DIY BenchSupply [2/2]
Вставка
- Опубліковано 15 жов 2024
- #189 In this video I continue working on my power supply by building and making the first tests; as with any design, nothing is perfect to begin with, some tweaks are necessary here and there, but at least it works.
Part 1: • DIY BenchSupply [1/2]
Part 2: • DIY BenchSupply [2/2]
Circuit schematic: drive.google.c...
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
![DIY BenchSupply [1/2]](http://i.ytimg.com/vi/ZZFIRvxPbyQ/mqdefault.jpg)
"Nothing bad happened...YET"
Love it! Makes me remember every time i made a circuit :)
Neat little unit. This would work just fine for most of my needs.
Now then, lets see what I have in my junk box.
There is always something interesting and useful there, just waiting :D
Nice tinkering and many thanks for showing the noise measurements.
Great video. Well explained.
Here's another trick. When using a toroid transformer you can extend the secundary winding by adding some turns manually. This will raise the voltage and can be used to compensate some losses, like the rectifier drop. That way you do not have to use an active rectifier with maybe extra cooling, it is cheaper and you can regulate the extra voltage yourself by the amount of turns. The maximum current will be a bit lower though. This trick does not work for EI transformers though.
Some try to use schottkey diodes to minimize losses, but there is a caveat, the voltage drop of those is about the same as normal rectifier diodes for high currents. Like the current peaks through the rectifier diodes. And these currents can be really high, like 10A or 20A depending on the load.
If you take a piece of wire and make 10 turns through the toroid. Then measure the voltage and devide by 10 (amount of turns) then you now how much voltage extra you get for each extra turn.
Looking forward to many more videos where you continually upgrade this design !!!
Some UPS units use a relay to switch in further transformer windings when the mains voltage sags. Just putting that out there.
Build the Bridge Rectifier with individual Schottkey Diodes. Parallel each Diode with 1 or 2 extras.
Add another 10,000 uf 25 ~ 35 volt, and at least 1 Mohm bleeddown Resistor.
I would rewind the toroid also.
There is one potential 'gotcha' in the circuit. The current for the -5V rail is actually sourced from the +18V rail via the loads on that rail, returning via the 470uF capacitor connected to the transformer. You have a 10K resistor minimum load but that will only provide 1.8mA, which may be enough for normal operation but perhaps not when actively discharging a large load capacitance when there is no load current. However the panel meter probably provides a load of perhaps 20 to 50mA (depending on the number of segments lit) which is likely to be sufficient for the -5V current demands for this design but replacing it with a low power LCD meter for example, might produce unexpected results.
470uF is looks to be much bigger than needed; 47uF would provide over 20mA, saving space and cost.
Блок питания сделан своими руками, эта работа заслуживает уважение! Видеоролик интересный и информативный! Спасибо!!!
Thanks for this video.🇷🇴
As always with your videos, it's a nice work ! However, there is one thing missing in the tests you're showing. I built since the 70's a number of linear PS and I noticed that they are prone to oscillate at high frequencies. The output cap will suppress most of this oscillation as seen on a scope at the output but it may be present inside the loop with quite a high level (try to probe at 10x inside the feedback loop e.g. at the output of the error amp). It's always a matter of frequency compensation and you may have to tradeoff frequency stability vs transient response.
Excellent explanations. Its funny how a real life design could deviate from the theoretical idea.
A bench power supply seems so easy to make. But I never succeeded to build one without a silly behaviour 😁
Felt like you had to contain yourself a lot and was held back in this project.
Hi 👋 I need to have multiple sepersted power supplies(separated grounds) in my designing at ltspice like multiple batteries but idont know how do that ??
Yea those diodes need to be changed or maybe you can replace 2 with fets in a quasi synchronous setup.
Also a proper PCB would reduce noise. regarding the 30mV pp noise i am not sure where that comes from but i think it is a GND loop thing since if i remember well you face a few points where you connect to the case.
On a side note where did you buy that transformer from? and about how much was it? (parca erai din bucuresti? )
This may be pretty straightforward, but how is the green LED circuited in? I know its tied to the output, but is it just simply attached in in parallel with a series 1k resistor? Would the current / voltage in the LED not fluctuate along with the Vout?
It is supplied from the DC voltage after the bridge rectifier (before the regulator circuit); so the supply voltage is more or less constant.
What I found as strange is that the unloaded output was loosing 1 Volt drawing 200mA, and 3.6V at 1Amp.. That seems to me like a lot for a regulated PSU. You explained that the output of the transformer will get lower under load and we cannot do a lot about it. BUT... that is exact why stabilised power supplies are made, to compensate for all the internal resistance of transformer and diodes. So I think this loss should not happen in that amount. Having no experience with switching supplies, and having no idea what design you used in the end, I can't explain it.
Another small problem with the design is that you used a bit thin wires for the current path, AND used a switch in that path that has doubtful resistance for such current path. Could you not cut off at the base of the output transistor ? As you suggested, your voltage losses will probably also be solved with the use of sense connections on the output, but at the base it seems better to use a low resistance output path in the first place.
And finally, the heat from the diodes will be halved when you use Schottky diodes, but that's seems obvious.
That's a very satisfying answer. I was thinking that the max. voltage was already set by a reference.
Very nice design! But the rectifier bridge is a bit too toasty imho, would definitely change them to the schottky type. Did you take the top cover off right before taking the temperature measurements or was it open all the time? This can also skew the measurement a lot. What about the transient response, can you measure the loop stability?
Indeed the PN type diodes where not having too much fun, but I did change the rectifier; I just started off with that particular one; Regarding the top cover - it was always off, so the airflow was even better than normal and it was still getting extremely hot.
It reminds me the painful days about, there is no constant current section, right? Mine is a fking oscillator under constant current mode........I never figured out why.....
waitung for the fix on the outpu! the voltage should not drop when there is load. please make a video