Hello all,
I've been reading what all of you wrote about smps and their problems but I am facing one that has not been touched.
First of all, my project started with Elliott's project 89.
I used IRFZ44N as mosfets, BYV32 as rectify diodes, a toroidal 40mm diam (300kHz max freq).
Without load I have a beautiful square wave in the primary.
Using 2 PC PSU as source for my smps and a load of 5 12V (21W) bulbs in series, the wave on the primary is still square with no overshoots. When I increase the load the overshoots appear but the psu's can't take the load and shutdown.
I then changed the power source with a car battery and overshoots appear on the raising edge and and short overshoots on the falling edge.
The overshoots on the falling edge are gone because I put some inductors after the rectifiers (before the caps).
I don't think that overshoots are normal but I can't get rid of them.
Tomorrow I will pot some photos from scope.
I am not shore that they are called overshoots or rings or something. I hope I made myself understood despite my bad english.
I've been reading what all of you wrote about smps and their problems but I am facing one that has not been touched.
First of all, my project started with Elliott's project 89.
I used IRFZ44N as mosfets, BYV32 as rectify diodes, a toroidal 40mm diam (300kHz max freq).
Without load I have a beautiful square wave in the primary.
Using 2 PC PSU as source for my smps and a load of 5 12V (21W) bulbs in series, the wave on the primary is still square with no overshoots. When I increase the load the overshoots appear but the psu's can't take the load and shutdown.
I then changed the power source with a car battery and overshoots appear on the raising edge and and short overshoots on the falling edge.
The overshoots on the falling edge are gone because I put some inductors after the rectifiers (before the caps).
I don't think that overshoots are normal but I can't get rid of them.
Tomorrow I will pot some photos from scope.
I am not shore that they are called overshoots or rings or something. I hope I made myself understood despite my bad english.
I have just posted some pictures that I promised.
Here they are:
http://ploter.home.ro/smpsprob/index.html
hope you can help me now.
Here they are:
http://ploter.home.ro/smpsprob/index.html
hope you can help me now.
Hi.
Are You using 56k+100nF on primary and on secondary ??
Also I can't see how big are overshots(spikes) on primary and on secondary ?
I have talked to some peoples and they say that spikes are normaly but they have to be reduced to lowest possible size using inductors, capacitors and good wounded transformer ...
byw: what shematics from ESP you are using first one with pwm regulation or second without pwm ?
Im for testing with loads using this amp http://valveaudio.tripod.com/images/schematics/mosfet50.JPG , preaty simple for testing and very cheap mosfets on output .
bye
Are You using 56k+100nF on primary and on secondary ??
Also I can't see how big are overshots(spikes) on primary and on secondary ?
I have talked to some peoples and they say that spikes are normaly but they have to be reduced to lowest possible size using inductors, capacitors and good wounded transformer ...
byw: what shematics from ESP you are using first one with pwm regulation or second without pwm ?
Im for testing with loads using this amp http://valveaudio.tripod.com/images/schematics/mosfet50.JPG , preaty simple for testing and very cheap mosfets on output .
bye
no, I am using 56ohms on primary and secondary.
on primary the spike is 20-30V with 100W-200W and almost no spike with no load.
the schematic si the second without feedback. i also tried optoisolated feedback and the spike increases.
I changed the 100ohm resistor for deadtime with 500ohm and I saw that the spike happens when the mosfet opens, because with a big dead time the spike can be seen distinct before the next mosfet close.
Is it normal???
Thank you very much for your help.
on primary the spike is 20-30V with 100W-200W and almost no spike with no load.
the schematic si the second without feedback. i also tried optoisolated feedback and the spike increases.
I changed the 100ohm resistor for deadtime with 500ohm and I saw that the spike happens when the mosfet opens, because with a big dead time the spike can be seen distinct before the next mosfet close.
Is it normal???
Thank you very much for your help.
Hi .
Well I'm not that experienced in smps so I could tell you how big spikes are normal, but I know this :
1) I saw on primary that some amps have instead 56+100n -> 2.2+100n or 4.7+100 . You can experement with that !
2) I checked my smps with friends scope with 2 different transformers and one haved greater spikes than other. Both winded with same number of turns only one was winded nicely .
In my smps (not finished fully) I tried with reducing 56 to less but resistors started to heat a lot.
And some peoples told me that there is no neeed for output inductors (one between fast-diodes and capacitors) when you are not using pwm .
Is yout 56ohm producing heat ??
How much current is your smps pulling without load ?
bye
Well I'm not that experienced in smps so I could tell you how big spikes are normal, but I know this :
1) I saw on primary that some amps have instead 56+100n -> 2.2+100n or 4.7+100 . You can experement with that !
2) I checked my smps with friends scope with 2 different transformers and one haved greater spikes than other. Both winded with same number of turns only one was winded nicely .
In my smps (not finished fully) I tried with reducing 56 to less but resistors started to heat a lot.
And some peoples told me that there is no neeed for output inductors (one between fast-diodes and capacitors) when you are not using pwm .
Is yout 56ohm producing heat ??
How much current is your smps pulling without load ?
bye
Hi.
I wounded 3 transformers .
With first wich have too little turns, radiators were very hot when smps loaded(90watt), but were warm with light load .
And I haved voltage drops at secondary side after capacitors ( at about 80watt).
With second and third transformer radiators are completly cold, and slightly warm when loaded about 150watts . My amplifier produce more heat than smps .
I haved no voltage drops at 150watt .
1. transformer -> 4x4/11x11 -> target voltage was 28x28 (bad)
2. & 3. transformer -> 4x4/19x19 -> target voltage was 35x35
Like I said I'm using pwm with optoisolator and I define voltage but zener diode connected to opto.
But i'm using PWM .
Does your 56ohm resistor heats up or is cold ?
Also what kind electrolitic capacitors you use ? Special or normal ?
bye
I wounded 3 transformers .
With first wich have too little turns, radiators were very hot when smps loaded(90watt), but were warm with light load .
And I haved voltage drops at secondary side after capacitors ( at about 80watt).
With second and third transformer radiators are completly cold, and slightly warm when loaded about 150watts . My amplifier produce more heat than smps .
I haved no voltage drops at 150watt .
1. transformer -> 4x4/11x11 -> target voltage was 28x28 (bad)
2. & 3. transformer -> 4x4/19x19 -> target voltage was 35x35
Like I said I'm using pwm with optoisolator and I define voltage but zener diode connected to opto.
But i'm using PWM .
Does your 56ohm resistor heats up or is cold ?
Also what kind electrolitic capacitors you use ? Special or normal ?
bye
they heat up a leattle but i changed them with 2,2 ohm and the spikes ar smaller.
with 56 ohm - 20V spikes
with 2,2 ohm - 7-8V spikes
same load.
I intend to do the same, to wind the transformer for a higher voltage and lower it with pwm. also with optoisolator.
but first I want to see how to solve this problem.
with 56 ohm - 20V spikes
with 2,2 ohm - 7-8V spikes
same load.
I intend to do the same, to wind the transformer for a higher voltage and lower it with pwm. also with optoisolator.
but first I want to see how to solve this problem.
I am using this core:
http://www.kaschke.de/pdf/ferrite/ringe/R40.pdf
Al=3400nH
Material: K2006
http://www.kaschke.de/pdf/material/K2006.PDF
On the net I have found diferent equations.I tried some but did not solve a thing. How should I use this data that the producer of the core is offering?????
http://www.kaschke.de/pdf/ferrite/ringe/R40.pdf
Al=3400nH
Material: K2006
http://www.kaschke.de/pdf/material/K2006.PDF
On the net I have found diferent equations.I tried some but did not solve a thing. How should I use this data that the producer of the core is offering?????
