high gain you have in your loop. One suggestion I have is avoid external voltage gain. In your circuit this can be done by using an emitter follower to drive the output tranies DIRECTLY (get rid of the IRF9530). Watch your power dissipation and SAO (safe area of operation) for the driver trannie (the BC649 will not cut it). Something like a BD135/7/9 is very cheap and ok for this. You likely will need a small heat sink on it. Doing this is easy in you circuit as once ther IRF9530 is gone the driver tranie as an emitter follower (i.e. collector to input voltage, emitter to base of output pass trannies, is at the right phase. You may need to lower R7 to 1k or so for more current because the BD139 has a low gain (but it is cheap and jelly bean common). You may need to put a schotcky diode (the cheap and also jelly bean common 1N5819 will do) from the BD139 emitter to the output base, to protect from reverse input voltage (which absolutely WILL blow up your op-amp, should it ever happen, remember reverse Vbe is 5V or so on all trannies). If you don't already have/use the free LTSPICE,simulator (just download from analog.com) , you can try your design without rework until ready. I suggest you start with your current circuit and see how the stability compares to your actual real world build. ...do keep in mind that a simulator gets you in the ball park, real world tweaking/testing still required. The BD___ trannies are not in the LTSpice libraries, just pick similar ones from what they have to get you started. You will have to get the LM358 model from a download ...use google to look for "LTSPice LM358 model". I made a suggestion about your voltage reference in another comment, if you are interested. BTW, I love the way you did the whole design, MCU, software, and all. Many do not bother or sometimes even know how, ...hat of to you :)
@@dominikworkshop6007 In your transient response testing; if you see ANY ringing, then it is likely that under some load conditions the PSU will oscillate ..just something I learnt the hard way in one of my designs that I thought was great, but as it turned out, no so much :)
To set the current limit, you can short the output of the power supply while measuring the current. Then you can turn the current adjustment potentiometer to regulate the output current to fit your needs. But if you built it correctly and calibrated it, then on the LCD will be displayed the set current value, without the need to short the output.
Mała uwaga dla budujących kupiłem tranzystory bc639 i nie działało po dokładnym sprawdzeniu okazało się że tranzystory mają inne wyprowadzenie Pętanie do was panowie czy można podwoić wszystko i zbudować zasilacz w kanałowy i ewentualnie łączyć kanały aby zwiększyć napięcie ? Czy będzie zwarnie przez Arduino jeśli tak jak to rozwiazać
Można zrobić wielokanałowy z możliwością łączenia szeregowo, ale tak jak piszesz, controll board by to zwierał. Rozwiązać to można by było stosując izolator I2C do przetwornika ADC, i mieć połączenie między płytkami tylko przez I2C. (zasilanie ADC też nie może być brane z controll boarda oczywiście)
@@siedshojaat4473 The output of this power supply is galvanically isolated from mains earth, so it is possible to for example, connect other power supply in series with this one
You would have to choose Q1 and power transistors that can handle that voltage and current, and adjust R12 from the voltage feedback voltage divider. And of course you would need a transformer with higher output voltage and possibly more tabs.
Once you are up into voltages that high you have to be extremely careful with safe operating area for the pass transistors. Second breakdown becomes a big concern. It is not unusual to use an array of cascoded transistors to deal with the second breakdown. The cascode arrangement essentially splits the voltage drop between two transistors). It is not a trivial undertaking. You would also need massive airflow for the heatsinks. If I were doing such a thing I'd seriously consider using power MOSFETs as the pass devices. There aren't many FETs on the market that are optimized for linear circuits and they are quite expensive, but have some definite advantages.
@@kudaty5535 Eksperymentowałem z wartością R11 i C2. Warto również wspomnieć że jeśli tranzystory mocy są u połączone jak u mnie przewodami, to długość i ich położenie również może mieć wpływ (pasożytnicze pojemności i indukcyjność), ja się starałem aby były jak najkrótsze.
@@dominikworkshop6007 Właśnie próbuję to naprawić, ale wyszedł mi jeszcze jeden problem. Przy większym obciążeniu spada napięcie o ~200mV W trybie CV(to nie wina transformatora).
Myślałem, ale chciałem żeby mieć jak najczystsze napięcie z zasilacza. Możliwe że kiedyś zbuduję warsztatowy zasilacz impulsowy, ale to już wyższa szkoła jazdy.
@@kudaty5535 @Kudłaty W najgorszym wypadku mój zasilacz musi rozproszyć około 150W, ale to się rzadko zdarza. Budowa zasilacza impulsowego może być prosta jeśli chcę sie użyć gotowych modułów, wtedy wystarczy zasilacz impulsowy AC-DC, potem przetwornica DC-DC i jakiś układ pomiarowy. Ale ja jakbym budował to bym próbował zbudować swój zaliacz AC-DC z regulowanym napięciem i prądem, bez osobnego modułu przetwornicy DC-DC wtedy jest potencjał na lepszą sprawność energetyczną, ale to jest nieporównywalnie więcej zachodu.
@@kudaty5535 Zależy co masz na myśli. Jeśli chcesz zasilić układ regulatora liniowego napięciem stałym 30V to nie musisz nic zmieniać, możesz tylko usunąć układ prostowniczy oraz układ przełączania odczepów transformatora bo byłby już nie potrzebny. Jeśli chcesz po prostu użyć transformatora o napięciu 30V to musisz zwrócić uwagę na tranzystor T2, bo może się mocniej grzać.
I use U to refer to voltage and V as a unit of votlage [Volt]. I thought that everyone does this, but I checked wikipedia now, and it says "Voltage or electric potential difference is denoted symbolically by V, especially in English-speaking countries or international U".
Why such good work and use a 7805 as a reference? At least put 100 ma. load on the 7805 for stability and a ceramic capacitor. I don't like the fet used in resistive mode to feed the pass transistors.
I used a 7805 because it's very cheap, and reasonably accurate. There are ceramic capacitors, but I guess that if I load this IC with 100mA it will heat up enough to change its output voltage significantly. Why don't you like the FET there? What do you suggest instead? Thank you for your feedback!
@@dominikworkshop6007 I assume Daves is referring to the issue of ultra high gain ...Daves??? I would love to know your thoughts on this, because I stick to 60Mhz, which make loop stability a small nightmare at times for me). Most commercial supplies also use slowish op-amps like TL062/072/082's. Your design uses an even slower op-amp? Is this due to low cost/availability, or was this by design for some other reason? BTW for the reference, you may want to consider a TL431? ...also VERY CHEAP but much more accurate, 0.5% to 2% depending on grade (and only requires a 1mA load).
Thanks. I added Eagle files and pdf files of 3 pcbs on my github ( github.com/Dominik-Workshop/Linear_power_supply/tree/main/PCBs ). These pdf files can be used to transfer traces onto PCBs if someone wants to make them at home as I did.
Hi Dominik, Is there a way in which I might be able to contact you? My email address is in the about section of my channel if it would help to get into contact with you please? **Note** you will only be able to see the address if you are looking from a PC browser, the android/ IOS apps won't show the address.