Good luck!
But first, plan and design the values!
I will help you if you are stuck.
The transformer windings seem to be the first decision.
I have used 36 turns primary with two 8 turn secondaries.
Is it ok to stay with this or do I need to change it ?
My resonant capacitor is currently 22nf.
What is the next step ?
What is the input voltage? Is there a PFC? What is the output voltage?
Output turn ratio is the first deicision not the turns.
Input voltage is 320 VDC.
No PFC just common mode choke plus 2 100nf capacitors for a filter.
Turns ratio needs to be 4:1. 320/4 = 80 volts max.
I am looking to get +/- 40 volts DC out but it is variable from 30 to 50 volts via a pot in the feedback loop.
Last edited:
You have a 230V AC line in England. Then input voltage will be 290V - 330V
(you ahve to yount with high line conditions, and the depletion of the primary capacitors).
Irs27951 has to stabilzie output even at high input votlage so for 80V output voltage you will need at least n=4,125 turn ratio.
For 40V you need a turn ratio of 8,25.
You won't able to set it in such a range with a pot. It is not a PWM converter. (it can set in a range of +/- 10% depending of gain of the LLC circuit)
The pot is on the secondary TL431 circuit to set when the opto coupler output goes low.
If all I can get is a set voltage then 40 volts should be fine.
I simply used a zener, and managed to get <5% voltage drop from no load to full load. With the pot you will be able to fine tune the circuit with the TL431, to get less than 2% voltage drop.
I decided to go with the app note example transformer and split the bobbin as you suggested.
I got 1.5 mH primary inductance with 500uH leakage inductance.
This is way higher than the app note got.
So I gapped the core and now get 768uH primary inductance and 500uH leakage inductance.
The circuit still wont run.
I got 1.5 mH primary inductance with 500uH leakage inductance.
This is way higher than the app note got.
So I gapped the core and now get 768uH primary inductance and 500uH leakage inductance.
The circuit still wont run.
These values are unrealistic. What do you measure your inductance with?
How do you short the secondaries? (you have to short all of them)
After you designed the resonant elements, you need to design the controll elements.
For the split bobbin there is an approximation for the leakage inductance. Measere the winding width of there half of the bobbin. For ETD34 the full width is 20,9mm, half of the bobbin width the split is about 9mm. Calculate the inductance with this gap. AL=(gap^K2)*K1 so AL=32nH for 9mm gap, so leakage inductance will be about 42uH.
This is only an approximation, and only stands for split bobbins.
So there must be an error with your measurements.
Last edited:
I have an LC meter.
Yes I shorted both secondaries at the same time.
I am using an ETD54 transformer so perhaps the extra size is giving me extra inductance ?
I put together another pcb.
Just my luck it wouldn't work at all, not even on low DC volts.
So I removed the 27951 and put one in a SOIC to dip converter and soldered that into the pcb, and it burst into life.
I can only guess I fried or static discharged the previous 27951.
I have 27952 in the circuit now and that works although the transformer buzzes a little.
I just need now to take a step back and ensure the transformer and series capacitor are correct.
An ETD54?????? That's for >10kW thing. I made >400W from an ETD29.
According to my approxiamtion your leakage should be about 100uH. 500uH can only be if you wind it really ugly.
500 uH is extremely large, there is something wrong with your transformer. Could you make a picture of it? I reccomend getting the error out first, since this is a hazrdous thing. Also you will not get regoulation without the proper setting of the resonant elements, and ALSO you can easily get ZCS or hard-switching without knowing it, and can fry your FEts (and controller also). ZVS switching canonly be set with the contorl elements, when you have set the resonant elements.
The transformer buzz is beacuase of the magnetizing current and the ugly winding.
I was hoping to get 800 watts hence the large transformer.
It looks like there is one winding but there is the primary about 2/3 across the width and 2 secondaries in the last third.
Can I not use the transformer as is but change the resonant C to match ?
It looks like there is one winding but there is the primary about 2/3 across the width and 2 secondaries in the last third.
Can I not use the transformer as is but change the resonant C to match ?
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
OMG, don't use that transformer. there is one hazardous errror and other errors.
- You did not use any insulation. Put the bobbin splitter at the middle of the bobbin, and use some plastic or paper as a splitter. Glue it with epoxy resin or other glue with good insualting properties.-
- After winding the core put some good insulation on the winding. Best is 3M yellow tapes, Kapton (by DuPont).
Without this insulation your transformer is LETHAL. The polyurethane insulation on the wire is only stated up to 500V. You need to get insulation over 2kV if you ground the secondary and the projected bos, otherwise at least 3kV insulation is needed.
- You used thick wire. Due to skin depth, this thick wire will only conduct at 10% or less of its area. Calculate the skin depth, use lots of thin wire wounded togethet 8strnaded wire). I usually use 10s of 0,2mm wire twisted together.
If you want to get over 800W you need to complete redesign the resonant circuit, the FETs and the controll elements.
What is the material of your core. For example an EPCOS ETD54 with N87 is stated for about 4 kW at 100kHz, and an N97 is stated for about 5kW at 100 kHz.
Your FETs can provide up to 300W.
Read the application note carefully, Nigel, everything is in it. All I know is from IT (okay, and from the UCC25600 apllication note and the TEA1713 application note). Try to understand the operation.
Feel free to ask questions, I will help.
Since its a converter working on line voltage, I reccomend you to take extra caution. I don't want you to have an injury or other.
What Cr value did you put in instead of 22nF?
In this way you cannot know if there is ZVS switching (unless you measure the gate signal with a signal at different load conditions).
Also load regulation will be bad.
Make a spreadsheet, calculate your conditions.
what happens if this converter goes into overcurrent state, does the voltage.on the half bridge caps start tilting, ie one goes up to rail and.the other to zero. I can imagine that could lead to catastrophic failure and loss of magic smoke.
(annhilating primary silicon
If the resonant network is properly designed, the IRS2795x overcurrnet protection will shut down the converter.
At a case of secondary short (after secodnary capacitors), the curent in the primary starts to increase. The speed of increase is slwoed down by the resonant network. Usually withon 2-3 cycles the urrent on the lower FET increases to, where its votlage drop is over 2V (or 3V for IRS27952), then the IC stops switching.
Just for curiosity: when the switching is stopped, the current in the resonant network circuilates (between the transofmer, resonant caps and FETs Coss) until the resistive (and other) losses causes it to go to about zero.
There are other overcurrent options (using diodes on slplit resonant capacitors, monitoring the resonant capacitor votlage, using the shunt) and the LLC is very easy to make short-circuit proof, due to the LLC network (it limits current ramping).
All of my supplies were tested to be short circuit nad overload proof. I also design them, taht at max contionous load, it shunts down when the FETs heat up to about 80°C.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.