About Chinese LLC SMPS

What could be the power of this power supply?
It has active PCF
 

Attachments

  • WhatsApp Image 2019-09-20 at 08.54.21(1).jpeg
    WhatsApp Image 2019-09-20 at 08.54.21(1).jpeg
    366.3 KB · Views: 840
  • WhatsApp Image 2019-09-20 at 08.54.21.jpeg
    WhatsApp Image 2019-09-20 at 08.54.21.jpeg
    295.2 KB · Views: 831
They have to cut corners somewhere to get the price down.

LLC SMPS are very good. As the transformer primary sees a sine wave and not a square wave there is much less radiation from the SMPS.

The resonant frequency is very important. If you move it by changing caps then you can shift the frequency into the capacitive high current mode and possibly destroy it.

I have the problem with simulating such a smps, not because I can not design it but the transformer who is from a pc supply I have calulated with ExcellentIT7300 looks like the voltages are way higher on the calculated coil values, I had 15 volts 15 amps, 300 volts 250 volts 250mA and 2 x 120 volts 250mA it is for a tube preamp to test the circlotron.

Maybe I did something wrong, it is a ETD 39/20/13 core, I have test all materials but did not change much, why is this? or has this core a gap? the pc supply was half bridge, I have used half bridge, for simplicity when making transformer.

I have simulate al kinds of freguencies, none did give satisfied results, always to high voltages, Maybe ringing on the rectifier I did used a full wave 4 diodes rectifier to spare to much windings.

thanks for eventual giving tips.

regards
 
Last edited:
the 500w china power supply is not bad but unter heavy load its thermal shuttdown due overheat of heatsink. Maybe a fan could help.


I think the moste heat comes from the rectifing diods. If the where doubled it would be much better. I hope I can sell my china PSU. I don't need it anymore.






My new 36V power supply has same L6599 but much more power and bigger transformer and double the rectifing diods. So less heat is generated.
 

Attachments

  • WhatsApp Image 2019-10-15 at 16.50.13(3).jpeg
    WhatsApp Image 2019-10-15 at 16.50.13(3).jpeg
    342.1 KB · Views: 706
I just received one of these ordered to be +-26VDC.

Not sure if it says anything, but I had a meter so I measured capacitance between input and output:
N/L to output GND 0,85nF@100Hz 0,23nF@10k
GND to output GND 0,55nF@100Hz 0,20nF@10kHz

Loading it with a 5,6Ohm resistor, voltage drops to 23,5V, which is more than I expected.
Ripple seems a bit high under load too. Do you think it would be safe to add some more capacitance to the output, or maybe LC?
 
I put in parallel wit R66(130kohm value on the board) one resistor with 680k value. According to my calculation, the value of the output voltage had to decrease from 2x61V to 2x52 V.
Because my target output voltages were 2x48V, I replaced the paralle resitor from the R66 with 470K. This should make the job of decreasing the voltage close to 2x48V but it doesn't. The voltage remains 2x52V. Any thoughts?
PS. I read on the internet that the feedback is made with TL431 even if i couldn't spot this IC on PCB.
 
Wel Piersma

What if I go use more voltages? like 15 volts 10 amps, 2 x 130 volts 250 mA 1 x 300 volts 100 mA and eventueel (dutch) 2 x 65 volts 15 amps.

In a resonance system, the feedback is needed then on 15 volts, become then 12.3 volts regulated for tube heater, or I can do open loop and use lineair regulator for tube heater.

Thanks in advance mine friend.

regards
 
Hi toghether,

iam starting a new projekt wich is TPA3255 based.
should be one Stereo Amplifier-board for 2 Klipsch R28F (8ohm / around 85 watts each)

At the moment iam thinking about different Power solutions, and i think for the Chinese Part of that, iam in the right Thread :)

Iam just wondering, whats the Difference between the here often used "meanwell lrs-350-48" MEAN WELL LRS 350 48 48V 7.3A meanwell LRS 350 350,4 W Einzigen Ausgang Schalt Netzteil|Schaltnetzteil| - AliExpress and the LLC SMPS version you are talking here about. ?

any difference there what you actually could hear?


Thanks for your helps ;)
 
Hey all,

Cool thread this one, thanks guys!

I am about to repair (again) but at a different place my SONY TA-F55. This time the PSU died, opened it, measured and understood: ok, there is a pair of hFE and Vbe matched transistors hard to find - amp is from 1978.

Now, since I already switched it from lamp to LED and swapped output transistors (high fT) by also high fT but Darlington from a Harman Kardon, I say: ok, let's change the PSU too.

I finally bought the SMPS500R, it's LLC soft switch based and uses an old friend the L6599A. L6599A - Improved high-voltage resonant controller - STMicroelectronics

A short review: Resonant SMPS for Audio | H i F i D U I N O

I took the +/- 50 V since it's what my SONY needs. By the way, it has a real sweet sound also the reason why I am not able to throw it away but keep it alive.

Sony TA-F55 on thevintageknob.org

BTW: I also own a very rare JVC A-X9 - if the Sony is concluding maybe I do same for the JVC. This one I also already swapped out all lamps and replaced by LEDs - needed so recalc of the circuits.

JVC A-X9 on thevintageknob.org

Let me know if you've experience with the SMPS500R from Connex:

SMPS500R | Connex Electronic

Chris
 
You will have a limited "bandwith" going up or down (voltagewise), the used transformer will determine this factor.

`Mine idea is now to use a separate smps for the poweramp, en for the high voltages.

I think that you mean more voltages lowering Q who is important in LCC smsp.

On the other hand, the smps is without feedback, maybe then bandwith is less importamt and can I still use one smps. for the tube I need then a 6.3 or 12.6 liniair. regulator.
 
Hello guy, i just received a new LLC power supply based on L6599. But it has big problem:

No load voltage is 42V as normal. But each little power draw will lower the voltage. For example if I plug my BRU5 TPA3255 amp voltage drops to 28V when music is listening it fals to 24V and lower... When more power is demanded it fails completly.
Have you an idea?

L6599D_2.jpeg
 
This is why it is a Chinese aliexpress supply, you need to switch the 42 to 24.

The products I do see on aliexpress are so bad, I did blow a lithium batteries with one of them
lithium protectors.

Just a;ways read the responses of the buyers first, otherwise you have a faulty supply, maybe it
is not on right frequency, we need the sweet spot to get a good stable voltage. I presume
this one is a feedbacked type, regulating and bad design can do nasty things.

measure the resonance coil and transformer and use a program like this to see what is happening.

I see it is a slotted bobine, so remove transformer and measure primairy with secondary shorted.

Auxiliary winding is only for the driver chip, just measure voltage on it, when supply is on needs to be 12 volts or so.
 

Attachments

  • RESONANT-CONVERTERS-LLC-DESIGN-TOOL-FOR-FSFR-SERIES.zip
    934.8 KB · Views: 66
Hi Hansueli.I put this smps and L15DX2 500w class d amp in this case(I am using it bridged).It plays well since 2 days.But there is a small sound (tırrrrrrr) from smps transformer.(not audible over 30cm)
It plays well now.For your information.I did not open a ventilation hole on top and bottom because it is not heating up.
I hope it will not break down.
Class D and smps can interfere, that is why these are synchronized sometimes. However with a resonance supply it is more difficult.
Also a resonance supply, special with feedback do not like low loads as with a amplifier who go as low as 400mA (idle) or less with class d (idle).

Much better system is a LCC these can do work very nice.

Burst mode can do make a sound, this will be entered as the amp is in low sound listening setting.
 
Last edited: