Even if you get the core you have working, you don't know anything about it. From the people I've helped, an unknown core (particularly one that was being used as a filter inductor in a commercially available piece of equipment) is almost always going to cause problems. The properties of cores change as they heat up. At higher temps, the core may cause the supply to fail due to a change in permeability (or other parameters). It's better to use a known core (at least one that was being used as a power transformer in a switching power supply). If you get the supply working, then you can try other cores. You want as few variables as possible.
Hello All,
Buying a toroid from overseas is out of guestion from the ground I stand the best material I can think is EI trafo from computer power supply or TV power supply.
EI from computer supply is to small I think it will difficult to me to put turn that I need to power a car amp ? any one with real experience with PC trafo use to made a car amp PS is welcome
EI from TV is bigger is easier to use. now I found a new problem the TV supply trafo work in 100Khz if I use lower freq let say 50 Khz could it generate a new problem ?
any suggestion how to calculated how many turn I need if I use lower freq for TV trafo ?
Buying a toroid from overseas is out of guestion from the ground I stand the best material I can think is EI trafo from computer power supply or TV power supply.
EI from computer supply is to small I think it will difficult to me to put turn that I need to power a car amp ? any one with real experience with PC trafo use to made a car amp PS is welcome
EI from TV is bigger is easier to use. now I found a new problem the TV supply trafo work in 100Khz if I use lower freq let say 50 Khz could it generate a new problem ?
any suggestion how to calculated how many turn I need if I use lower freq for TV trafo ?
I'm still having trouble getting hold of suitable cores this end. I've been trying to dismantle some quite large E types from printer switchmode supplies. These things are rated at 700VA if the information printed on the tape is to believed. However i haven't managed to get one apart successfully yet. I've tried paint stripper and nitromorse. The latter softened everything up but caused the windings to swell and crack the core. I tried heating a transformer in the oven at 200'c. This resulted in cracking the core also. It seems that these transformers are too well bonded to be able to use. If anyone has any suggestions on how to dismantle these transformers please do share your experiences.
Leigh
Leigh
The first link points to a post with links to 3 USA based online shops. I've placed several successful orders with one of them. I've also got ferrites from Farnell and RS Amidata sometimes.
On the other hand, almost all the ferrites, even the less suitable ones, will work fine in your application if you just operate them under 150mT. This leaves gentle room for temperature changes. There are not that many things that can go wrong with an unknown core.
On the other hand, almost all the ferrites, even the less suitable ones, will work fine in your application if you just operate them under 150mT. This leaves gentle room for temperature changes. There are not that many things that can go wrong with an unknown core.
Yay!
I finally managed to melt an old switchmode transformer into submission using nitromorse in a tin. The thing just fell apart in my hands without breaking anything. I'll try winding it today with 5:25 windings. Hopefully this core should work better than my other's and hold its output voltage. If not i'm gonna have to seek professional help to de-mystify the subject of transformers and their cores.
Leigh
I finally managed to melt an old switchmode transformer into submission using nitromorse in a tin. The thing just fell apart in my hands without breaking anything. I'll try winding it today with 5:25 windings. Hopefully this core should work better than my other's and hold its output voltage. If not i'm gonna have to seek professional help to de-mystify the subject of transformers and their cores.
Leigh
I don't recommend E cores and the like for 12V push-pull applications, winding them optimally is is a true pain and sometimes a problem with no good solutions. Note that the cores that you will find in most power supplies are gapped and not suitable for a 12V push-pull either. It's easier to get a ferrite toroid from a junk car amplifier instead (or to buy it online).
Hi Eva,
May I ask question about switchmode trafo.
1. why in big SMPS I found mostly use EI trafo and I see you dont like it use in car SMPS the topology perhaps........... what about SMPS for PA off course it will use 220V ac supply use are you still suggest to use toroid ?
2. In china made amp I measure mostly use 33Khz swithing freq I try to bring it up to 47Khz but its seem nothing changed ( heat, voltage output ) what is optimal freq use in car amp in your experience.
off the topic. here in indonesia with 30 to 60 USD you can buy an china made amp with attractive looking and bad sound my goal is made this thing sound better any sugestion especially in PS term.
