Hi people, I need to make a power supply converter to convert 200-250 VAC 50Hz to 115 VAC 60Hz with a useable power output of 800 Watts and a surge capability 300mS of 1200W, I assume the best approach is a flyback converter, I have built the output side of this unit thinking that this was the hard part and the flyback would be the simple part, I am having trouble calculating core size and getting reliable info on a suitable Flyback converter, ..some direction would be greatly appreciated..
TeeJay.
TeeJay.
Given the fact that both the input and output voltages are AC, wouldn't be much simpler to use an industry standard 60Hz autotransformer? Such autotransformers are not so big, as half the copper is shared between the input and the output.
Note that if you want to go for a SMPS, the most straightforward solution would be to convert 230V AC to DC first, then get a lower DC level through a buck or push-pull converter, and then convert the output DC voltage back to AC. The resulting circuit is going to be quite complex.
Note that if you want to go for a SMPS, the most straightforward solution would be to convert 230V AC to DC first, then get a lower DC level through a buck or push-pull converter, and then convert the output DC voltage back to AC. The resulting circuit is going to be quite complex.
Thankyou kindly for your reply Eva,
This is my FIRST ever FORUM so please let me know if I am not doing it correctly.
I discounted any normal transformer because of the frequency change needed 50Hz to 60Hz which has to be done to keep frequency sensitive timing operations (in the load to be driven) correct. I already assummed the mains to DC was required to use a Flyback converter, which I hope would supply a 2:1 transformer ratio to get my 115 V RMS 160V P giving me a current doubling as the voltage is halved to the secondary.
I have built 2 banks of MOSFETS driven by those nice HCPL3120 isolated drivers for the output side of things. These are driven by a micro which I have pulse width modulated to a sinusoidal effect for each bank in turn, separately for the top half and bottom half of the sinwave,... this section seems to work OK with a 60VDC up to 5 amps from a bench supply fed to it, I am hoping the high inductive load will be enough to filter the high frequency modulation out to end up with an averaged normal looking sinwave.
Q1. Is the most suitable/cheapest converter a Flyback?
Q2. Am I correct in assuming 1600uF Electros would be the smallest I could get away with after the mains rectifier due to acceptable ripple?
Q3. Can I gain a 2 to 1 current advantage with any other kind of converter?
This is my FIRST ever FORUM so please let me know if I am not doing it correctly.
I discounted any normal transformer because of the frequency change needed 50Hz to 60Hz which has to be done to keep frequency sensitive timing operations (in the load to be driven) correct. I already assummed the mains to DC was required to use a Flyback converter, which I hope would supply a 2:1 transformer ratio to get my 115 V RMS 160V P giving me a current doubling as the voltage is halved to the secondary.
I have built 2 banks of MOSFETS driven by those nice HCPL3120 isolated drivers for the output side of things. These are driven by a micro which I have pulse width modulated to a sinusoidal effect for each bank in turn, separately for the top half and bottom half of the sinwave,... this section seems to work OK with a 60VDC up to 5 amps from a bench supply fed to it, I am hoping the high inductive load will be enough to filter the high frequency modulation out to end up with an averaged normal looking sinwave.
Q1. Is the most suitable/cheapest converter a Flyback?
Q2. Am I correct in assuming 1600uF Electros would be the smallest I could get away with after the mains rectifier due to acceptable ripple?
Q3. Can I gain a 2 to 1 current advantage with any other kind of converter?
Have you considered an unregulated half-bridge converter with a 1:1 transformer? It's really simple, but the microcontroller would have to work harder in order to get rid of the 100Hz ripple in the top of the 60Hz squarewave.
The next option is a regulated half-bridge with a transformer having slightly more turns in the secondary than in the primary.
Concerning the flyback, it looks simpler at first sight, but have you calculated the magnitude of the peak currents involved? It's true that in a flyback only a single switch is required, but this switch has to turn off when it is conducting far more current than the ones from push-pull topologies in similar conditions, and has to whitstand far more voltage. Also, a single output diode and a single transformer are required, but the diode is subject to quite high peak currents requiring a more expensive device and yielding high losses, and the transformer is actually a coupled inductor that requires both high inductance and high saturation current, thus it's bigger than any push-pull transformer for a similar power level.
The next option is a regulated half-bridge with a transformer having slightly more turns in the secondary than in the primary.
Concerning the flyback, it looks simpler at first sight, but have you calculated the magnitude of the peak currents involved? It's true that in a flyback only a single switch is required, but this switch has to turn off when it is conducting far more current than the ones from push-pull topologies in similar conditions, and has to whitstand far more voltage. Also, a single output diode and a single transformer are required, but the diode is subject to quite high peak currents requiring a more expensive device and yielding high losses, and the transformer is actually a coupled inductor that requires both high inductance and high saturation current, thus it's bigger than any push-pull transformer for a similar power level.
For your 800W continuous / 1200W peak, a flyback seems not the most suitable topology. Your primary side switches will about 600V (leakage inductance spikes not included), and must carry peak currents around 15 amps or so. Given the conduction losses of high-voltage switching elements, this is not optimal. On the secondary, currents are double and voltages are half. Still quite a lot. Plus, you need either a massive core or very high switching frequencies.
