#MEASURING OUTPUT VOLTAGE# Ideal laboratory power supply - YOURS requirements

[MacIntoshCZ]

And the circuitboard that can handle 50A? I looked at making one recently and prices from a Chinese supplier in quantities of 1000 were in the range of $15 each.

I dont have pcb yet. But my previous forward smps (diferent topology) was able to supply 50A. Biggest problem was efficiency. It was around 80-85% at full load. PCB was not any hot. I used a lot of tin. Hottest parts were snubbers (primary/secondary) and diodes on secondary.
I will do some test with thermal camera and wattmeters, when project will be completed.

 

[ChrisP58]

i am aiming to 95% efficiency at full load. I am using phase shift fullbridge topology with synchronous rectification, at low load i am switching from ccm do dcm. At very light load burst mode is used. I believe it would be ok. Also i want all parts below 60C degrees.

I'm a bit confused. In post # 1, it definitely sounds like you're asking for recommendations for a power supply.
But here you're talking about how you're planning to build one.

Please clarify.

 

[MacIntoshCZ]

I'm a bit confused. In post # 1, it definitely sounds like you're asking for recommendations for a power supply.
But here you're talking about how you're planning to build one.

Please clarify.

I want to make power supply that would have everything,that must have power supply need. =)

 

[gophert]

I dont have pcb yet. But my previous forward smps (diferent topology) was able to supply 50A. Biggest problem was efficiency. It was around 80-85% at full load. PCB was not any hot. I used a lot of tin. Hottest parts were snubbers (primary/secondary) and diodes on secondary.
I will do some test with thermal camera and wattmeters, when project will be completed.

What kind of load were you supplying that took 50A? Or, took 50A long enough to heat the circuit board?

At 50A, a 0.05 ohm resistance of the board and solder connections will cause a 2.5V drop on the circuit board which corresponds to 125W of heating. Even a 0.01 ohm resistance is a 0.5v drop and 25W of heating. Good luck trying to build anything that handles 50A on a pcb. It can be done but not without some intensive thought and design effort.

 

[MacIntoshCZ]

What kind of load were you supplying that took 50A? Or, took 50A long enough to heat the circuit board?

At 50A, a 0.05 ohm resistance of the board and solder connections will cause a 2.5V drop on the circuit board which corresponds to 125W of heating. Even a 0.01 ohm resistance is a 0.5v drop and 25W of heating. Good luck trying to build anything that handles 50W on a pcb.

once i tested small SSTC it draws 30A on 30V DC. I dont know many applications. Just wirreless transfer =D, or you can make hydrogen
Also 50A are provided after secondary Caps, Irms on secondary is at maximum load aprox 20A. BTW it could be always connected with wires... So i am really not aware of that kind losses. Everything will be solved, if not i am not gonna sell something that suck.But you got a point

 

[schmitt trigger]

For that kind of current, remember two words: buss bars.

 

[MacIntoshCZ]

I was also thinking about other power supply with multi chanells. Question is how much power supplies do you need? All chanells isolated.
My idea is 3 channell 50V 150W fixed for one channel, not sure about amps. 150W topology would be flyback with active snubber for zvs...
How much current is needed for output voltages 3v3 5v 12v? I guess 12V could be used for gate drivers. 3v3 5v for IC.
How much current can flow through banana clips anyway? 10A max?
wiki says 15A

 

[gophert]

I was also thinking about other power supply with multi chanells. Question is how much power supplies do you need? All chanells isolated.
My idea is 3 channell 50V 150W fixed for one channel, not sure about amps. 150W topology would be flyback with active snubber for zvs...
How much current is needed for output voltages 3v3 5v 12v? I guess 12V could be used for gate drivers. 3v3 5v for IC.
How much current can flow through banana clips anyway? 10A max?
wiki says 15A

If you're just doing 3.3, 5 and 12, you may as well sell ATX power supplies for PCs. They are not all isolated but when is that needed?