Please can somebody post some photos with their primary wave, so I can see the diference. Please.
I tried everything I could think of. Rewind the transformer, I don't know how many times. No change. I don't kow if it is working well or not.
Should I try to change the topology with full brigde pushpull??? I don't need more than 300W or so but If I can't make this one work I try again.
I'm in the dark here.
I tried everything I could think of. Rewind the transformer, I don't know how many times. No change. I don't kow if it is working well or not.
Should I try to change the topology with full brigde pushpull??? I don't need more than 300W or so but If I can't make this one work I try again.
I'm in the dark here.
SMPS problem
Hi alexv,
Greetings from the USA. As far as the power supply goes, switch-mode power supplies (SMPS) are tricky and take a great amount of time to really master. I've been designing such as part of my job for many years, and I still have to really be careful. I've designed six to seven dozen, and I'm still learning new stuff with each new design. I would recommend that you visit web sites such as Texas Instruments, National Semiconductor, Linear Technology, etc. and download some application notes re SMPS. For your first project, start with a simple non-isolated (transformerless), "buck" dc-dc converter. Then, I would suggest studying the boost, and the buck-boost converters. Once familiar with these circuits, tackle the transformer-based networks, starting with flyback, and simple forward converters. Then proceed to the symmetrical converters (push-pull center-tap, half-bridge, full-bridge). That wil take more time than most are willing to spend, but troubleshooting an SMPS over the web takes very long and produces much frustration. Also, I would use a newer control IC. The 3524/3525 have served us well for more than two decades, but many newer controllers have come out since then, and offer much better performance. Also, I would steer clear of any SMPS that doesn't have an inductor (except flybacks, since the transformer stores energy the way an inductor would). Inductors reduce noise and stress on the parts, and increase efficiency. I'm not trying to discourage you from experimenting, just trying to spare you some frustration. As far as the current SMPS you're working on goes, read the SG3524 or SG3525 data sheet and observe the example circuits shown. These use inductors. They work well. Please be careful. Best regards.
Hi alexv,
Greetings from the USA. As far as the power supply goes, switch-mode power supplies (SMPS) are tricky and take a great amount of time to really master. I've been designing such as part of my job for many years, and I still have to really be careful. I've designed six to seven dozen, and I'm still learning new stuff with each new design. I would recommend that you visit web sites such as Texas Instruments, National Semiconductor, Linear Technology, etc. and download some application notes re SMPS. For your first project, start with a simple non-isolated (transformerless), "buck" dc-dc converter. Then, I would suggest studying the boost, and the buck-boost converters. Once familiar with these circuits, tackle the transformer-based networks, starting with flyback, and simple forward converters. Then proceed to the symmetrical converters (push-pull center-tap, half-bridge, full-bridge). That wil take more time than most are willing to spend, but troubleshooting an SMPS over the web takes very long and produces much frustration. Also, I would use a newer control IC. The 3524/3525 have served us well for more than two decades, but many newer controllers have come out since then, and offer much better performance. Also, I would steer clear of any SMPS that doesn't have an inductor (except flybacks, since the transformer stores energy the way an inductor would). Inductors reduce noise and stress on the parts, and increase efficiency. I'm not trying to discourage you from experimenting, just trying to spare you some frustration. As far as the current SMPS you're working on goes, read the SG3524 or SG3525 data sheet and observe the example circuits shown. These use inductors. They work well. Please be careful. Best regards.
Hello Claude,
In the past few weeks I've read hundrets of application notes form all the websites you mentioned and more irf, on maginc .. and so on.
I have a dial-up connection (if you remember what that is, stone age for a normal country, but Romania is not )and I've dounloaded almost 1Gb of documents, pictures and schematics.
I want to tell you a leattle story to show you how hard it is to find component in R.