Personally I would doubt between a full bridge and a half bridge.
If this is a commercial design, you would need a PFC stage to pass the tests anyway. You could try to use that stage to obtain the required DC bus voltage. Maybe a SEPIC?
Personally I would doubt between a full bridge and a half bridge.
If this is a commercial design, you would need a PFC stage to pass the tests anyway. You could try to use that stage to obtain the required DC bus voltage. Maybe a SEPIC?
Eva said:
Hi Eva, I am a bit confused about the 100Hz ripple I assume thats from the 50 Hz mains in full wave rect,
I need to end up with a Sinewave output and I need to keep things EMI clean as I have lots of radio receiving equip not far from this when it is finished.
Q1. Can I treat this as 2 separate sections, Section 1.(240VAC 50Hz to 160DC) Section 2. (Sinewave converter)
or
do you think I need to re-look at the sinwave conversion method with respect to the converter type?
DaBit said:
Hi DaBit, I need this to be as small as possible, Cool as possible, definately no fans! and clean EMI wise because of its proximity to my radio equip, I will be completing this unit as fast as I can get problems solved, thanks so much for your reply.
Q1. What is your best guess of Topology ?
In EMI terms, a half bridge is usually much quieter than a flyback at the same power level.
Adding regulation to the half bridge would probably increase EMI a bit, since the output diodes would now be always conducting (provided you chose continuous mode operation). However, producing a sine wave from a regulated supply voltage seems easier to me than producing it from an unregulated supply. Are you using any kind of feedback into the microcontroller in order to correct sinewave linearity?
Anyway, the worst thing that could happen with an unregulated supply is to have the 60Hz wave slightly intermodulated by the 100Hz ripple from the 50Hz mains rectification (and this is not EMI, it's just LF ripple). What are you going to power with that converter?
Adding regulation to the half bridge would probably increase EMI a bit, since the output diodes would now be always conducting (provided you chose continuous mode operation). However, producing a sine wave from a regulated supply voltage seems easier to me than producing it from an unregulated supply. Are you using any kind of feedback into the microcontroller in order to correct sinewave linearity?
Anyway, the worst thing that could happen with an unregulated supply is to have the 60Hz wave slightly intermodulated by the 100Hz ripple from the 50Hz mains rectification (and this is not EMI, it's just LF ripple). What are you going to power with that converter?
TeeJay said:
We all need that
The easiest way to get this going is a mains rectifier, followed by a half bridge converter which in turn feeds the sinewave PWM. But do you need to comply with some regulations, or is it for private use? I can imagine that filtering out conducted EMI would be troublesome at these powerlevels with the regular full bridge AC rectifier and storage caps.
If you must have your design certified, you might want to use a PFC preregulator. Normally they use a boost type converter to a DC voltage of 350-400 volts, and a flyback/halfbridge/fullbrdge/whatever to go down to the required voltage.
In your case the DC bus voltage is pretty high, so I was wondering if it would be possible to collapse the PFC preregulator and actual DC/DC into one stage. I have seen an appnote somewhere which uses a SEPIC to go down to a lower DC bus voltage.
About regulated or unregulated: Personally I would use a regulated supply, simply because I don't like stage A influencing stage B. Optimizing two different stages separately is easier than optimizing two stages together.
Eva said:
Attached Is what I have made so far for sinewave production, should I trash this for a bridge vesion?
The micro supplies a sinusoidal PWM signal to the load which when veiwed on the CRO with an Inductor in series with the load looks like a clean sinewave! No feedback, will I need some?
DaBit said:
Sounds like a good starting point, it seems from all the great feedback that I should have started with the AC to DC side first, I will start building a proto of a half bridge version for the front end AC to DC section and see how effective that is, yes DaBit ... I do need to have this unit meet standard EMC specs This unit is to power an incorrect voltage machine that I accidently ordered a number of with the wrong voltage 115VAC instead of 240VAC it is to costly to return them so I am trying to recover my sanity as well as serious losses, at the same time as learning more about my No 1 hobby!! ELECTRONICS.
Q 1. What do you mean COLLAPSE THE PFC regulator?
Q 2. Is the PFC a stand alone circuit that can be applied to what ever type of converter I use or is it specific to the converter type?
Ah. Take a course in voodoo at your local medicine man, that will help you coping with the Eliminate Minor Competitors guidelines.
Oops. That's the same 'you f**cked up; you solve' that costed me my weekend.
First of all I want you to know that I never ever designed or built a PFC preregulator. In my last SMPS project I solved that problem by specification of the PSU.
Now, there is an output power limit above which power factor (cos phi) restrictions apply. Check your local regulations for the exact numbers; the power supplies I normally buy, modify or, if nothing else is possible, design are for the military, for ships, etc, so other regulations apply. The average AC line -> EMI filter -> bridge rectifier -> storage caps setup needs a serious filter to get the PFC at the required values. Thus, above some tens of Watts active PFC circuitry become cheaper and more compact to implement than passive PFC circuitry.
