so basically you have to cycle powerto reset the converter. LLC converters is voltage mode so that figures .
One thought: as it is difficult to get an exact leakageinductance in the xformer and therefor Lr, how about adding an external inductor for the fine tuning of Lr? say 20-50% of total Lr? yes, more components, more board area but perhaps in the end less hassle. Only problem is that the.flux swing will be quite large and therefor demands low loss material or airgaps. (first thoght was a simple iron core but loss is probably to big) .
to bad I cant read hungarian...
One thought: as it is difficult to get an exact leakageinductance in the xformer and therefor Lr, how about adding an external inductor for the fine tuning of Lr? say 20-50% of total Lr? yes, more components, more board area but perhaps in the end less hassle. Only problem is that the.flux swing will be quite large and therefor demands low loss material or airgaps. (first thoght was a simple iron core but loss is probably to big) .
to bad I cant read hungarian...
Yes for IRS2795x you need to reemoce it's supply and put back to restart.
For bigger power designs it is usual to have a standby supply (a quasiresonant flyback) to supply basic stuff at secodary, and the IRS2795x. If you put a switch to the IRS2795x suply then you can activate and daeactivate standby mode, and also reset after fault.
You can use extra inductor, but that costs more, takes more space.
If you have a bit expereince it is very easy to get the exact leakage inductance you want. Also you can change the controll elements for your resonant network
(you can change magnetizin inductance and resonant capacitance easily)
Also there is a special method of putting some magnetic shunt into the transformer to set exact leakage inductance. (google for 1MHz LLC converter)
Its quite difficult to get an exact leakage inductance.
I just wound the transformer then measured it.
With a secondary wound on top of primary I got quite a low leakage inductance. With the secondary wound next to the primary it became very high. I guess the compromise is to wind the secondary partly on the primary.
I also found I was getting a high primary inductance I guess due to using an ETD54 core. I had to gap the transformer to bring the inductance down.
It just needed a very small gap.
I am open to suggestions for getting the right primary and leakage inductance using an ET54 transformer.
I guess I could partly overlap the windings to lose a bit of leakage inductance.
What is the material of your ferrite? Epcos N87?
At the frequency range of the application note (60-150 kHz), you will find it hard to deisgn it to a low power level of 240W.
Nigel, you are not deisgning, just quessing. Do calculations, use math.
I am working on the translation of my excel spreadsheet to English for you.
I built the transformer as suggested in the app notes so I am not guessing.
It should be pretty easy just matching Lr, Lm and Cr.
I have Lr and Lm so just need to match Cr.
The control components have already been given to me by the app note.
I used open circuit secondaries to measure primary inductance.
I used short circuit secondaries to measure leakage inductance.
This method is in the application notes.
Since i have no idea about checking the resonant signal for tat chip, or as described in the AN,
If you let me know how they confirm the resonance in that SMPS, i will be able to help you save lot of time make it working
please show waves
regards
I would assume that the capacitor and leakage inductance make a tuned circuit.
At resonance the circuit will pass much more signal than it would off resonance giving control of how much power is given out.
The primary inductance is 400uH and the leakage inductance is 365uH.
So the resonant frequency is:
1/ 2 pi sqr(lc)
Therefore:
c=1/4 pi pi l f f
Or c= 7nf for 100kHz
At resonance the circuit will pass much more signal than it would off resonance giving control of how much power is given out.
The primary inductance is 400uH and the leakage inductance is 365uH.
So the resonant frequency is:
1/ 2 pi sqr(lc)
Therefore:
c=1/4 pi pi l f f
Or c= 7nf for 100kHz
It is a PFM controller, so varies the frequency. There is no tuning to one frequency.
Since there are lot of input values, tere is always some degree of freedom at scaling.
Scopeing the signal on the resoanant capacitors shows a sine from light to heavy loads (at idle or small loads its triangle like, due the magnetizing current), confirms the resonant operation.
Scopeing the half bridge voltage and gate signals confirms ZVS switching.
However by calculation ZVS switching can be guaranted over full load range.
I have just looked over that chip, its LLC.
I am still not sure where to measure, in order to make sure that the resonant is happening, and the circuit is running well.
Also the transformer you are using will waste your time making this circuit stable, even if you made it working.
You need the split type bobbin, side by side windings.
Read the application note again. There is NOT ONE, but two resonant frequencies.
Your induectance values are not good (your magnetizing inductance is smaller then the leakage). You need to rewind your transformer.
attached you will find tha half-translated excel calculator.
As you can see, I put in 300-330V input, 50V and 20A output. This is quite small from an ETD54 (please Nigel, tell the material of your ferrite, it is really important).
You can see, that 13 turns of primary is needed. This will take down your leakage inductance to a normal values. For 1kW the leakage inductance should be 18 uH (for this frequency range). (I must mention that your FETs will not take over 200W, due to the internal protection of the IRS IC)
Rewind the transformer primary for 13 turns. For 50V secondary you need 4 turns. Then measure the leakage inductance and the magnetizing inductance.
Then we can move forward.
The leakage is definitely smaller it is 365uH.
The main inductance is 400uH.
This gives a k of slightly more than 1 which isn't brilliant.
I wound the transformer primary -= 36 turns and secondary -= 4 turns.
From this I get 45 volts output.
If I have time Wednesday night I will rewind it.
On my previous SMPS which was a fly-back SMPS I had problems stopping the transformer whistling at high loads. I think the problem was the transformer being over wound and causing the switching frequency to be very low.
It looks like I have made the same mistake with the half bridge SMPS of over winding the transformer. I took the windings from the app notes of a example SMPS.
It looks like I have made the same mistake with the half bridge SMPS of over winding the transformer. I took the windings from the app notes of a example SMPS.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.