| Offline SMPS duty cycle, core & bridge - Click HERE for Original Thread |
| JohnT_diy |
Hi.
I am looking into building an offline SMPS to replace multi-secondary multi-tap expensive amplifier power transformer... I am familiar with SMPS design with using 12V primary and in this case need to move onto 230VAC input to +50v_0v_-50v output, amongst other output voltages. Power range will be 500W or less.
I need a couple of things confirmed:
1. Can I "stack" transformer "E" cores (say 2 to ensure 500W throughput) inside the windings to give a larger core area? This gives a "rounded corner rectangle" shape to the windings. Also with the heaps of former and core I have, board space footprint will stay the same, just transformer gets higher. I have seen a 230VAC input, 12V/40A output design where this was done... Is it good practice??
2. Design will initially be half-bridge, using the usual 2 switching MOSFETs. Is it better to use 2x470uF in series (with parallel resistors) to generate the bridge center voltage, or, as I would prefer, 2x 1uF in series connected in parallel across a single 470uF/400V cap which is across the 320VDC. The center of the two 1uF's would give the bridge center voltage.
3. And lastly in the 12V input SMPS, I run them at nearly 50% duty cycle with unregulated output. This works OK for many of us and I am wondering if the same can be done for an offline half-bridge SMPS. This would simplify feedback (none) and DC output filtering.
Thanks in advance,
JT. |
|
|
| guitar_joe |
1. stacking cores is no problem... however you will have to fabricate your own bobbin... and keeping the transformer safe is more difficult. (more in accordance to minimum creepage distances and all those regulations)
2. I'd go with the first solution, most pc power supplies do it that way and they seem to work good.
3. Commercial power amplifiers have unregulated switchmode supplies, some have regulated. They both seem to work good. You can always start the design off without feedback and implement it later...
hope that helps |
|
|
| walkura |
Good evening .
Your question about the capacitors made me remind something .
Have a look at the following link and then especialy post #6 & #35 .
http://forum.radiospec.ru/index.php...opic=5879&st=20
Sorry for the Russian , but the schematics speak for itself .
(while your at it don't forget to have a quick peek at Post #34 cause it is beautyfull in its intention ;D )
As you see they do it both ways ,either 2 x 220 uF or 470 + 2 x 1 uF .
Basicly i think it boils down to the same .
Although i must say i didnt see the aproach of elco + 2 caps before ,i don't see a serious objection for this second aproach .
I will certainly try the 3 capacitor option myself for the simple reason that over the years i builded the impression that elco's age faster then non polarised (non elko) capacitors .
Have a nice eve Wakura |
|
|
| EWorkshop1708 |
Sounds great. An SMPS for an amplifier would also lighten the weight and may make it run cooler.
What about using a toroid for this as well as E-cores?
What turns ratio were you planning for this type of power supply for 50-0-50V?
My guess is 30 turns or 15-0-15 for the primary and 10-0-10 turns for the secondary. What were you going to use? |
|
|
| guitar_joe |
| quote: | Originally posted by EWorkshop1708
Sounds great. An SMPS for an amplifier would also lighten the weight and may make it run cooler.
What about using a toroid for this as well as E-cores?
What turns ratio were you planning for this type of power supply for 50-0-50V?
My guess is 30 turns or 15-0-15 for the primary and 10-0-10 turns for the secondary. What were you going to use? |
you don't have to worry about centre tapping the primary with half or full bridge |
|
|
| JohnT_diy |
Thanks everyone!
I'll go for the 50% duty cycle as it's so simple! I'm thinking the 230VAC is "reasonably" stable between 0->500W so output regulation should be OK. Yes there will be sags & peaks in the input voltage but I will only use an output over-voltage protection.
I don't think a single core will handle 500W - 300W was easy in a car amp... Center of E core is 12.54mm diameter. I won't stack the E cores yet, but will do later on as the board layout will allow it. This will make core area increase simple!
Will try 2x 200uF/200V in series to get the split bridge voltage. Being vertical mounting, they will be a smaller board size than 2x 1uF or 2uF
Transformer primary hasn't been calculated yet. Maybe around 20 turns or so, but definately won't be center-tapped... The single winding of half-bridge primary maximises copper usage for each switching phase. Again, so somple! |
|
|
| guitar_joe |
Do run it at lower voltages to observe ringing and spikes at the primary side.
If you don't sort these out then the supply won't have a chance at full line voltage! |
|
|
| megajocke |
| Computer PSUs use BOTH center tapped electrolytics and a small film cap in series to the center point. The time constant won't be short enough without the film cap. Using a small cap there in a unregulated supply has the additional benefeit of being able to tune the leakage inductance-series cap-resonance a bit higher than the switching freqency and get ZCS operation :) |
|
|
| Eva |
| Two film capacitors in series and in parallel with a big electrolytic will result in a nice resonator. Use two electrolytics in series and a film cap instead. |
|
|
| megajocke |
| That is a very good point :) |
|
|
| TechGuy |
[QUOTE]1. Can I "stack" transformer "E" cores (say 2 to ensure 500W throughput) inside the windings to give a larger core area? This gives a "rounded corner rectangle" shape to the windings. Also with the heaps of former and core I have, board space footprint will stay the same, just transformer gets higher. I have seen a 230VAC input, 12V/40A output design where this was done... Is it good practice??[QUOTE]
You can if you can wind them. Another option is to double the switching speed to permit double the power. Of course you need to check to see if your core material can support the much higher switching speed. If your going to buy a second core, perhaps you can purchase a core with a much higher switching speed.
The downside risks are:
1. Can the existing output diodes handle the higher speed.
2. your switching losses will increase (your efficiency will likely drop).
3. your controller may not be able to operate at a higher frequency. |
|
|
|
|