The most common PFC technique I see around is the boost preregulator. Thus, the incoming sinewave is filtered and rectified, and then followed by a boost type converter which steps up the voltage to some 350VDC..400VDC. The special thing about those circuits is that they try to mimic a resistor, thus they sense the line voltage and draw more current when the voltage is higher and less current when the voltage is lower.
Now, the boost type circuit is quite simple, but is has a disadvantage: the DC bus voltage must be higher than the maximum momentary input voltage (Example: If you accept 230V +/- 20%, that would be 230 * 1.2 * sqrt(2) = 390V).
Now, a DC bus of, say, 400VDC is workable, but you must follow it with a step-down regulator to get to your output voltage. If your output voltage is 5VDC, you don't have much choice either.
But in your case the output voltage is 115 * sqrt(2) + some extra to cover up losses, thus about 170V. I'm not sure about it, but I feel in my belly that it would be in the range of an alternative PFC regulator structure. And then the first topology that comes in mind is the SEPIC since that is quite close to the boost, but capable of going to both a higher and a lower output voltage than the input voltage (buck-boost).
A flyback has about the same properties, so I suppose that would be useable too.
Now, should you investigate this further? If you are the kind of guy who likes to figure out a nice circuit to go from A to D directly instead of knitting together blocks to go from A to B to C to D, yes. If you plan to produce many circuits, then yes, since the extra engineering effort might lower the production costs.
But if you only need a few or some 10's of the converter, I would save myself the trouble and focus my effort on building a 1200W PFC preregulator, a DC/DC which does the 350..400V -> 170V conversion, and the sine shaper. And when finished, I would connect these blocks in series, and voila, a working device.
the PFC is a standalone circuit which has the job of producing a preregulated DC voltage and drawing a sinusodial current, in-phase with voltage, from the (mains) input. So yes, it can be used as a frontend to whatever topology you choose for the final downconversion of the voltage.
About your sinewave converter: are you sure it works like it should? It seems that it only generates a nice sinewave when loaded resistively.
peranders said:
Hi PA, I did mention a bit about what I am powering it is more than a bit embarrasing ... I ordered the incorrect voltage/freq machines from the US. the frequency is critical because of very Synchronous motors used .... If I cannot put together a compact efficient power supply to solve this problem I am in serious trouble, the return freight would cost mega $ so I am trying extremely hard to do this, not being an expert in this field, only a very keen tryer, learning fast.
They need to be reasonable good devices ... tuff call but I do not give up easily.
Hi DaBit, I have only tested it with resistive loads you are absolutely correct, I figured I could put an Inductor in series with the load output and a bit of C across the load terminal to clean away the high frequency PWM used to get my sinewave!! I was not up to seeing the effect of a large Inductive load yet with current and voltage phase diff, but have since decided from a lot of reseach and feedback from YOU GUYS (This is good fun here)...... that the general concenses is to re attack this problem.DaBit said:
1. Design a full bridge converter getting the 2:1 current advantage ....
(the mains current is in danger of being to high for the standard power outlet in this country especially if the PFC is not corrected)
.... followed by a bridge arrangement for the sinewave production using the same micro to drive these FETS and see how close I can get things with this approach!!!
I have not been able to locate any full bridge controllers that are able to drive the high side fets above 75V
Am I to assume that I use one of these controllers and make it drive some Isolated FET drivers for the FULL bridge converters,
OR ..
DOES ANYONE KNOW OF A FULL BRIDGE CONTROLLER WITH BUILT IN ISOLATED DRIVE CAPABLE OF OFF LINE levels 350VDC feeding the FET bridge?
Any comments on this direction for this project ...regards TJ
For such high voltages, you have to use a discrete high side drive solution or a separate driver IC.
Also, have you considered modifying the appliance for 230V 50Hz operation instead of building converters? Why does your load rely so much in 60Hz power? If it has synchronous motors, then you can just drive them with 60Hz square waves and find a simpler solution for powering the rest. Is a sine wave really required? Note that 120V synchronous motors may be driven with a 230V square wave having just 50% duty cycle.
Also, have you considered modifying the appliance for 230V 50Hz operation instead of building converters? Why does your load rely so much in 60Hz power? If it has synchronous motors, then you can just drive them with 60Hz square waves and find a simpler solution for powering the rest. Is a sine wave really required? Note that 120V synchronous motors may be driven with a 230V square wave having just 50% duty cycle.
peranders said:Considering your knowledge and the voltage in mind I suggest that you get yourself a good old transformer.
C'mon PA, dont be like that, you have to get standard transformers out of your head definately, you keep forgetting the serious frequency issue .... I may not be a design engineer, but I have been servicing Radio TV & HiFi gear for 30 Years, Along with Industrial custom designed embedded controllers Past history ... I no longer do that now.
So forget about high voltages being a worry, I have been working on live chassis and live 3 phase switchboards 415VAC half my working life,
Anyones serious help where ever they can is extremely appreciated.
I am doing this regardless, I am so Impressed with the bits and pieces of help from some professionals on this forum ... It gives me a good feeling to think that professional people are willing to spend some of there time helping others .... thanks to all so far TJ
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.