Solution Finder Get Tech Support

lnk6777e pad

Posted by: bougnoul on

Terry:
I cannot get back to our long discussions on this forum. They have changed the website. It is much less user friendly than before including mangled screens where lines/texts overlap.
In an event: I am using LNK6777E.
I have mentioned before, I desinged using values close to the app note. I have tested my 4-layer assembly first by putting an external MOSFET & driving it. No controller yet. With this, I checked voltage at the BP pad, voltage at the feedback pad & polarity etc.
I then checked drain waveform. Excellent. All appeared to be consistent with design.
So then I decided it was safe to install the controller.
One thing, since it is hand assembled, I have not soldered the exposed pad yet.
I powered up: BP node now is 5.75V which is correct.
But no switching.

I suspect I have to solder the exposed pin to the return path.
Before I remove the controller, I need to know if not connecting the exposed pad is the reason or something else is preventing it from switching.
-robin

Comments

Submitted by PI-Terry on 12/11/2014

Hi, Robin:

I am sorry to hear that you have difficulty to get back to the previous post because of the new website. We are just in the beginning of the new website, and I am send this feedback to our IT team for continuous improvement. Thanks for that.
Looks like you using the E package of LNK HP from use, so the exposed pad you mentioned is the pad on the back side of the package, is it correct? This pad is actually connected t the source pin internally, it is not required to connect it to return source. But you can connected to the source if you want to.
For the no switching behavior, the reason could be the line UV fault that disable the switching. The current from the FB pin when the primary side MOSFET turned on is detected to implement the UV and OV protection. You can refer to the data sheet for more detail information.

Hope this is helpful for you.

Submitted by bougnoul on 04/14/2015

Terry:
After several months of being away, I am back again trying to make this 10W flyback work.
I have checked so far the following
-pcb : correct as per the schematic in the d/s
- component values: correct as per the PIDesigner Table. But no input current limiter resistor as xmfr DCR is 5 Ohms.
- transformer: tested by installing an external; MOSFET without the chip. Drove it from a source. Checked polarity of secondary rectified output, bias winding outputs both before the diode & bias node.
All consistent with what one would expect. Turns ratio, leakage all correct. I loaded secondary fully, no primary spike at all. Albeit, the source is 50 Ohm PWM, one would expect clean switching.
I put back the controller. Slowly ramped up DC, can see bias at correct value all the way to 230V
Then bam, turned on DC all the way to 230V
Blast.
Chip fried.
I am adding back 10 ohm input current limit.
One thing bothers me: the DC source takes 140ms to ramp up to 230V.
This is too slow but would it blow the controller like a hole in it?
Appreciate any critical comment to get this running asap ....
thnx

Submitted by bougnoul on 04/14/2015

Terry

As I am trying to prevent another blast of the new controller, I read with interest the data sheet. How is it possible to get FB pin to source current when there is no switching? Is it not true that first as DC level at drain goes above UVLO, there has to be several switching cycles. Due to these switching pulses on the drain, secondary windings will get both + & - voltage levels. Clearly then FB pin can source current into the bias winding when MOSFET is ON.

All this is not happening in my circuit.

I did not have a clamp initially because my lkg did not seem to need. But I do not know if  during soft-start, it is possible to get HV on drain  which might have blasted my chip.

I added on the clamp.

what else?

thnx

Submitted by bougnoul on 04/14/2015

In reply to by PI-Terry

After blasting out a chip, got smarter. I added the clamps + input current limiter.

Now I see very clean switching but of extremely low frequency. It also does so at some voltages, shuts off at all others.

Definitley, there is hysteresis & long delay...I have to wait a while for it to give me a pulse or 2..

I find there is a mix of of Rfb2 with value at PD pin.

Will fix that tomorrow.

But certainly, UVO or Over current ...is sensed & it shuts down.

Let me write tomorrow again.

MAybe I will be lucky. It is a clean 4-layer pcb so it must work nice.

robin

 

Submitted by bougnoul on 04/15/2015

Terry

I had it almost working when I blew the chip & the clamp R

This forced me to look into the PI Design suite again.

I noticed that I had been ignoring that the Design was selecting LNK6774V while the part I installed is LNK6777V

A big difference in transformer!!!!!!!!!!!!!

Also, Lpri is vastly different. Which means I did not have enough gap for LNK6777V

That may explain all the troubles I am having.

6774 comes up with very different turns than what I have. I am attaching the file for your reference. My transformer is 108:16:15  with Lpri as 1910uH. Mine was 5880uh!!!!!!!!!! all this time.

I had to add significant gap in the RM8ILP core to get 1.91mH

In this config now, I have exactly the values of turns, ratio & RF1 & RF2 as in the spread sheet here.

I will post again if this works

robin

 

Attachment Size
LinkSwitch_HP_16V5_6777v.pixls 50 KB
Submitted by PI-KSAN on 04/16/2015

Thanks for Posting your reply into and we appreciate that you find the root cause for the IC blew up.  Our PI Expert designs the transformer based on the ILIM Selection of the Parts.  We have to be very careful  about the warnings, generally PI Expert issues warning when the part and transformer doesn't match. 5.89mH for LNK6777V is too high for 10W Output power , as Inductance is too High , the leakage inductance is also too high and LNK6777V part current limit is higher than LNK6774V part, the leakage nenergy is too High with the Combination of LNK6777V Part and 5.89mH, Compared to 1.9mH almost three times. So with Higher inductance the Peak Drain to Source voltage might have over the BVdss of our part and might caused the Blew up.

