Helo.
I am working at schematic below. this supply will be used in audio amplifier, and must pass such test.
Output voltage +/- 80 V DC
Load used, Halogen lamp 1000W 220V.
Connected the load (one lamp) for 5 hours with no problem, stable,
Output voltage +/- 80 V DC, 3.5AMP
Connected two lamps 1000W Mosfets Blown in 7 minutes.
Output voltage +/- 76 V DC, 6.65AMP
The transformer temperature is 85C when connecting the two lamps as the load.
I need help in:
1- Is it normal for the transformer to heat that much?
2- Is there any way to minimize the heat of the transformer?
3- Why the mosfets blow up?
4- Primary Turns 20 1mm , Secondary 12, 12 Turns 1.2mm
Appreciate the help
I am working at schematic below. this supply will be used in audio amplifier, and must pass such test.
Output voltage +/- 80 V DC
Load used, Halogen lamp 1000W 220V.
Connected the load (one lamp) for 5 hours with no problem, stable,
Output voltage +/- 80 V DC, 3.5AMP
Connected two lamps 1000W Mosfets Blown in 7 minutes.
Output voltage +/- 76 V DC, 6.65AMP
The transformer temperature is 85C when connecting the two lamps as the load.
I need help in:
1- Is it normal for the transformer to heat that much?
2- Is there any way to minimize the heat of the transformer?
3- Why the mosfets blow up?
4- Primary Turns 20 1mm , Secondary 12, 12 Turns 1.2mm
Appreciate the help
I don't know the exact reason, butyou may like to correct some thing:
- mosfets may need individual D-S RC snubbers
- bootstrap capacitance is very high
otherwise it looks OK for me, nice design (unless I overlooked something)
The things you describe may have something to do with winding technique and transformer parasitics.
- mosfets may need individual D-S RC snubbers
- bootstrap capacitance is very high
otherwise it looks OK for me, nice design (unless I overlooked something)
The things you describe may have something to do with winding technique and transformer parasitics.
I would use 2x 1mm pro primary and secondary
Can you say that does heat, core itself or wires and core with it?
half power, does core heat up?
If you really have edt59, use 2x 1mm primary, 3x 1mm secondary, more wire the better in terms of heat, I think core itself does not heat up..
Also that 1uF cap in series, you could add one more
And yes, do use 100uf instead 22uf for IR
You could use snubber right on trafo, not behind rectifiers
What is your freq.?
Can you say that does heat, core itself or wires and core with it?
half power, does core heat up?
If you really have edt59, use 2x 1mm primary, 3x 1mm secondary, more wire the better in terms of heat, I think core itself does not heat up..
Also that 1uF cap in series, you could add one more
And yes, do use 100uf instead 22uf for IR
You could use snubber right on trafo, not behind rectifiers
What is your freq.?
Transformer
Heloo.
Thanks for you all,
I think that I have some problems with the transformer winding, I am not an expert in transformers, but will Re wind it as LUKA recommends.
Also, there is a snubber circuit for mosfets, and protection diodes at the gates for transients and surge protection.
I have noticed that the Upper mosfet in the schematic, the one connected to the +320V line, Heats more than the other one.
I think that the transformer design in any SMPS is 70% the success and efficiency of it.
I have a design from professional Amplifiers manufacturer using two IGBTs, and delivers 3000W RMS, with a few components, really simple.
but the transformer in there is not simple. (other wise any one can do it)
Any advice about transformer winding?
Heloo.
Thanks for you all,
I think that I have some problems with the transformer winding, I am not an expert in transformers, but will Re wind it as LUKA recommends.
Also, there is a snubber circuit for mosfets, and protection diodes at the gates for transients and surge protection.
I have noticed that the Upper mosfet in the schematic, the one connected to the +320V line, Heats more than the other one.
I think that the transformer design in any SMPS is 70% the success and efficiency of it.
I have a design from professional Amplifiers manufacturer using two IGBTs, and delivers 3000W RMS, with a few components, really simple.
but the transformer in there is not simple. (other wise any one can do it)
Any advice about transformer winding?
Amp
hi all...
Today I will receive some faulty Audio Amplifiers, (from known audio manufacturers) 2KW to 3KW
I will explore them part by part, make all the tests I want, to discover all the hidden secrets, Transformer type, winding, isolation, every thing
to make a power full SMPS.
I will try to post some pictures.
hi all...
Today I will receive some faulty Audio Amplifiers, (from known audio manufacturers) 2KW to 3KW
I will explore them part by part, make all the tests I want, to discover all the hidden secrets, Transformer type, winding, isolation, every thing
to make a power full SMPS.
I will try to post some pictures.
I don't like the turns ratio.
the original design (ignoring the fact thats its centre tapped) is 24 turns primary, to 18 turns secondary, which is 1 : 0.75 which should be able to give +-100volt rails, too much!!!
you've done the transformer differently, making it 20 turns primary to 24 turns secondary! your powersupply has the ability to give more output voltage than input! the switchmode controller will have to be working overtime to keep the rails regulated.
I think 20 turns primary is too few windings as well.
the original design (ignoring the fact thats its centre tapped) is 24 turns primary, to 18 turns secondary, which is 1 : 0.75 which should be able to give +-100volt rails, too much!!!
you've done the transformer differently, making it 20 turns primary to 24 turns secondary! your powersupply has the ability to give more output voltage than input! the switchmode controller will have to be working overtime to keep the rails regulated.
