Now that I have had more study I am inclined to try to unify the PFC and transformer step, rather than a separate boost PFC, bulk capacitor then LLC driven transformer.
There is some benefit in an obvious, step by step solution but the more direct path looks better here.
It's what Cuk proposed and now I see it more clearly.
Not sure of the details, presumably resonant to minimise losses so the principle should still apply.
I have done sims of switch mode amplifiers and GaN was much more efficient so I have a bit of a prejudice.
But that was around a MHz so the benefits will be less at lower frequencies of a power supply.
On the other hand, it's no so much the frequency as the speed of the transitions.
This will need sims to decide once I have settled on the details of the circuit.
Best wishes
David
It's not just three phase, Fourier says a half rectified sine wave has plenty of harmonics.
They may even be worse than the three phase, I meant to check this when someone else raised the same point earlier.
But it didn't seem sufficiently important to do yet - I just see it as a non-problem, a decent amp will have plenty of PSRR and it's not hard to filter anyway.
Maybe you could fire up LTspice and compare the harmonic spectra?
Best wishes
David
I have been busy at work and had this on the back burner but now have a little time to continue on the project.
I have decided to rectify the 3 phase and reference to Neutral so I have + and - 340 V or so, and run a converter on each side.
Essentially two separate 240 V AC converters, which simplifies life if I ever move to a spot with no 3 phase, at less scary potentials rather than the full ~700V, and I can use commodity inverters and circuits.
So I have a new thread on the inverter circuitry, or one aspect of it to start with - https://www.diyaudio.com/community/...power-electronics-errors.387414/#post-7052679
Kind of surprised it's had little response so far, I expected plenty of "what would I know compared to a respected author like Erickson?" criticism
But anyone can make mistakes, so it will be fun to see where the truth is.
Bet wishes
David
I have decided to rectify the 3 phase and reference to Neutral so I have + and - 340 V or so, and run a converter on each side.
Essentially two separate 240 V AC converters, which simplifies life if I ever move to a spot with no 3 phase, at less scary potentials rather than the full ~700V, and I can use commodity inverters and circuits.
So I have a new thread on the inverter circuitry, or one aspect of it to start with - https://www.diyaudio.com/community/...power-electronics-errors.387414/#post-7052679
Kind of surprised it's had little response so far, I expected plenty of "what would I know compared to a respected author like Erickson?" criticism
But anyone can make mistakes, so it will be fun to see where the truth is.
Bet wishes
David
Filter capacitors ripple current is rated for the mains frequency application (max 120Hz). You need to get a different type of capacitor for higher ripple frequency.
Or The voltage rating of the capacitor must be increased to limit that ripple current at a higher frequency.
Regards.
Please show an example of a "filter" capacitor.
As far as I'm concerned, a "filter" capacitor is a capacitor that is being used in a filter. There is no such thing as a "filter capacitor" as far as product catagories on sites like Mouser go, for instance.
Here is an example of a capacitor I use as a filter capacitor; It is rated for 1.15A of ripple @ 100 kHz and must be derated to 40% of that for 50 Hz operation. It is a 100uF/500V long life 105°C capacitor.
I don't read Chinese, but it's obvious to me what this means:
https://datasheet.lcsc.com/lcsc/1912111437_Ymin-LKMJ3552H101MF_C443409.pdf
As far as I'm concerned, a "filter" capacitor is a capacitor that is being used in a filter. There is no such thing as a "filter capacitor" as far as product catagories on sites like Mouser go, for instance.
Here is an example of a capacitor I use as a filter capacitor; It is rated for 1.15A of ripple @ 100 kHz and must be derated to 40% of that for 50 Hz operation. It is a 100uF/500V long life 105°C capacitor.
I don't read Chinese, but it's obvious to me what this means:
https://datasheet.lcsc.com/lcsc/1912111437_Ymin-LKMJ3552H101MF_C443409.pdf
Of course, you can use most of the capacitors anywhere one likes. When need to think of longer life, things are different to look.
I come from different region and for me a filter capacitor is power supply filter capacitor. I also say some as decoupling capacitor as the function applies, some as power supply decoupling capacitor. I won't be looking for a decoupling capacitor from the manufacturer search.
Now that many manufacturers have stopped mentioning certain application specific products. For example, MOSFET output transistors. The IRF series were never meant for audio applications in their data sheets and many use as a output device successfully. So if someone's application permits, you can wisely choose a device and use it.
At lower frequencies the ESR is low and higher frequencies the ESR is higher. This makes a ripple current specified at 100 Khz to be derated at lower frequency to maintain power lost or maintain temperature (Power losses are to the Square of current times ESR, Losses higher at lower frequency)
The ripple currents are specified at 20 deg C and capacitor life is shortened at higher temperatures even for the 105 deg C types.
Two 5,000 mfd instead of one 10,000 mfd capacitor gets longer life.
Higher voltage capacitors (larger size and easier heat dissipation) will have their operating temperature lowered if used in this application.
Typical number of hours of standard capacitor at 80-90 deg C is about 2000 hours.
Regards.
I come from different region and for me a filter capacitor is power supply filter capacitor. I also say some as decoupling capacitor as the function applies, some as power supply decoupling capacitor. I won't be looking for a decoupling capacitor from the manufacturer search.
Now that many manufacturers have stopped mentioning certain application specific products. For example, MOSFET output transistors. The IRF series were never meant for audio applications in their data sheets and many use as a output device successfully. So if someone's application permits, you can wisely choose a device and use it.
At lower frequencies the ESR is low and higher frequencies the ESR is higher. This makes a ripple current specified at 100 Khz to be derated at lower frequency to maintain power lost or maintain temperature (Power losses are to the Square of current times ESR, Losses higher at lower frequency)
The ripple currents are specified at 20 deg C and capacitor life is shortened at higher temperatures even for the 105 deg C types.
Two 5,000 mfd instead of one 10,000 mfd capacitor gets longer life.
Higher voltage capacitors (larger size and easier heat dissipation) will have their operating temperature lowered if used in this application.
Typical number of hours of standard capacitor at 80-90 deg C is about 2000 hours.
Regards.
I don't really need mains frequency filter capacitors at all.
I can just run a converter directly off the rectified 3 phase, there is always at least one phase to supply power.
A Cuk converter can do boost as well as buck, looks attractive to help track and stabilise the rectified input.
That's part of the reason to use 3 phase in this application - possible to completely avoid expensive and life limited mains frequency capacitors.
But probably better to have a least some minimal input filter capacitors.
I don't think the difference in ripple capacity between 50 Hz (actually 100 Hz and harmonics once it's half wave rectified) and three phase rectified is much of a problem, unless you have data otherwise?
Best wishes
David
Last edited: