Oh yeah, almost forgot- I like the thick aluminum stock, too. I understand about not wanting to re-re-re-rewind the xfmr yet another time. Perhaps of you do re-do the board, if you get superior performance from the toroid, dimension the baord to accommodate the toroid. There's just something visually stunning about lots of toroids........
Hi
You could ask Eva and she would said that toroid won't have superior performance, if you look at everything. But I like them, because all windings are in contact with air so better cooling.
I find lots of toroids stunning too, but it is hard to use 2 or more for mains powered applications(not needed, one is enough).That could just happen in case that there is not enough space for all windings.
Aluminum stock is cool for this application, since there is so high efficiency + they are air cooled(with fan that is)
You could ask Eva and she would said that toroid won't have superior performance, if you look at everything. But I like them, because all windings are in contact with air so better cooling.
I find lots of toroids stunning too, but it is hard to use 2 or more for mains powered applications(not needed, one is enough).That could just happen in case that there is not enough space for all windings.
Aluminum stock is cool for this application, since there is so high efficiency + they are air cooled(with fan that is)
Luka,
If you're refering to her post regarding toroids (I think in another thread), I think what she said is that they're not well-suited for applications where the pri-sec ratio is larger than, say, 7:1 (poor pri-to-sec coupling), or where the number of turns, like on a 5V secondary, would have so few turns that it would have a hard time covering the entire core.
But for say, a full-, or half-bridge, putting out , say, +/-40V, or +/-50-60V, with a ratio of 2:1, or even 3:1, where the secondaries are 8T + 8T or greater, a toroid would fit the bill nicely.
If you're refering to her post regarding toroids (I think in another thread), I think what she said is that they're not well-suited for applications where the pri-sec ratio is larger than, say, 7:1 (poor pri-to-sec coupling), or where the number of turns, like on a 5V secondary, would have so few turns that it would have a hard time covering the entire core.
But for say, a full-, or half-bridge, putting out , say, +/-40V, or +/-50-60V, with a ratio of 2:1, or even 3:1, where the secondaries are 8T + 8T or greater, a toroid would fit the bill nicely.
Hi
That is great news, I will keep that in ming for nex time, but Pri must be on whole toroid, what about Sec?
That is great news, I will keep that in ming for nex time, but Pri must be on whole toroid, what about Sec?
I have told him several times to spread all windings around the toroid for minimum leakage inductance and stray fields, but to no avail 😉 I give up.
Hi
Now I will remember it, it is the first thing that will pop to my mind about toroids. I must be getting old, I don't even remember that that was said, but I'm sure it was. But thanks for saying it again
Now I will remember it, it is the first thing that will pop to my mind about toroids. I must be getting old, I don't even remember that that was said, but I'm sure it was. But thanks for saying it again
Hi
That shouldn't be happening to me, well not yet anyway 😀, First you have to learn all in school, then comes time to forget all. It seems that that time has come for me too early 😀😀 or more like
That shouldn't be happening to me, well not yet anyway 😀, First you have to learn all in school, then comes time to forget all. It seems that that time has come for me too early 😀😀 or more like
I learned nothing about SMPSs in school (either high school or college). But then, SMPSs weren't common when I was in HS! 😀
I learned almost everything I know from Chryssis', Pressman's, Hnatek's, Brown's, and one other's book (Sorry, can't remember his name). The rest, from datasheets, app notes, and, of course, here!
I learned almost everything I know from Chryssis', Pressman's, Hnatek's, Brown's, and one other's book (Sorry, can't remember his name). The rest, from datasheets, app notes, and, of course, here!
You know that, same here, smps's guess still aren't important 😀, so I learned all from diyA, you guys and girls, and those books...
Wouldn't be where I am now if there wouldn't be for diyA and its great users... may thanks yet again
Wouldn't be where I am now if there wouldn't be for diyA and its great users... may thanks yet again
N-Channel said:
Me too,
Pressman´s and Chryssys´ books and a lot of exploded boards were my smps teachers.
Simulation of a halfbridge converter with opto feedback
Luka
How are you its been along time and I see you have completed your project and there seems to be some loose ends along with some small problems. I have uploaded some files that might interest you to the below link. If you like you can download the software and the files and used them as you wish. I think you will see that if you only monitor one rail that the other will track closely also the 50% load variation on either rail has little overall effect. While this is not professional it does allow you to get a snapshot of your supply. If you have any questions let me know.The supply has a 100kHz FS and dual outputs of 35 volts @10amps.
chas1
http://groups.yahoo.com/group/LTspice/files/ Temp/HalfBridge_Converter_files.zip
Luka
How are you its been along time and I see you have completed your project and there seems to be some loose ends along with some small problems. I have uploaded some files that might interest you to the below link. If you like you can download the software and the files and used them as you wish. I think you will see that if you only monitor one rail that the other will track closely also the 50% load variation on either rail has little overall effect. While this is not professional it does allow you to get a snapshot of your supply. If you have any questions let me know.The supply has a 100kHz FS and dual outputs of 35 volts @10amps.
chas1
http://groups.yahoo.com/group/LTspice/files/ Temp/HalfBridge_Converter_files.zip
Hi
Do you have any other place to upload files, or send me directly so that I don't have to register there... would be much appreciated
Do you have any other place to upload files, or send me directly so that I don't have to register there... would be much appreciated
PM
Luka
I don't mind, send me a PM with your email address and I will send the zip file as an attachment but you still need LTspice software.
chas1
Luka
I don't mind, send me a PM with your email address and I will send the zip file as an attachment but you still need LTspice software.
chas1
zip files and simulation
Luka
Did the corrections work and did you receive the .wmf schematic file?
chas1
Luka
Did the corrections work and did you receive the .wmf schematic file?
chas1
HI
Jup sure did. There seems that everything is ok, sim is running and current limit works as well. I just don't know why you have 1.2mH for each sec, as this generates 165Vp in each sec. Isn't this too much? I have lower it to 0.529mH or 14 turns.
