I will see how I get on with my current pcb.
It should work with one cut and one link to sort out the star grounding.
Thanks for the offer.
Befora checking the whole circuit just check the controller.
Use an external 12V supply to supply it. Be sure to use a <10 ohm resistor insted of the lower FET drain and source connection (otherwise the IC's OC protection turns on). Apply some dummy cpaactiros (4,7nF ceramic) instead of the FET gates, and check the capacitor votlage with scope probes.
Make a screen capture (for digital scope) or photo the bildschirm (analog scope).
I did that previously using 30 volts to supply the mosfets and 18 volts to supply the 27952. I got a square wave out on HO and LO.
I got my smd rework gun today.
I soldered in a new chip.
I checked it on 30 volts to mosfets and 18 volts to 27951 and it outputs a square wave on LO and also a signal on HO.
I now have a transformer with 200uH primary inductance and 75uH leakage inductance.
The resonant capacitor is 22nf.
This works out at resonant frequencies of 60KHz and 100KHz the same as the app note project.
I get a brief voltage on the output on power up then the 27951 goes into power down mode.
What should I try next ?
I soldered in a new chip.
I checked it on 30 volts to mosfets and 18 volts to 27951 and it outputs a square wave on LO and also a signal on HO.
I now have a transformer with 200uH primary inductance and 75uH leakage inductance.
The resonant capacitor is 22nf.
This works out at resonant frequencies of 60KHz and 100KHz the same as the app note project.
I get a brief voltage on the output on power up then the 27951 goes into power down mode.
What should I try next ?
I have also ordered a etd49 transformer bobbin and cores to try to get he correct inductance. The etd49 is used in the app note.
Here is the LO output of the 27951
Here is the LO output of the 27951
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
You have a primary inductance of 200uH without gap? If you have gap, try to decrease the gap to get an inductance of 4-500uH. Then you will have a k about 5, so a resonant capacitor of 33nF will match to 250W.
Can you measure the parasitic capacitances of the transformer primary? (use the LC meter as a capacitor meter on the primary) It would be helpful to set the maximal switching frequency. Your minimal switching frequency (with 75uH 500uH and 33nF) combo will be about 70kHz.
It is 250uH without a gap. and 95uH leakage inductance.
With my meter cross the primary I get 70pf.
My new transformer transformer will be here on Monday, I will try the etd49 with 13 primary turns as you suggested earlier.
I worked out the timing resistor values and now the circuit outputs 40 volts ok.
See the excel sheet I sent you previously. I use that to calculate turns.
Your MATH is BAD!
The inductance (and all impedance related parameter) is proportional to the SQUARE of the turns.
Your other math is bad. The two resonant frequencies are NOT 60,8kHz and 65kHz.The lower resonant frequency should be calculated from the resonant capacitor and the total primary inductance (that is 60,8kHz), HOWEVER the upper resonant frequency should be calculated from the leakage inductance and the resonant capacitance (it gets you 173kHz). The last one is the MAIN resonant frequency.
Also the timing resistors are not set to these frequencies, Nigel, have you really understood the application note?
First you must calcuate the magnetizing current and from that the needed dead time according to FET's equivalent Coss and gate charges. This gives you the timing capacitor. Then you must calculate the minimum frequency required to maintain output stabilization (its neither the one of the previous freqs), to set the minimum freq. Then you must set the max frequency. A good rule of thumb is to set it 1,5 times the main resonant freq. It is highly limited by the transformer parasitic capacitances.
Then you must set the softstart freq and time. I reccomend you a soft start time of at least 20ms, and the soft start freq should be at least 1,7 times the main resonant freq (the IRS has a reccomended max of 500kHz).
The circuit wont power up because the lack of soft start. At power up there is a short at secondary (due to secondary el. capacitors) and the IRS protection turns on. It has almost noting to do with star ground.
Please, please, please! Use the excel spreadsheet I sent you. It does the math for you.
Look at the ferrite material datasheets, if you do you will see that there is some problem with your inductance metering.
Hello people,
through this thread I have become aware of the IRS2795x. I will also try to build a resonant switching power supply.
Do you know this German site:Dimensionierung von Schaltnetzteilen
It is for dimensioning of switching power supplies. You do not need speak German to use this site.
through this thread I have become aware of the IRS2795x. I will also try to build a resonant switching power supply.
Do you know this German site:Dimensionierung von Schaltnetzteilen
It is for dimensioning of switching power supplies. You do not need speak German to use this site.
Nigel, please use the spreadsheet (thanks lorylaci, it is very useful 
to get it right.
Or atleast do it right on paper first.
and thoroughly read through teh app note.
Without softstart the circuit will never get running, because it will trip on overcurrent on the first few switchcycles. It only runs at low voltage because then the peak currents will never trip the overcurrent. Otherwise you would have had to replace primary silicon repeatedly .
regards.
Rickard
edit: I think that Prof Schmidt-Walters webpage has been up on discussions before, as there was some errors in the calculations or somehting like that but I can be misstake as my memory is very short and probably incorrect. However it is useful as a conceptual design base (ie test different topologies)
to get it right. Or atleast do it right on paper first.
and thoroughly read through teh app note.
Without softstart the circuit will never get running, because it will trip on overcurrent on the first few switchcycles. It only runs at low voltage because then the peak currents will never trip the overcurrent. Otherwise you would have had to replace primary silicon repeatedly .
regards.
Rickard
edit: I think that Prof Schmidt-Walters webpage has been up on discussions before, as there was some errors in the calculations or somehting like that but I can be misstake as my memory is very short and probably incorrect. However it is useful as a conceptual design base (ie test different topologies)
Last edited:
O know this site, the professor has some lecture notes to there. However there is no lecture note about LLC converters.
If you can understand the operation of an LLC converter, can understand the ZVS switching, then building a resonant supply is very easy.
Due to the ZVS resonant switching layout is less important, you do not need to get rid of spikes etc.
spreadsheet (thanks lorylaci, it is very useful
You're welcome! The spreadsheet has lot of extras, whcih are not in IRS application not, and sometimes I use different claculations.
For example the AN from IR calculates with the nominal Coss of the FET. Since MOSFET Coss is highly nonliner, it is reccomended to at least calculate an approximate linear equivalent ( Cosseq=sqrt(V_datasheet/V)*Coss_datasheet), or use the equivalent values from the datasheet: Coss_er, Coss_tr.
Also the application has an error calculating the MOSFET switching off losses. MOSFET switchoff losses are not really related to Coss, however it is related the current flowing before the switch off. So far I have not been able to find a good approximation with a simple equation.
The spreadsheet also calcluates core losses, whcih is very helpsome. You can put in ferrite data on a different sheet. Also MOSFET losses is made more punctual (Rdson mosfet die temperature rise dependence also calculated). You can get the needed heatsink info etc...
I always use spreadsheets wherever I can, because as you can see there are lot of variables, so it is good to see what changes what, this can be easiyl done in the spreadsheet.
I am only sorry that it is halfway translated, maybe later I finish the translation. (and also google translator is your friend till then)
I have looked at the spreadsheet but all it does is input values.
I cant see where Rmax and Rmin etc are calculated.
Thanks for spotting the resonant frequency mistake.
I have used the control components from the app note so soft start should be fine. This is also why I set min resonant frequency to around 60000 to match the app note transformer.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.