UCC28780: DC24V 400W output with two LMG3410

Ulrich-san,

No.

Many customers network standby load is 1.6W.

So Flyback cct. must clear 80% AC to DC efficiency at 1.6W DC load.

I have measured UCC28780EVM-021 with high accuracy power meter, the result is under 80% at 1.6W DC load.

You have to understand the condition of AC230V is very low efficiency AC120V is very easy.

The rules should be measured at AC230V , it's a European rules.

Japanese maker's Flyback IC can clear this. So we have no chance for W/W network standby consumer product.

In few years, all products will be connected to internet.

Hello Doi-san,

I am aware that efficiency is measured at 230Vac 50Hz in most of the world, but I did not know that network stand-by load is typically1.6W. 

Can you also tell me please what is the typical full-load rating of the DC5V output?  Using the UCC28780EVM-021 board, 1.6W represents ~3.6% of full load of a 45-W, 20-V output.  This EVM is a general purpose evaluation module and was not designed for network equipment, so it is not optimized for those specific conditions.  The controller alone does not determine a power supply's efficiency, rather its control methods plus appropriate power stage design interactively contribute to the efficiency performance over line and load.  In the case of network equipment applications, the power stage may be designed differently than the EVM.  Comparing the EVM to a supply with a Japanese controller that is already optimized for the application puts the EVM to a disadvantage.

If you able to, kindly provide the full design requirements (specification) of the auxiliary supply being considered (DC5V, IC bias outputs, etc.).  We can see whether a revision of the power stage using the UCC28780 controller can be optimized for this application and meet the 2-W rule in stand-by, and all of the other requirements of the supply spec at their respective operating conditions.

Regards,
Ulrich

  

Ulrich-san,

I think you are misunderstanding.
Aux. flyback IC, now meet 2W rule, is very cheep.
Even though UCC28730 is expensive.
UCC28780 and two LMG3410R150 is planning for 300kHz. 24V 400W output.
400W cct.itself, can not meet 80% eff. At 1.6W load.

And you understand the 2W rule is 2019 rule.
Evaluating cct. design status condition is not open.
Only TI assigned customer can contract NDA with TI.

Hello Doi-san,

Yes, I agree, I must be misunderstanding something.

This is how I understand the situation based on your first and second postings:
1. There is a network system main power converter rated for 400 W at 24 V output, intending to use PFC, ACF, and GaN Fets.
2. A part of this converter is a flyback auxiliary supply to provide bias power to the control ICs of the main converter and to provide DC5V to some load. Full load for DC5V = XX-W (unknown).
3. When in stand-by mode, the main converter is turned off, the bias power is also turned off, and the DC5V load is reduced to 1.6W.
4. When in stand-by mode, total input power must be < 2W.
5. When in stand-by mode, input efficiency must be > 80%, since eff. = (P_24V + P_bias + P_DC5V)/P_in = (0 + 0 + 1.6W)/2W = 0.80.

The question whether the Aux supply can deliver 1.6W at >80% efficiency or not depends on the full power rating of the AUX supply. If 1.6W stand-by power represents ~25% or more of the total Aux power rating, then I think the UCC28704 controller can meet the 2-W spec. If (XX-W + P_bias) > ~8W, then I am not sure whether one of the UCC287xx family of controllers can meet >80% without a synchronous rectifier for the DC5V output.

So, to my understanding, the total rated load on the Aux supply will influence its efficiency during stand-by, when its load is only 1.6W.
If the rated load is 8W or less, then 1.6W is > 25% of rated load, and >80% should be achievable.
If the rated load is more than 8W, then 1.6W is < 25% of rated load and the Aux supply is operating in a light-load range where efficiency begins to drop off. In this case it is difficult to predict whether 80% can be achieved or not.

Since you have a Japanese controller that does achieve 80% at 1.6W output on DC5V, I suspect the full rated Aux power is somewhere around ~8W. And I think that TI has controllers that also can help achieve the required efficiency. The controller alone does not determine the efficiency of the Aux supply, but the control method together with a power stage design that is optimized for that controller.

Regards,
Ulrich

Ulrich-san,

Main major MFP maker never accept PSR. They request accuracy.

And the 5V normal requirement is much over 8W.

So TI do not have cost competitive Flyback IC for MFP(Multi Function Printer).

Few month ago, I already said this situation to HVP manager.

Hello Doi-san,

I did not know PSR was unacceptable, or that the application is for MFP. These are new information.

In any case, I cannot contest your conclusion without attempting to make a design to the full spec, and I am not in a position to do so.
Neither can I concede that TI does NOT have a solution for your application, for the same reason.
Perhaps TI's Power Design Services can contacted for a reference design.

Given this present situation, I will close this E2E thread.

Regards,
Ulrich

A power unit suplire, world big four MFP power unit making, said PSR desnot meet customers spec.