May I ask question about switchmode trafo.
1. why in big SMPS I found mostly use EI trafo and I see you dont like it use in car SMPS the topology perhaps........... what about SMPS for PA off course it will use 220V ac supply use are you still suggest to use toroid ?
2. In china made amp I measure mostly use 33Khz swithing freq I try to bring it up to 47Khz but its seem nothing changed ( heat, voltage output ) what is optimal freq use in car amp in your experience.
off the topic. here in indonesia with 30 to 60 USD you can buy an china made amp with attractive looking and bad sound my goal is made this thing sound better any sugestion especially in PS term.
I re-wound the switchmode transformer last night that i salvaged from a computer PSU. It's only a small one but i can now pull over 200W out of it. It seems the core material is definitely the limiting factor here. I'll now keep my eyes open for a larger E-transformer in old switchmodes as they are in abundance around here. It would be nice to try a toroid of good core material but it's looking as if i'm not going to be able to find one in the general junk i have access to. If anybody else is struggling with their cores dropping out too early like i was earlier in this thread i can recommend the switchmode E-types if you can get them apart in one piece. Although not as efficient as toroids, almost any reclaimed e-type out of old PSU's will have the correct material for what we are trying to do here. Up to now all the toroids i have salvaged have all been constructed from unknown and unsuitable material. The colours of the toroids dont seem to help either.
Hope my experience helps others out here. I'll report back when i find a larger E-core or a suitable toroid to get me up and over 500W
Regards
Leigh
Hope my experience helps others out here. I'll report back when i find a larger E-core or a suitable toroid to get me up and over 500W
Regards
Leigh
Hi Eva,
May I ask question about switchmode trafo.
1. why in big SMPS I found mostly use EI trafo and I see you dont like it use in car SMPS the topology perhaps........... what about SMPS for PA off course it will use 220V ac supply use are you still suggest to use toroid ?
2. In china made amp I measure mostly use 33Khz swithing freq I try to bring it up to 47Khz but its seem nothing changed ( heat, voltage output ) what is optimal freq use in car amp in your experience.
off the topic. here in indonesia with 30 to 60 USD you can buy an china made amp with attractive looking and bad sound my goal is made this thing sound better any sugestion especially in PS term.
If eva get to busy to answer my question can any one else anwer it for me ...............please..........................................................
May I ask question about switchmode trafo.
1. why in big SMPS I found mostly use EI trafo and I see you dont like it use in car SMPS the topology perhaps........... what about SMPS for PA off course it will use 220V ac supply use are you still suggest to use toroid ?
2. In china made amp I measure mostly use 33Khz swithing freq I try to bring it up to 47Khz but its seem nothing changed ( heat, voltage output ) what is optimal freq use in car amp in your experience.
off the topic. here in indonesia with 30 to 60 USD you can buy an china made amp with attractive looking and bad sound my goal is made this thing sound better any sugestion especially in PS term.
If eva get to busy to answer my question can any one else anwer it for me ...............please..........................................................
- Toroids have a low profile that fits better the internal height of the case of car amplifiers.
- Toroids may be small in size yet they can still accomodate heavy windings for high power output.
- Toroids produce less stray magnetic fields than E cores even when they are pushed close to saturation.
- Symmetrical (bifilar) high-current low turn count windings are much easier to make on toroids.
- Toroids result in lower leakage inductance than E cores, unless windings are interleaved, but this is quite difficult to apply to high current low turn count windings.
E cores are much more suitable when the input voltage is high because turn counts are high too and current requirements are much lower, thus allowing to easily split and interleave windings for reduced leakage inductance.
On the other hand, mains isolation is difficult to apply to toroids because the isolation barrier results in spaced windings and higher leakage inductance, but it is easy to apply to E cores and it makes little difference in performance.
If you do some investigation on car amplifiers, you will notice that all recent designs use toroids, only a few older amps used E cores. There are plenty of pictures of these amplifiers on the web.
- Toroids may be small in size yet they can still accomodate heavy windings for high power output.