I think 20 turns primary is too few windings as well.
luka said:
EPCOS AG (for example) gives approximate maximum power limits for each core type. But these values do not take into account delivering high power at low voltages; required wire cross-sectional area will therefore determine the suitable core (window) size.
I have designed transformers starting with the following equation (that can be derived from basic identities):
Npri = D*ULpri,max / 2*Bmax*Ae*fs
Here D is the maximum pulse width at Uin,max and ULpri,max is the maximum voltage across the primary coil. Bmax is the selected maximum flux density, Ae the core effective area and fs switching frequency.
It can be seen that the number of primary turns is dictated by the maximum input voltage: On the other hand, the primary wire gauge is selected on the basis of primary RMS current. The coil turns and wire gauges should be selected so that completely full layers are wound on the bobbin. The secondary turns & gauges are defined by the selected transform ratio and RMS currents.
Core heating means that there is large ac flux within the core. Check the manufacturer's core data sheets, there are graphs that estimate the power loss density vs. ac flux or vs. frequency. Increasing the turns will reduce the flux density, but will increase copper losses. So there is a trade-off between these two losses.
I'm working on a design that gives +/- 63 V @ 8 A (total 1 kW) from a PFC stage, so the input voltage will vary somewhere between 300 - 400 volts. I'm using ETD54 bobbin with N97 core. At 50 kHz switching frequency the estimated (rough) core ac flux is 170 mT peak. Npri = 54, Nsec = 2 x 16. Wire gauges: primary bifilar 0.9mm, secondary maybe bifilar 120/0.1 LITZ wire (120 strands of 0.1mm each). Copper losses make up roughly 4 watts.
I started with the ETD49 but it turned out that in order to meet the isolation requirements, all the wires did barely fit within the bobbin! So I moved on to the next larger bobbin. By the way, using vertical bobbin usually improves the heat transfer, although vertical bobbins at this size are not easy to find.
Cheers,
mflorin said:
Indeed it is, I wrote "maximum pulse width at Uin,max" meaning simply the maximum pulse width that happens at maximum input voltage. I usually check it from the simulator and it tends to be somewhere between 0.4 - 0.7 (when the maximum duty at Uin,min is 0.9).
If you look at the equation, there is the maximum volt-second product D*ULpri,max/fs = D*ULpri,max*Ts. This was my point.
I'm sorry but I won't give the complete design here as I intend to sell this design onwards. However, I will help you as much I can!
Maybe these are of some help:
-Topology: Full bridge (conventional)
-Switching frequency: 50 kHz (25 kHz per switching device)
-Input voltage: 320 V - 420 V (from PFC stage)
If your local grid supplies 110 Vac, you might find a suitable PFC stage giving roughly 190 Vdc. That means you'll have twice the primary side current compared to my design.
Get yourself LTSpice simulation software from Linear, that's what I'm using for rough simulations. I will attach a link to a file comprising a full bridge converter with dual 63 V output. The circuit has a very crude voltage-mode control circuitry so you can simulate various dynamic changes both at the input and output. The reference voltage V2 must equal the voltage across R5 divided from the output voltage via R1, R6 & C11.
To be of best help, I would like to know something about your background in electronics/power electronics.
p.s. I would still like to remind that designing a high-power offline converter isn't easy task and it requires broad knowledge and understanding in theory and in practical issues as well.
Full bridge simulation model
Maybe these are of some help:
-Topology: Full bridge (conventional)
-Switching frequency: 50 kHz (25 kHz per switching device)
-Input voltage: 320 V - 420 V (from PFC stage)
If your local grid supplies 110 Vac, you might find a suitable PFC stage giving roughly 190 Vdc. That means you'll have twice the primary side current compared to my design.
Get yourself LTSpice simulation software from Linear, that's what I'm using for rough simulations. I will attach a link to a file comprising a full bridge converter with dual 63 V output. The circuit has a very crude voltage-mode control circuitry so you can simulate various dynamic changes both at the input and output. The reference voltage V2 must equal the voltage across R5 divided from the output voltage via R1, R6 & C11.
To be of best help, I would like to know something about your background in electronics/power electronics.
p.s. I would still like to remind that designing a high-power offline converter isn't easy task and it requires broad knowledge and understanding in theory and in practical issues as well.
Full bridge simulation model
Thanks Bootstrapper for offering to help. 
Although I am an EE that has been playing with electronics since I was 12 years old, I need all the help I can get with SMPS.
For the last 20+ years I make my living as an acoustical consultant and electronics became my hobby/therapy.
I already got the best books available and along with those I have been studying application notes from the major players. I learned quite a bit from other members here, but my practical experience is close to none when it comes to SMPS.
My interest is basically to make SMPS for audio applications - mainly subwoofer amplifier power supplies.
I typically use Multisim, but I will get LTSpice (again) and will play with the simulation file you kindly offered.
Thanks again and I wish you all the best in getting your design efforts "converted" to Euros.
===========================
Sorry to all for the OT.
Although I am an EE that has been playing with electronics since I was 12 years old, I need all the help I can get with SMPS.
For the last 20+ years I make my living as an acoustical consultant and electronics became my hobby/therapy.
I already got the best books available and along with those I have been studying application notes from the major players. I learned quite a bit from other members here, but my practical experience is close to none when it comes to SMPS.
My interest is basically to make SMPS for audio applications - mainly subwoofer amplifier power supplies.
I typically use Multisim, but I will get LTSpice (again) and will play with the simulation file you kindly offered.
Thanks again and I wish you all the best in getting your design efforts "converted" to Euros.
===========================
Sorry to all for the OT.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.