 

[MacIntoshCZ]

If you're just doing 3.3, 5 and 12, you may as well sell ATX power supplies for PCs. They are not all isolated but when is that needed?

Selling ATX power supplies will be probably more chalenging. But it will open doors for more customures. Every kid wanna gaming rig =)

 

[MacIntoshCZ]

If i wanna sell electronic device i need to pass some test and get certification. What if i wann sell DIY kit that is not working when device is sold.
Do i need still certification? Kit is aimed for hobbyst.
What do you think?

 

[gophert]

If i wanna sell electronic device i need to pass some test and get certification. What if i wann sell DIY kit that is not working when device is sold.
Do i need still certification? Kit is aimed for hobbyst.
What do you think?

Ask your regulatory office.
Some offices around the world are ok and some would say you are trying to skirt the law. What is the limit of "disassembled". Is a TV "disassembled" if the power cord is not attached and you call the assembly "a DIY project". Who defines "disassembled".

It is best to just ask if clear rules are not published.

 

[MacIntoshCZ]

It is best to just ask if clear rules are not published.

I asked on the czech forum. They responsed that even with DIY kit i need certification. I read law and found something. So to byppas the law i will sell it for offgrid applications. Dont judge player, judge a game =P. No certification needed anymore.

 

[Nigel Goodwin]

Is a TV "disassembled" if the power cord is not attached and you call the assembly "a DIY project".

No

I can offer 'some' explanations for TV's.

Even sending disassembled TV's here (from China or where ever) is subject to import duties - so to avoid it the PCB's can't be complete. A common solution is to not fit one of the large components - the tuner is a classic case (on the board that contains the tuner), or for a CRT set the LOPTX wouldn't be fitted, for a PSU presumably the transformer wouldn't be fitted.

By leaving off the larger components it not only avoids the import duty, but makes the PCB physically smaller, allowing more boards to be packed in a smaller space.

 

[MacIntoshCZ]

Does anyone has schematic of passive emi filter for 100khz smps or better method of calculatioj? I cant find any procedure to calculate that.
Thanks
***
i used this values

 

[MacIntoshCZ]

Actually working on current feedback loop. If current sense resistor is 10mOhms then with 0,1A there is 1mV across it.
QUESTION is:
What should be lowest output current? after 0A... =)
0,1A is easy to do.
I think 10mA is lowest i am capable design.
At lower currents im aware of noise.
Max is 5A.

 

[gophert]

You should be able to easily add an op amp with a capacitor across the feedback resistor to amplify the voltage drop across the sense resistor and filter the noise. If your sense resistor is connected to ground (not supply) you should be able to get a pretty stable voltage without much noise across a low value resistor.

 

[MacIntoshCZ]

You should be able to easily.

Do you think that sensing 10uV for achieving 1mA resolution is easy? I should use two 10x gain amplifier than i guess. low drift / low offset
My idea for current adjust is to sense current, when is above some value then stop pwm. But with big "B" if cap is charged and load is connected it will lead to current peak and then its should stabilise current. This will also lead to automatic voltage regulation for load.
To prevent current peaks i will need transistor probably mosfet in series with mosfet (with BJT it will work as emitor follower and i have only 12V supply for logic).

 

[MacIntoshCZ]

I wonder where to measure output voltage and how much i should care about precision.
If i measure it across output pins of pwoer supply its still wont be correct voltage on load due contact resistance and resistance of wires.
If i measure it at PW_vout to GND i will get like 0,1V drop on 5A due mosfet RDS and shunt resistance.
PW_VOUT to PW_GNDout requires diferential amplifier etc but provides better voltage reading.
PW_Vout to PW_GND is easier and cheaper.
BTW display resolution will be 50.00V and 5.000A so 10mV step and 1mA step.
Also i sets voltage on load with multimeter. Its the best way
measured voltage is for feedback loop + readings on psu
Is this usable https://www.ti.com/lit/ds/symlink/ths4561.pdf for feedback loop