I tried to buy a toroidal form the only store that seals them in the capital city of R. after they order it form Germany they call me in a week and tell me to come and pick up a diferent core made of difrent, not sutable material. Today I descovered that kaschke produces the ferrites in Romania, and from 50 products they manufacture I can only order 3-4 and form thouse 4 I get the wrong one.
hope I did not bore you, not my intention, but I am angry. sorry.
About that inductor. I put inductors on the secondary after the rectifiers and before the caps, but I saw in many circuits another inductor placed on the center tap of primary. Not the filter on power supply but an inductor. Should I try to place one or were you refering to other inductors.
thank you very much for your answer.
In the past few weeks I've read hundrets of application notes form all the websites you mentioned and more irf, on maginc .. and so on.
I have a dial-up connection (if you remember what that is, stone age for a normal country, but Romania is not
)and I've dounloaded almost 1Gb of documents, pictures and schematics.I want to tell you a leattle story to show you how hard it is to find component in R.
I tried to buy a toroidal form the only store that seals them in the capital city of R. after they order it form Germany they call me in a week and tell me to come and pick up a diferent core made of difrent, not sutable material. Today I descovered that kaschke produces the ferrites in Romania, and from 50 products they manufacture I can only order 3-4 and form thouse 4 I get the wrong one.
hope I did not bore you, not my intention, but I am angry. sorry.
About that inductor. I put inductors on the secondary after the rectifiers and before the caps, but I saw in many circuits another inductor placed on the center tap of primary. Not the filter on power supply but an inductor. Should I try to place one or were you refering to other inductors.
thank you very much for your answer.
the spikes are when the mos turns off. so if the turnoff time si very short the spikes are lower.
I tryed to rewind the transformer again and i get the best resaults when the primary wires are layed side by side. for example 6 strands that form one primary are layed one after the other not crossing and then the other primary the same. I will try the same with the secondary because nou the secondarys have 4 strands twisted together.
Is iti normal??? What is improved in this way?
I also experimented with difrent numbers of turns for the primary. There is no change. also I changed the freq and there was still no change. One formula stated that the number of turn are tightly related to freq and B (magnetix flux). So changing the variables in the equation, something must change, at least for the worst if not for better. The core can go up to 300khz but with the same number of turn and very diferent f the flux shuld change and afect the wave form, doesn't it.
I tryed to rewind the transformer again and i get the best resaults when the primary wires are layed side by side. for example 6 strands that form one primary are layed one after the other not crossing and then the other primary the same. I will try the same with the secondary because nou the secondarys have 4 strands twisted together.
Is iti normal??? What is improved in this way?
I also experimented with difrent numbers of turns for the primary. There is no change. also I changed the freq and there was still no change. One formula stated that the number of turn are tightly related to freq and B (magnetix flux). So changing the variables in the equation, something must change, at least for the worst if not for better. The core can go up to 300khz but with the same number of turn and very diferent f the flux shuld change and afect the wave form, doesn't it.
Hi.
Well about type of winding Im also confused.
In Elektor magazine (electronics magazine), they published car smps, but they used copper foil for primary, and couple of normal wires for secondary (secondary was wounded first - EE core).
On Internet you will find primary first then secondary and revers.
Whats best I dont know.
All of them use winding side by side.
One guy toled me that its best to wind primary parallel (that bot primaries at one time), and secondary serial (that x turns - center tap - x turns).
I havent tried that, I dont have nervs to wind anymore.
Good luck, and report your progess
Well about type of winding Im also confused.
In Elektor magazine (electronics magazine), they published car smps, but they used copper foil for primary, and couple of normal wires for secondary (secondary was wounded first - EE core).
On Internet you will find primary first then secondary and revers.
Whats best I dont know.
All of them use winding side by side.
One guy toled me that its best to wind primary parallel (that bot primaries at one time), and secondary serial (that x turns - center tap - x turns).
I havent tried that, I dont have nervs to wind anymore.
Good luck, and report your progess
To reduce the spike i've found this: TOSHIBA
What do you think about this???
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
What do you think about this???
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.