We are glad that you find out and make sure whenever your designing you are using the right part and inductance on the prototype board.

 

Regards,

PI-KSAN

Submitted by bougnoul on 04/16/2015

KSAN

thnx for the alert.

Yes, at first I was not paying attention to the auto selection of LNK6774 which gets designed using PI_Designer very differently.

Once I selected LNK6777V and selected proper input DC min/max, output voltage & power, I got values that match what I built.

I emailed you the pixl in the last email- you can see inductance is 1910 uH. I got now 2014 uH which is very close.

Secondary is 46uH again very close to what PI_designer gives you.

Windings are 108T:15T:16T, lkg<58uH

At first I tested the assembly with an external MOSFET & a gate drive. Everything looked as expected.

Then I installed a device WITHOUT THE exposed pad soldered to source pins.

With RFB(top)= 118k, RFB(bottom) =16.2k, Rpd=10k no caps.

I got it working  for a bit. Quickly measured outputs & bias. They were both approx 36V but should be only 16V

But it would shut down..& I would turn off because of high outputs.

This morning, I changed Rp to 121k (no caps). rolled up DC, heard a pop. The input 10 Ohm resistors popped. This is not even needed in the LNK6777 circuit.

I shorted that out. Tried to roll up DC, there was no switching. But around 170V , the device blew up.

Why so without any switching?

I am definitely doing something wrong here.

I am email all files to Mike Jones. Let us see if we can make it work quickly.

thnx

KSAN

thnx for the alert.

Yes, at first I was not paying attention to the auto selection of LNK6774 which gets designed using PI_Designer very differently.

Once I selected LNK6777V and selected proper input DC min/max, output voltage & power, I got values that match what I built.

I emaled you the pixl in the last email- you can see inductance is 1910 uH. I got now 2014 uH which is very close.

Secondary is 46uH again very close to what PI_designer gives you.

Windings are 108T:15T:16T, lkg<58uH

At first I tested the assembly with an external MOSFET & a gate drive. Everything looked as expected.

Then I installed a device WITHOUT THE exposed pad soldered to source pins.

With RFB(top)= 118k, RFB(bottom) =16.2k, Rpd=10k no caps.

I got it working  for a bit. Quickly measured outputs & bias. They were both approx 36V but should be only 16V

But it would shut down..& I would turn off because of high outputs.

This morning, I changed Rp to 121k (no caps). rolled up DC, heard a pop. The input 10Ohm resistors popped.

I shorted that out. Tried to roll up DC, there was no switching. But around 170V , the device blew up.

Why so without any switching?

I am definitely doing something wrong here.

I am email alll files to Mike Jones. Let us see if we can make it wokr quickly.

thnx

roibn

 

Attachment Size
LinkSwitch_HP_16V5_6777v.pixls 50 KB
Submitted by PI-KSAN on 04/17/2015

Could you please send us the schematic. And also please let me know what are your exact specifications?

Regards,

PI-KSAN

Submitted by bougnoul on 04/20/2015

KSAN I can only email you encrypted files. I did email them to Mike Jones Thursday. These are self decrypting files. Meaning, once you svae the file in a folder, double clicking will make it ask for a passphrase. I have also emailed the instructions to Mike. Can you find out if he can forward these to you?

Submitted by bougnoul on 04/21/2015

here is a short form spec

input: 165 to 330V

Input has 50uF very close to this module

output: 16.5V @ 500 mA . In order to reduce ripple significantly, I have an LT3080 at the output with very little voltage drop.

that is it.

Submitted by PI-KSAN on 05/01/2015

Thanks. I will speak with Mike Jones to forward all of your related material to me. What are the Issues that you are encountering now? Once i get the material from Mike, I will go through it.

Regards,

KSAN

 

KSAN

 

Oh, I have not been checking the support page for a while...

 

Mike has been good in communicating on this matter.

 

Pl check what I emailed yesterday(May6)

 

My main concern is that from what you are pointing out--that auxiliary diode current conduction time has to be more than 1.2usec is not possible in this load-variable application.

 

I have made Raytheon invest significant time($$) in chasing this design. So I hate to tell them I may drop this circuit.

 

But yet, something tells me this is absurd!

 

We are misunderstanding something here.

 

I have a complete LTSPICE model of of LT1070(in cascode mode) with my style of winding & it works fine.

 

Only  thing is LT does not have isolated winding like PI & besides, I already have PCBs built....

 

so help me find a solution.

 

thnx

 

Submitted by PI-KSAN on 05/21/2015

Hi ,

So are you using LTPart as an LDO or to design a whole Power Supply.  I think mike Jones is working with some one in our Office for your query. I will talk to Mike Jones again and will find out what exactly is happening. We will definitely help you out finding a perfect solution for you .

 

Regards,

PI-KSAN