솔루션 찾기 기술 지원

TOP262EN output current limit why 12A

Posted by: r_x_p_u on

Trying to design an ac-dc converter, 300W using TOP262en

Input range : 220Vac-240Vac

Output: 7V 40A

PI expert gives an error messege:

Selection: input parameter IO1_CONT is equal to 40A

Problem: input is greater than the max limit of 12A

Fix: modify input parameter and repeat

 

My question is: Why is the output limit 12A? Can we change it to 40A? (how?)

7V @40A is 280Watts which seems that TOP262EN can handle in open frame with fan cooling.

 

Best regards.

 

 

 

댓글

Submitted by PI-Wrench on 11/10/2021

40A output is rather unreasonable/impractical for a flyback supply, as the secondary will see in excess of 80A peak when the primary switch turns off. The output ripple current in the secondary filter caps will also be enormous, and will require a veritable forest of caps for reasonable lifetime.  Secondary layout will also be difficult due to the high peak current. This application would be better addressed with a forward-based design using Hiper-TFS with a Hiper-PFS front end. The design is still challenging due to the high DC output current, but at least the secondary peak currents will be within reason, and the output capacitor ripple current will be much lower. 

Submitted by r_x_p_u on 11/11/2021

Thank you for the detailed explanation.

We tried to switch to TOP262EN rather tahn HiperTFS because of delivery problems.

Mouser,Farnell and Rs have long delivery times.

Do you have any information where to get the parts quicker? 

(We are in Turkey)

Best regards.

 

Submitted by PI-Wrench on 11/12/2021

I have an inquiry out about TFS2 availability. Another possibility might be using the Hiper-LCS with a PFS-4 front end. The disadvantage of using an LLC approach would be the need for large, high strand count Litz wire for the secondary winding, with all the associated handling/manufacturing issues, as well as an output ripple current of ~17A for the filter caps. Both the forward design and the LLC design would have lower output losses using synchronous rectification. One possible approach would be a design similar to the DER-385, which is a 12V, 255W design using synchronous rectification and a PFC front end. 

Submitted by r_x_p_u on 11/13/2021

If we stay by flyback topology is it possible to use MLCCs at the secondary output to face the high ripple and then use afterwards a small inductor + electrolytics caps.

 

(any negative issues with MLCCs?)

 

We have several applications such as 2.5V output and 7V output.

 

If we stay by flyback (lower cost, easy to implement etc) is 7V @25A with TOP262EN reasonable? (instead of 40A as yo do not reccomend)

 

We have also another application 2V @50A , Would you reccomend using TOP262EN @50A secondary output if the secondary Vmax is 2V

 

Is there a way to tweak the output current limitation of 12A in PIexpert software for lower Voltages (such as 2V to 1.5V) to see the results

 

Can it be changed to 25A?

Submitted by PI-Wrench on 11/15/2021

MLCCs would not be a good choice for a flyback output. They have high ripple current ratings, but there would not be sufficient capacitance for an affordable capacitor array to support the output during the primary switch on-time, where the output current is supplied solely by the output filter capacitors. Delta V =( i*delta t)/C. Aluminum polymer capacitors would be a better choice, with higher capacitance values and high ripple current rating for a small can size.

Having said that, these low voltage/high current applications are not well served using a flyback topology. The high peak currents magnify the effects of any stray inductance on the secondary side, and can cause problems with the primary snubber, as the reflected stray secondary inductance adds to the effective leakage on the primary side. 

As a "rule of thumb", the peak current on the secondary side of a flyback converter will be at the very least twice the output current, and the capacitor ripple current will at the very least be equal to the output current. This is illustrated by a limiting case thought experiment with 50% duty cycle and high primary inductance, so the current looks like a square wave. During the 50% primary switch on-time, the output current is supplied entirely by the output filter capacitors. When the primary switch turns off, in order to maintain a constant output voltage, the current supplied to the output filter capacitor must be sufficient to not only replenish the charge in the output filter capacitor, but also to supply the output current. This makes the required current twice the output. This situation is the best case for an impractically continuous flyback at 50% duty cycle. More practical designs will fare worse.