Jup sure did. There seems that everything is ok, sim is running and current limit works as well. I just don't know why you have 1.2mH for each sec, as this generates 165Vp in each sec. Isn't this too much? I have lower it to 0.529mH or 14 turns.
transformer turns
Luka
Good observation, You can now simulate what if's , Like what happens when the input power dips to 200 volts or below and /or transients on the line, remember in front of the rectifiers you always need twice the output voltage the load requires. Try pulsing the input voltage (320-180) monitor the voltage in front of the diode and the inductor current making sure you don't go into DCM. a slight overwind of a transformer protects against this. If you were designing for low voltage like 3.3 VDC ,5 volts and where you might have a 12 volt winding and you need to use schottky diodes then the diodes would dictate the winding of the transformer in our case we are dealing with diodes that can handle the voltages and the currents so the biggest factor is coreset size , a couple of extra turns will not make a big difference in that regard and it insures that you have enough voltage to keep the supply in CCM even with light loads ; have you bothered to simulate the supply with a light load to see where in transitions into DCM and then you will see the feedback compensation is on the hairy edge because of the change in the output frequency poles and zero locations. If I remeber correctly this will occur about 2 amps and also note the supply was designed for .05v ripple measure it you will see it's very close. However I changed the ESR of the filter caps (raised it) so the ripple in my supply is about .7 volts but compensation is easier.If you like I will send a file that will help with compensaton calculation and plots the results thru the power section.
chas1
Luka
Good observation, You can now simulate what if's , Like what happens when the input power dips to 200 volts or below and /or transients on the line, remember in front of the rectifiers you always need twice the output voltage the load requires. Try pulsing the input voltage (320-180) monitor the voltage in front of the diode and the inductor current making sure you don't go into DCM. a slight overwind of a transformer protects against this. If you were designing for low voltage like 3.3 VDC ,5 volts and where you might have a 12 volt winding and you need to use schottky diodes then the diodes would dictate the winding of the transformer in our case we are dealing with diodes that can handle the voltages and the currents so the biggest factor is coreset size , a couple of extra turns will not make a big difference in that regard and it insures that you have enough voltage to keep the supply in CCM even with light loads ; have you bothered to simulate the supply with a light load to see where in transitions into DCM and then you will see the feedback compensation is on the hairy edge because of the change in the output frequency poles and zero locations. If I remeber correctly this will occur about 2 amps and also note the supply was designed for .05v ripple measure it you will see it's very close. However I changed the ESR of the filter caps (raised it) so the ripple in my supply is about .7 volts but compensation is easier.If you like I will send a file that will help with compensaton calculation and plots the results thru the power section.
chas1
Hi
In forward C I have tryed output current changing from 0.1A to 9.9A in one rail, other the same but is delayed so that it has peak when other is lowest. Output voltage changes only 0.6v with 2200uF
And yoe please send me that file too.
In forward C I have tryed output current changing from 0.1A to 9.9A in one rail, other the same but is delayed so that it has peak when other is lowest. Output voltage changes only 0.6v with 2200uF
And yoe please send me that file too.
output circuit
Luka
Checkout Pressman pg 448-460 , it's very good for output filter calcuations which is the first thing that should be done before supply is designed IMHO and yes I will send you the files, they might take some explanation but they are straight forward concerning the results. I think if the schematic of your supply is the one you have posted there is a lot of capacitance and all in parallel which will decrease the ESR (figure close to 0) and the phase will drop to almost 180 at flc. I think if you use two caps of about 500u in parallel on each rail which will give you about 1000u combined with an inductor of about 40-50uH which in effect is about 100u if wound on common core, if not then you should increase them to about 100uH in each rail will give you a good ripple component and be easier to compensate (don't be fooled use typeIII compensaton) a few extra components but for VCM is highly recommended.
chas1
Luka
Checkout Pressman pg 448-460 , it's very good for output filter calcuations which is the first thing that should be done before supply is designed IMHO and yes I will send you the files, they might take some explanation but they are straight forward concerning the results. I think if the schematic of your supply is the one you have posted there is a lot of capacitance and all in parallel which will decrease the ESR (figure close to 0) and the phase will drop to almost 180 at flc. I think if you use two caps of about 500u in parallel on each rail which will give you about 1000u combined with an inductor of about 40-50uH which in effect is about 100u if wound on common core, if not then you should increase them to about 100uH in each rail will give you a good ripple component and be easier to compensate (don't be fooled use typeIII compensaton) a few extra components but for VCM is highly recommended.
chas1