- Toroids produce less stray magnetic fields than E cores even when they are pushed close to saturation.
- Symmetrical (bifilar) high-current low turn count windings are much easier to make on toroids.
- Toroids result in lower leakage inductance than E cores, unless windings are interleaved, but this is quite difficult to apply to high current low turn count windings.
E cores are much more suitable when the input voltage is high because turn counts are high too and current requirements are much lower, thus allowing to easily split and interleave windings for reduced leakage inductance.
On the other hand, mains isolation is difficult to apply to toroids because the isolation barrier results in spaced windings and higher leakage inductance, but it is easy to apply to E cores and it makes little difference in performance.
If you do some investigation on car amplifiers, you will notice that all recent designs use toroids, only a few older amps used E cores. There are plenty of pictures of these amplifiers on the web.
HI eva,
Thanks for the input thats really help.
if you don't mind I still have question to you.
In some amp I have opened I found mostly do not use a regulated power supply why ?
the dropping voltage in this PS is big 40% or more in peak bass ( the car battery just dro 0.5 volt in case like this what is your suggestion ?
changed the primary and secondary wire to smaller one but use many maybe????????????????
What is benefit or lose when we rise the freq of PWM osc let say from 33Khz to 48Khz with same parts ........
Thanks
Thanks for the input thats really help.
if you don't mind I still have question to you.
In some amp I have opened I found mostly do not use a regulated power supply why ?
the dropping voltage in this PS is big 40% or more in peak bass ( the car battery just dro 0.5 volt in case like this what is your suggestion ?
changed the primary and secondary wire to smaller one but use many maybe????????????????
What is benefit or lose when we rise the freq of PWM osc let say from 33Khz to 48Khz with same parts ........
Thanks
Rembulan,
That's the problem i've been having with my PSU's. The regulation is awful when nothing is done about it. Changing the core material helped a lot at first but now i've got to a point were changing the core dose not help. Using lower RDS fets has little effect too. Seems the transformer's are losing efficiency at high power. To combat this i've had to produce a transformer that gives 30% more voltage than i need and then regulate the voltage back down using PWM in a feedback loop. This is a double edged sword though because when the PWM is very tight ( low power ) the switching efficiency of the fets become compromised causing heating of the fets. I'm currently experimenting with missing pulse regulation instead of PWM to keep things under control. I'm sure we must be missing somthing with these transformers tho. Any clues Eva??
Keep D.I.Ying..
Leigh
That's the problem i've been having with my PSU's. The regulation is awful when nothing is done about it. Changing the core material helped a lot at first but now i've got to a point were changing the core dose not help. Using lower RDS fets has little effect too. Seems the transformer's are losing efficiency at high power. To combat this i've had to produce a transformer that gives 30% more voltage than i need and then regulate the voltage back down using PWM in a feedback loop. This is a double edged sword though because when the PWM is very tight ( low power ) the switching efficiency of the fets become compromised causing heating of the fets. I'm currently experimenting with missing pulse regulation instead of PWM to keep things under control. I'm sure we must be missing somthing with these transformers tho. Any clues Eva??
Keep D.I.Ying..
Leigh
Perry,
I think the cause of the problem is the back EMF being clamped by the FET's internal diode at low duty. This causes the FET to overheat due to the high volt drop across the diode. When running at higher duty the FET itself clamps the back EMF with next to no volt drop thus same clamping effect but much lower heat. I don't think there is a workaround for this. This is just how things are. Also the PWM can become so short as to cause the FET to spend more time in transition causing more heat. A minimum pulse period would be handy here but i may as well just go to missing pulse. Thats what i'm currently playing with.
As for the core i'm using i've got to be honest i don't know anything about it. All i can tell you is it came out of a 200W computer PSU. I'vere-wound it and funnily enough i can draw about 200W ( 170W actually ) out of it.
Thats about were i am now.
