UCC256404: ucc256404 STARTUP

robin gangopadhya18

Part Number: UCC256404
Other Parts Discussed in Thread: UCC256402, UCC28740, UCC25600, UCC28782, UCC28780

Hello:

While looking into the eval kit of UCC256404 LLC controller, is there a possibility of starting it up from the input power line itself as it is done in almost all controllers.

In the Web BAsed design, there is "battery" like symbol. Of it is not practical to do so is any commercial product.

Also, some applications show an elaborate "small" converter to get 12v from AC. As in the Eval kit.

The paradox is that the present design using UCC256404 will be itself a " small" converter of 30W, cost-competitive. Furthermore, the dc power input is 600V. The paradox becomes acute since small biasing converters are flyback...which can start by trickle current from the input line...but 600V input DC is problematic in flyback mode even if we can use 1200V device.

It is not an option here.

What approaches have you tried and is there an App Note on the startup for UCC256404?

Any help in the matter will be appreciated.

over 3 years ago

0 Manikanta Pallantla over 3 years ago

TI__Genius 16155 points

Hi Robin,

Thanks for reaching out to us.

UCC256404 uses HV pin for both high voltage startup and x-cap discharge. The controller needs to see zero crossing at HV pin to see AC input voltage has been disconnected. If it doesn't see any zero crossing at the HV pin, it injects x-cap discharge current.

So, though you can do the startup using DC voltage, it keeps injecting x-cap discharge current (since zero crossing is missing) which would increase the temperature of the IC.

So the best way would be using UCC256402 or UCC256402A which has high voltage startup but no x-cap discharge feature. In this case you can directly start the controller using DC input voltage.

please see the device comparison table below:

Please let me know if you have any further questions.

Regards

Manikanta P

0 robin gangopadhya18 over 3 years ago in reply to Manikanta Pallantla

Mastermind 6296 points

Manikanta:

Thanks a lot for drawing our attention to this fact- is not at all obvious from d/s review.

Really really appreciate that.

Is there a way to stop that injected current from flowing?

Pl bear in mind the situation is abominable! TI parts have been hijacked and hoarded: you cannot simply buy what you want.

I found hoarded parts in China *04 version, I just bought them for our immediate need. I cannot choose 02.

So right now: our design will have AC- it is not a problem.

Next yr, will change over to 02 IFF customers must feed 600V DC only.

BTW: where exactly does the controller sense line ZC? I do not see it unless, rectified ripple is sampled, freq determined & ZC estimated!

Can you send me your email so that I can send you the finished schematic we will be using to make our own after studying the Evalkit we are expecting next week?

I cannot post it in the forum. Hope you understand.

0 Manikanta Pallantla over 3 years ago in reply to robin gangopadhya18

TI__Genius 16155 points

Hi Robin,

We can not stop injecting the x-cap discharge current in 404 version.

The zero crossing is detected at the HV pin by looking at the rectified voltage which is shown below. This same circuit is also used for IC startup.

I can definitely review your schematic. You can send me those files through personal message through e2e.

Regards

Manikanta P

0 robin gangopadhya18 over 3 years ago in reply to Manikanta Pallantla

Mastermind 6296 points

thnx much! Will do so asap

0 robin gangopadhya18 over 3 years ago in reply to Manikanta Pallantla

Mastermind 6296 points

Manikanta

Problem!

I tried to massage the Lr. Cr, Lm values in the UCC256404 design calculation Spreadsheet but it seems there is no convenient practical values one can obtain in this topology with high turns ratio & low power.

I see Lr in 500uH range, Cr approx 1nF, Lm 3000uH....Then Vcr may be 3000V!

These values will not make a consumer /industrial grade module at a convenient price range.

So I got to take another look at UCC28740.

But that is another forum , another question.

I will post if I find a way to proceed with LLC given that I bought out the Eval board, 100 PC of the controller etc!\LOL

We do have project to make LLC modules...not right now.

Do you have some other controller? prefer one without integrating output "dc inductor".

Let us see where it goes from here.

0 Manikanta Pallantla over 3 years ago in reply to robin gangopadhya18

TI__Genius 16155 points

Hi Robin,

May I ask what's the power rating of the LLC you are designing at? In general half bridge LLC is recommended up to 600W.

UCC25640x is our latest generation half bridge LLC controller which is based on the charge control. We do have UCC25600 which is a voltage mode controller.

Regards

Manikanta P

0 robin gangopadhya18 over 3 years ago in reply to Manikanta Pallantla

Mastermind 6296 points

Hello Manikanta:

Nominal output power is only 5.25v@ 30W, it can go up to 45W so we designed for 45W

0 Manikanta Pallantla over 3 years ago in reply to robin gangopadhya18

TI__Genius 16155 points

For such a low power, I would recommend using flyback topolgy.

You can share the design calculator if you still want to use the LLC controller in this application.

Regards

Manikanta P

0 robin gangopadhya18 over 3 years ago in reply to Manikanta Pallantla

Mastermind 6296 points

Hello Manikanta:

Yes, we tried: each has a challenge.

UCC256404 requires huge inductance values and Vcr is also very high. See attached LLC worksheet- errors I do not have access to fix.

On the other hand, 28740 flyback will generate Vfly too high to be handled even if we have 1200V devices.

UCC28782 seems almost a good fit : excepting ratings of the Depletion mode startup FET.

Although that can be more easily addressed by putting 2 devices in series.

Yet, I am convinced THERE must be an optimum fit with either. More likely with 28782. Attached its xl calc.

Any help in the matter will be highly appreciated.

r4670.UCC25640x Design Calculator Rev4.0.xlsx

UCC28782_20v_600V_48_3_3.xlsx

0 Manikanta Pallantla over 3 years ago in reply to robin gangopadhya18

TI__Genius 16155 points

Hi Robin,

Could you let me know input voltage range spec for your application?

FYI, UCC25640x HS pin max voltage is 600V. You can not exceed that voltage.

Regards

Manikanta P

0 robin gangopadhya18 over 3 years ago in reply to Manikanta Pallantla

Mastermind 6296 points

bam!

You found a serious error here!

Yes, LLC nodes will likely be >600V although VDCin is =600v,

BUT THERE IS A WAY OUT.

And this being highly competitive- to show how small this module can be done- cost is not strictly an issue right now.

We can insert an Isolated gate driver. Say, UCC2154.

There is this little issue that kicks in now: output of UCC25640X IS 12V, ucc2154 max input is 5v....we know that ucc2751x series drivers can take 12v input while connected to 5v supply. That is a way to "level shift" 12v to 5 v without power loss- not too much.

All these can be addressed. If only Lr, Cr, Lm can be made practical consistent with the overall pkg of the intended module.

Any means to accomplish that will be highly appreciated.

And this will be our most favorite controller to use. LOL!

let me know

0 Manikanta Pallantla over 3 years ago in reply to robin gangopadhya18

TI__Genius 16155 points

Hi Robin,

We can not have half bridge gate driver to this controller. This is because HS pin needs to be connected to the switch node for adaptive dead time generation.

We can only have two single channel gate drivers if you want to connect external gate drivers. So, HS pin still sees more than 600V in your application which is not acceptable.

Regards

Manikanta P