These considerations are why PI Expert limits the output current to 12A. It may be possible with a lot of struggle to get a 50A output design to work, but the results will be less than optimal. With a forward-based design, the peak current requirements are relaxed, and the output ripple current requirements are also considerably relaxed. For your 2V application, it would be a good idea to add synchronous rectification to get around the power dissipated in the output rectifier forward drop, which will be high compared to the output voltage and will have a significant impact on efficiency.

Submitted by r_x_p_u on 11/16/2021

Thank you for the detail answer.

3 suggestions:

1 - Adding a non dissipative LC + diode snubber to the primary (this will allow more room for the reflected secondary )

2-  Adding a cascaded output filter just like   cheap Film cap (lower value that faces the tide wave 1uF 16V x1) + cheap Mlccs (16V 1uF x20) + a pcb trace based small inductor + bulk cheap eletrolytic

3- Allowing a higher output ripple. Our apllication is tolerant to output current ripple (it is a kind of precise heating instrument) , So the effective power output is much important than current ripple.

 

According to 1+2+3  , How are the advantages of flyback (transformer acts like an inductor, simple to implement, better feed back stability etc.) 

 

In fact if Power Integrations would design a new SicMosfet based flyback family (with a 1200V or 1700V SicMosfet) operating up to 1 Mhz, some drawbacks of flyback will change into advantages.

(very small and cheap caps even for high ripple at secondary output @ 1Mhz  etc.) . Or a GaN alternative with a non dissipative LC snubber at the primary.

 

Would be a perfect hit.

 

Thanks for your help.

Submitted by PI-Wrench on 11/18/2021

Any problems present at 100 kHz or so will be magnified by operation at 1 MHz. This application is better served with a forward topology than a flyback. An illustrative example in point is a personal computer power supply. They are generally forward - based because of the requirement for high current at low output voltage. A Hiper-TFS2 design can be run at high line without a PFC, a te cost of some efficiency, though the design will run better with a PFC to supply a relatively constant DC input voltage.

Submitted by r_x_p_u on 11/20/2021

I totally agree. The main problem is the delivery issue of TFS7708H. (very long delivery time). We can find a few TFS7707H for prototyping purposes. 

I hope we find a way to get reasonable amount of chips.

We have the ability to use forced fan cooling with oversized Heatsinks. What may be the practical continious power with an oversized Heatsink (Fan cooled)

We have 3 major application:

Are they possible with above oversized Heatsink with fan cooling using TFS7708H (what abaout TFS7707H ?)

 

Application 1 : 7V 50 A   (350W output)

Application 2 : 5V 70A   (350W output)

Application 3 : 2V 175A  (350W output)

 

Are they feasable? Is there more room above 350W for an oversized Heatsink with fan cooling?

 

Thank for your help.

 

 

 

Submitted by PI-Wrench on 11/23/2021

The big problem on all these designs will be the reduced efficiency at low output voltages due to the voltage drop of the output rectifier. This will be especially evident for the 2V design. They will all need a substantial air-cooled heat sink for the output rectifiers. I have included  3 preliminary design spreadsheets for the 7V, 5V, and 2V designs, as well as a PFS front end preliminary design that has sufficient power output to drive all 3 of the forward DC-DC converters. The forward spreadsheets assume use of a TFS7708, but could be run with a TFS7707 at reduced power. The spreadsheets assume use of a 125u Sendust toroid with magnet wire for the output inductor. These high current designs in my opinion would best be served instead instead by a Sendust E core with a foil winding. Choose a core with similar AL to that of the toroid called out in each design. The standby supply chip included in the Hiper-TFS2 device can be used to generate a supply for the cooling fans as well as bias power for the TFS2 and PFS4. I would encourage you to download PI Expert and use it to verify the designs as well as doing any necessary changes according to your needs.

Submitted by r_x_p_u on 11/23/2021

Thank you for the useful and detailed information.

Best regards.