Regards
Leigh
I think the cause of the problem is the back EMF being clamped by the FET's internal diode at low duty. This causes the FET to overheat due to the high volt drop across the diode. When running at higher duty the FET itself clamps the back EMF with next to no volt drop thus same clamping effect but much lower heat. I don't think there is a workaround for this. This is just how things are. Also the PWM can become so short as to cause the FET to spend more time in transition causing more heat. A minimum pulse period would be handy here but i may as well just go to missing pulse. Thats what i'm currently playing with.
As for the core i'm using i've got to be honest i don't know anything about it. All i can tell you is it came out of a 200W computer PSU. I'vere-wound it and funnily enough i can draw about 200W ( 170W actually ) out of it.
Thats about were i am now.
Regards
Leigh
The way in which windings are placed around the core (and thus leakage inductance) has far more impact on the resulting voltage drop under load than any of the factors mentioned (except voltage drop at the input which is equally important). This subject is being discussed in the power supply section and some interesting papers were presented a couple days ago.
For high powers I recommend using several small transformers (but properly wound), preferably 1:1 or 2:1, with parallel primaries and series secondaries to enforce current sharing.
Ultimately, following the voltage with the oscilloscope will tell exactly where it is being dropped most.
Edit: A few parallel low-ESR capacitors are a must at both sides of the converter. Proper ground planes help too due to their low AC and DC resistances.
For high powers I recommend using several small transformers (but properly wound), preferably 1:1 or 2:1, with parallel primaries and series secondaries to enforce current sharing.
Ultimately, following the voltage with the oscilloscope will tell exactly where it is being dropped most.
Edit: A few parallel low-ESR capacitors are a must at both sides of the converter. Proper ground planes help too due to their low AC and DC resistances.
The core itself has nearly no impact on coupling and energy transfer, it's there just to add magnetizing inductance and avoid a huge current from flowing when 12V are switched onto the primaries. The wrong core will still result in good coupling if windings are laid out properly, but magnetizing current may be high leading to excess heating of everything.
Every time current flow is reversed, leakage inductance has to be discharged and energized in the opposite direction. This process takes some time and little or no energy is being transferred to the output while it lasts. Increasing the operating frequency results in more of these reversing events per second, more of this "dead time" and thus more leakage-inductance related voltage drop and losses. The lowest safe frequency should be used (its choice is what depends 100% on the core type).
Magnet wire AC resistance is also higher than DC resistance, particularly for heavy gauge solid wires. This implies that the effective resistance of the wires is higher than expected for a brief period after every current flow reversal event, thus conduction periods should be as long as possible.
Every time current flow is reversed, leakage inductance has to be discharged and energized in the opposite direction. This process takes some time and little or no energy is being transferred to the output while it lasts. Increasing the operating frequency results in more of these reversing events per second, more of this "dead time" and thus more leakage-inductance related voltage drop and losses. The lowest safe frequency should be used (its choice is what depends 100% on the core type).
Magnet wire AC resistance is also higher than DC resistance, particularly for heavy gauge solid wires. This implies that the effective resistance of the wires is higher than expected for a brief period after every current flow reversal event, thus conduction periods should be as long as possible.
The following waveforms are from a power supply at various duty cycles. As you can see, there are no spikes that need to be clamped by the FETs. As long as the gate drive voltage is above ~5v, the FETs will run cool. If the transformer is wound with too high of a ratio, the drive may be cut back so far (due to the regulation) that the FETs may be operating at the threshold and that can cause them to get hot.
http://www.bcae1.com/temp/aP1010031.jpg
http://www.bcae1.com/temp/bP1010028.jpg
http://www.bcae1.com/temp/cP1010029.jpg
http://www.bcae1.com/temp/dP1010030.jpg
Note:
This is without a significant load. The supply was connected to an audio amplifier (only drawing current due to the biasing of the outputs). The duty cycle was varied by adjusting the output voltage of the power supply.
http://www.bcae1.com/temp/aP1010031.jpg
http://www.bcae1.com/temp/bP1010028.jpg
http://www.bcae1.com/temp/cP1010029.jpg
http://www.bcae1.com/temp/dP1010030.jpg
Note:
This is without a significant load. The supply was connected to an audio amplifier (only drawing current due to the biasing of the outputs). The duty cycle was varied by adjusting the output voltage of the power supply.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.