if you have to isolate the primary and secondary ground systems (which I don't really think is necessary in this case), consider an auxiliary secondary winding on the transformer, rectify, filter and use this "sense winding" to detect the error in the secondary output.MaXiZ said:
But explain to me why isolation is necessary. I can see this in the case of ground leakage in an A.C. system, for an off-line switcher, but not for a car audio system.
The problem
Boys and girls,
I hope this finds you all well, I was playing with these 12V smps for a while, I think that the feedback that your trying to use with this chip isnt quite correct. Your problem is that the SMPS is at maximum duty cycle because the feedback network isnt triggering the chip properly to cut down the duty cycle. Thus why you are seeing maximum voltage at all times, this is "wide open throttle" as I like to call it
Now, try perhaps using something like a 10K resistor in your inverting input to ground instead of a 1K(The one that you input your optocoupled feedback into).
Now guys, play with this and you shall fix it. I have also found that maybe using the VREF might overload it, try perhaps using the VIN voltage to trigger it. This might work a little better.
If your problems persist please email me and ill attempt to help you out..
Many thanks
Matt Donohoe
Boys and girls,
I hope this finds you all well, I was playing with these 12V smps for a while, I think that the feedback that your trying to use with this chip isnt quite correct. Your problem is that the SMPS is at maximum duty cycle because the feedback network isnt triggering the chip properly to cut down the duty cycle. Thus why you are seeing maximum voltage at all times, this is "wide open throttle" as I like to call it
Now, try perhaps using something like a 10K resistor in your inverting input to ground instead of a 1K(The one that you input your optocoupled feedback into).
Now guys, play with this and you shall fix it. I have also found that maybe using the VREF might overload it, try perhaps using the VIN voltage to trigger it. This might work a little better.
If your problems persist please email me and ill attempt to help you out..
Many thanks
Matt Donohoe
Re: The problem
I tried making that point at the beginning of the thread, and similar threads -- seems that folks are just fixated on using the web-circuit without reading the supporting materials from teh chip manufacturers.
mdonohoe said:Boys and girls,
I hope this finds you all well, I was playing with these 12V smps for a while, I think that the feedback that your trying to use with this chip isnt quite correct. Your problem is that the SMPS is at maximum duty cycle because the feedback network isnt triggering the chip properly to cut down the duty cycle. Thus why you are seeing maximum voltage at all times, this is "wide open throttle" as I like to call it
Now, try perhaps using something like a 10K resistor in your inverting input to ground instead of a 1K(The one that you input your optocoupled feedback into).
Now guys, play with this and you shall fix it. I have also found that maybe using the VREF might overload it, try perhaps using the VIN voltage to trigger it. This might work a little better.
If your problems persist please email me and ill attempt to help you out..
Many thanks
Matt Donohoe
I tried making that point at the beginning of the thread, and similar threads -- seems that folks are just fixated on using the web-circuit without reading the supporting materials from teh chip manufacturers.
SMPS
Well,
What can I say, somebody has used the wrong chip in that circuit, I believe that the SG3524 and SG3525's are compatable but that circuit needs SLIGHT changes before it can work in the other circuit. This is one such change. I believe the problem is that by putting a 1K resistor to ground you load up the VREF so badly that it is unable to reach the required potential to trigger the inverting amplifier to cut the duty cycle. Thus why I used the VIN. end of story.
Get an Oscilloscope!!!!!!!!!!!!!!!!!!!!! I bought mine for $150 used and about 30 years old, 100mhz bandwidth, heavy as a brick, but invaluable in circuits like this!!
Any further queries please do not hesitate to PM me or reply here I am quite willing to help out anybody venturing into the world of SMPS'.
Matt Donohoe
Well,
What can I say, somebody has used the wrong chip in that circuit, I believe that the SG3524 and SG3525's are compatable but that circuit needs SLIGHT changes before it can work in the other circuit. This is one such change. I believe the problem is that by putting a 1K resistor to ground you load up the VREF so badly that it is unable to reach the required potential to trigger the inverting amplifier to cut the duty cycle. Thus why I used the VIN. end of story.
Get an Oscilloscope!!!!!!!!!!!!!!!!!!!!! I bought mine for $150 used and about 30 years old, 100mhz bandwidth, heavy as a brick, but invaluable in circuits like this!!
Any further queries please do not hesitate to PM me or reply here I am quite willing to help out anybody venturing into the world of SMPS'.
Matt Donohoe
At no time did I say they were pin for pin compatable, I said they were compatable. Which means that essentially they are the same, except for a few things like a few different connections need to be made.
At any rate, any person that looked at the datasheets(easily available, try a search on excite.com) would figure this out in only a short time.
Matt Donohoe
At any rate, any person that looked at the datasheets(easily available, try a search on excite.com) would figure this out in only a short time.
Matt Donohoe
mdonohoe said:
about 0.188 amps, integrated over the full duty cycle, so the instantaneous current pulse is many, many times higher -- as donohoe points out the problem is the imprecision of the transformer -- could be a bit "scramble wound", the windings might not be balanced, a lot of things -- you might want to consider adding a bit of inductance on the secondary side -- 100uH
i hate to dampen enthusiasm for SMPS, but you really need a good scope to track down these problems.
oh, one more thing, are you running the supply loaded or unloaded ? the amount of inductance you need to store energy is inversely related to the load.
hello once again,
by the way, anybody whos wondering why am i fit to give advice please ask i'm quite happy to answer and share my experiences with SMPS supplies. They are extremely challenging, but at the same time fun!!!
Ok... here goes...
your snubber components shouldnt be that warm at idle or no load, I say this because that optoisolated loop you installed(you did fix it didnt you!??!?!?!) should be running the duty cycle somewhere near 1% once the capacitors are fully charges(ie. you get extremely small pulses every second or so to top them up!!)
Thus, you are generating no real spikes or the need to "snub" said spikes..
I suggest if you have not fixed your optoisolator problem you perhaps contact me or display your current circuit as well as the chip that your using and I may be able to lend some assistance.
SMPS' are wonderful, however, the hardest part can be getting a smooth and ripple free output.. ferrite beads are a godsent here! as well as a reasonable amount of inductance.. ferrite beads oppose quick changes in current i think and thats why they work so well, much like a inductor.. or is it that they block noise? i cant remember... somebody help me here.. I use them and they do a good job at any rate..
A well built SMPS even from something like a 50V-0-50V 400VA transformer that you plug into the wall can be very good.. Because you can get virtually any voltage you like by replacing your fixed zener diode with a potentiometer setup(I've also done this, it works a treat) and you get extremely high efficiency when stepping down 50 to 100V down to 1VDC+ as well as lots of current should you require it.. Also easy to implement a OPAMP based current limiter... which shuts down either the inverting input or toggles the shutdown line..
anyways, let me know what you think
Matt Donohoe
donohoem@iprimus.com.au
by the way, anybody whos wondering why am i fit to give advice please ask i'm quite happy to answer and share my experiences with SMPS supplies. They are extremely challenging, but at the same time fun!!!
Ok... here goes...
your snubber components shouldnt be that warm at idle or no load, I say this because that optoisolated loop you installed(you did fix it didnt you!??!?!?!) should be running the duty cycle somewhere near 1% once the capacitors are fully charges(ie. you get extremely small pulses every second or so to top them up!!)
Thus, you are generating no real spikes or the need to "snub" said spikes..
I suggest if you have not fixed your optoisolator problem you perhaps contact me or display your current circuit as well as the chip that your using and I may be able to lend some assistance.
SMPS' are wonderful, however, the hardest part can be getting a smooth and ripple free output.. ferrite beads are a godsent here! as well as a reasonable amount of inductance.. ferrite beads oppose quick changes in current i think and thats why they work so well, much like a inductor.. or is it that they block noise? i cant remember... somebody help me here.. I use them and they do a good job at any rate..
A well built SMPS even from something like a 50V-0-50V 400VA transformer that you plug into the wall can be very good.. Because you can get virtually any voltage you like by replacing your fixed zener diode with a potentiometer setup(I've also done this, it works a treat) and you get extremely high efficiency when stepping down 50 to 100V down to 1VDC+ as well as lots of current should you require it.. Also easy to implement a OPAMP based current limiter... which shuts down either the inverting input or toggles the shutdown line..
anyways, let me know what you think
Matt Donohoe
donohoem@iprimus.com.au
MaXiZ,
why do you use a fullbridge rectifier at the transformers secondary and what is the duty cycle?
The fullbridge rectifier both the "forward" and "flyback" energy comming from the transformer.
At small/no load, if the circuit cant regulate over full 100% duty cycle range, the voltage can actually rise verry much as you discovered with your first schematic presented in this thread(I guess you did connect in your feedback then...?).
why do you use a fullbridge rectifier at the transformers secondary and what is the duty cycle?
The fullbridge rectifier both the "forward" and "flyback" energy comming from the transformer.
At small/no load, if the circuit cant regulate over full 100% duty cycle range, the voltage can actually rise verry much as you discovered with your first schematic presented in this thread(I guess you did connect in your feedback then...?).
Hey MaXiZ, just curious if you ever solved your prob.
I did notice in your schem you didn't have any drains on your outputs. A resistance for unloaded supply situations on your filtered sec is required to keep it from walking up or down the zero point when running isolated. So instead of having 50-0-50 you may get 35-0-65. I have seen this in my own supplies.
I did notice in your schem you didn't have any drains on your outputs. A resistance for unloaded supply situations on your filtered sec is required to keep it from walking up or down the zero point when running isolated. So instead of having 50-0-50 you may get 35-0-65. I have seen this in my own supplies.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.