I have one of these LLC resonant half-bridge converter tube amp supplies. It provides
Here's level vs frequency for the HT supply
The bias supply looks worse.
I was expecting to need voltage regulators to clean up switching noise at ~100kHz+, but not mains harmonics or the humps at 1.5kHz, 2.8kHz 4kHz etc (no idea what's causing them).
The mains is rectified and filtered with 3 470uF capacitors before the resulting DC plus ripple is switched in the vicinity of 100kHz. 100kHz HT is rectified and filtered with a capacitor/inductor filter. The bias supply is half-wave rectified, filtered and regulated with a simple single transistor regulator.
The schematic for the converter is standard. The control daughter-board is coated so it's not possible to identify the controller and supporting components. Output level would be regulated with a feedback loop to the LLC resonant controller. No idea what it's time constant/s are. Perhaps the rectified mains ripple would go straight through.
Switch-mode power supplies can be noisy but LLC resonant converters with zero voltage switching of the MOSFET switches and zero current switching of the rectifiers shouldn't be. With good design and PCB layout performance should be acceptable. They're widely used in flat screen TVs, EVs, solar and even commercial HiFi.
The voltage regulators on the HT and bias do a good job but... would you expect mains harmonics out of such a converter?
- HT: 200V to 520V 300W maximum.
- DC1 and DC2: 2.5V-6.5V adjustable 15W peak 7.5A.
- DC3: fixed output 2.5V-15V, customizable, maximum 10W (not sure how it can be both fixed and customisable!)
- DC4: -10V to -120V 50mA maximum
Here's level vs frequency for the HT supply
The bias supply looks worse.
I was expecting to need voltage regulators to clean up switching noise at ~100kHz+, but not mains harmonics or the humps at 1.5kHz, 2.8kHz 4kHz etc (no idea what's causing them).
The mains is rectified and filtered with 3 470uF capacitors before the resulting DC plus ripple is switched in the vicinity of 100kHz. 100kHz HT is rectified and filtered with a capacitor/inductor filter. The bias supply is half-wave rectified, filtered and regulated with a simple single transistor regulator.
The schematic for the converter is standard. The control daughter-board is coated so it's not possible to identify the controller and supporting components. Output level would be regulated with a feedback loop to the LLC resonant controller. No idea what it's time constant/s are. Perhaps the rectified mains ripple would go straight through.
Switch-mode power supplies can be noisy but LLC resonant converters with zero voltage switching of the MOSFET switches and zero current switching of the rectifiers shouldn't be. With good design and PCB layout performance should be acceptable. They're widely used in flat screen TVs, EVs, solar and even commercial HiFi.
The voltage regulators on the HT and bias do a good job but... would you expect mains harmonics out of such a converter?
Well resonant converters are used because they are more efficient, they still pump out tonnes of RFI if not properly shielded and screened, although should be much less at higher frequecies (think ~10MHz) as waveforms are rather more like sine waves and less like square waves.
Those big humps could mean the supply is in the chaotic regime with lots of period-doubling effects leading to energy at subharmonics. If the load isn't constant it will interact with the converter and intermodulate it too. It looks like this design of converter isn't suitable for audio use.
Because mains switched converters run from rectified mains (which the minimum of smoothing needed to bridge the gaps between half cycles), they will carry plenty of mains harmonic components. You'd have to use a linear regulator after the smoothing cap to eliminate that, negating the whole point of having an efficient SMPS.
Thanks Mark.
I am planning to test a linear regulator between the mains rectification and filtering and the converter to confirm whether or not the limited filtering is the problem. Hopefully, the loss of the regulator's dropout voltage won't be sufficient to lower the frequency of operation into the capacitive ZCS zone.
There are more reasons than just efficiency to use a SMPS. My motivation is avoiding the weight, volume and cost of linear supply. I may yet have to resort to a linear supply though.
Dave.
I am planning to test a linear regulator between the mains rectification and filtering and the converter to confirm whether or not the limited filtering is the problem. Hopefully, the loss of the regulator's dropout voltage won't be sufficient to lower the frequency of operation into the capacitive ZCS zone.
There are more reasons than just efficiency to use a SMPS. My motivation is avoiding the weight, volume and cost of linear supply. I may yet have to resort to a linear supply though.
Dave.