0-50V 0-10A Variable SMPS

Status
Not open for further replies.

If you look at PFC app notes of ON Semiconductor, they always earthed the negative rail of PFC output. Yea i just used different symbols. Each one is isolated to each other.

Yea, i realized that later but it should work anyway. This way must disable that 2.5V reference pin.
Click to expand...

This one was about connecting the VFB and COMP pin of UC3844. Normally as voltage follower, we need to access non-inv pin but here we are not able to access. In my schematic, non-inv pin has 2.5V reference voltage. Thats why i said it should work but i am not sure anway.

My mistake, PFC output should not be earthed. Now i saw in PFC app notes of On Semi and i think only the stage before rectifier must have an earth.
 
Last edited:

Not enough drive current for Q8 to fan. Use a 1V Nch "logic level" Fet <<100mOhm. and consider Thermistor on hot spot to regulate bias a fan speed with injected triangle wave to get PWM.

V CE(sat) Collector-Emitter Saturation Voltage
(I C =500mAdc, I B =50mAdc) Vce(sat) = 0.75Vmax

Note : Ic/Ib=10 is common switch configuration, unless special low VceSat
 
I got the fan control circuit from PFC app notes of ON Semi. I have used here as 24V whereas they used 12V for fan.

- - - Updated - - -

Here are the updated schematics below. Now i seperated all grounds, earths etc.
 

Attachments

  • voltage, current control and fan.jpg
    631.4 KB · Views: 142
  • supply.jpg
    626.3 KB · Views: 146
  • pfc.jpg
    827.3 KB · Views: 130
  • control, power, output.jpg
    729.9 KB · Views: 152

One of my post has been removed i think. Anyway, i updated the schematic. I seperated the grounds and removed unnecessary 24V since now there is a pwm controlled fan. This fan sets its speed according to output current of SMPS. Its speed is proportional to output current.
 

Attachments

  • schematic.jpg
    2.5 MB · Views: 145
Last edited:

fethiyeli said:
One of my post has been removed i think.
Click to expand...

Looking back 48 hours, no post of yours appears to have been deleted. (It would remain visible to moderators if it were.)

Every so often the system pauses longer than normal after you click 'Reply'. This might be an indication of a glitch of some sort. It might be in the system interface, your internet connection, or your device's compatibility with the interface.

If you think your message might be caught in a glitch, then it is a good idea to quickly copy it, in case the system fails to post it.
 

Maybe it was my fault. I thought i had posted it but i am not sure. It is not important.

What do you think about the schematic ? Do you have any suggestion ? opcorn:
 

fethiyeli said:
I dunno what to say. Here is the app note of Fairchild. They used the same ground everywhere. Which one is better ? Seperated or combined ground ?

https://www.fairchildsemi.com/application-notes/AN/AN-4165.pdf
Click to expand...
I believe they are improperly using the earth ground symbol,it's not actually connected to the earth safety ground, otherwise it would not work. Many people, including major IC manufacturers, incorrectly use the dashed earth ground symbol as a general purpose reference ground symbol in their documentation. In that document, the solid triangle ground seems to represent a high power ground, and the dashed symbol represents signal ground for the controller, and they are tied together at the negative output terminal of the PFC. Neither of those symbols are the "true" earth from the mains.

Often documentation will indicate earth ground using the chassis ground symbol (the slanted fork one), which is also tied to earth ground. This is the case with the design in your document. This net is usually not connected to any of the internal circuitry (except for Y caps) for safety and EMC reasons.
 

Exactly. Many IC manufacturers use earth symbol everywhere and thats why i used the earth connected to real earth. It is so confusing. I have also thought that why they had grounded the negative rail of PFC bus. Now i seperated all of them. I forgot to add an operational amplifier to increase the sense voltage of 0.5V to 1.5V for fan control circuit.

Here is the updated one below.
 

Attachments

  • schematic.jpg
    2.6 MB · Views: 133
Last edited:

Here is the updated schematic. Updated one has redesigned fan control circuit.
 

Attachments

  • schematic.jpg
    2.7 MB · Views: 177

VFB and COMP still connected together...remember that opamps try to make their inputs the same......so you have a ref of 2.5v on the noninv pin of the internal opamp, and so it will try and drive the other pin to 2.5v, which wont do a good job of load regulation for you.
 


That is the what we want for regulation in deed. When i change the voltage of the noninv pin, inv pin also needs to be changed. This is how output regulation works. Voltage and current control circuits works good in simulation. When i tie comp and vfb pins together, these pins get disabled. So voltage of the vfb pin goes through error amp output. Do u have any suggestion ?
 

if you are bypassing the internal error amplifier in uc38xx, then you usually use it as an inverting amp stage, or instead, ground the FB pin and have like a common emitter-lke stage with collector to COMP....that may be an op amp acting as external error amplifier with its output to the comp pin.
 
Does anyone have suggestion for supplying ICs and fan ? I always hear that TOP ICs are pretty vulnerable. Experienced guys say that they always had damaged TOP ICs in their projects so they didnt use them anymore.
 

Actually i dont want to use TOP246YN. Instead of using that IC, bias winding from choke coil is suitable to feed IC and opamps but there are some issues about that. As you see in attachment, we need between 12.25V and 14.5V. If we choose 2 turns for bias winding, and 51 turns for primary, we get (2/51)*390V = 15.29V. This voltage may reduce down to 14V due to forward voltages of diodes at output of bias windings. So there is no guarantee to feed the NCP1653 at startup since it needs 14.5V at most. I need to use something more practical solution for supplying both IC and fan. Also i need isolated voltage for other fan opamps. Does anyone have suggestion ?
 

Attachments

  • 1.jpg
    36.9 KB · Views: 115

I would use a micrel integrated chip boost converter (or sepic etc) to provide the startup voltage to your chip.....then it can disable itself, or you just leave it in there. I think its MICXXXX or MIC33XX or something....good and few components, as the fet is in the control chip,
that is your voltage is not quite right, as u described, so use the micrel to get it to what you want.
 

It might be hard to find it here. I also need isolated supplies for UC3844, LM324, LM358 and LM393 apart from NCP1653 supply. Do you have suggestion else ?
 

fethiyeli said:
I got the fan control circuit from PFC app notes of ON Semi. I have used here as 24V whereas they used 12V for fan.
Click to expand...

Why drop 24V to 5V for 5V fan.

  • Select a 24Vdc fan.
  • Delete 5V linear Reg as it will waste almost 4x power as the fan.
  • https://www.bing.com/images/search?q=24VDC+Cooling+Fans&FORM=IRBPRS&=0
  • Use Nch-FET with Thermistor bias on hot spot.set to 50'C over 5'C range. with low side regulation.
  • 2 fans are more reliable than 1.
  • Use PC-film plenum to control air flow over hot spots for low RPM but max linear surface speed over hotspot.
 
Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…