Solution Finder Get Tech Support

LCS708HG

Posted by: gobooth on

I have a LLC supply using a LCS705HG that ios working.

To reduce the temperature of the LCS705HG I changed it to a 708.

I did not change any other component values.

With the 708 circuit has no output I put the 705 back in and I get an output. 

I checked PI Expert and switching for the 705 to the 708 doesn't change and resistor or capacitor values.

Why is the 708 not outputting? 

I am including the LLC current scan.

The load is a pwm 60 HZ inverter.

The peaks of +and - 7A seems well within  the 15.6 instantaneous repetitive rating.

Is that correct? 

Files

Attachment Size
llc.pdf 36.69 KB
LLC current.jpg 213.79 KB

Comments

Submitted by PI-NANO on 03/10/2014

Hi,

 

Ideally it should work.

 

Could you please check another part and another board.

 

Could you please send your PIxls file so that I can review the design?

 

Regards,

PI-NANO 

Submitted by gobooth on 03/11/2014

I didi check another part 708 and it did the samne thing. Then I put a 705 back in and it worked. I onlyy have one board as it is an R&D board. I have been using it a lot and the 705 has had no robelems. I just wanted a cooler part.

The llc secondary current on the pixls is 1.6 A  which is what my scope measures my current probe rms at full load.

Submitted by PI-Tucker on 03/13/2014

Pls. place a differential probe between VCCH and HB. Reduce the bandwidth of the scope channel below 200 kHz if possible. Check if the voltage drops below the hi-side driver UVLO(-).

 

Also place a probe between VCC and GND and check that too.

 

 

Submitted by gobooth on 03/20/2014

I placed a meter across VCCH -HB and it read 10.55 V which is just at the edge of UVLO.

VCC measures 12.6 V.

How do I fix the UVLO with the 708? 

Submitted by gobooth on 03/20/2014

I see VCCH has a lower start threshold. I don't have a differential probe but when I use the math function and subtact i see ripple that could go below VCCHB-UVLO.

File of scope. The red difference of VCCH-HB is noisy because my scope's math function is not very good.

File of scope. The red difference of VCCH-HB is noisy because my scope's math function is not very good.

Submitted by gobooth on 03/20/2014

I raised VCCH to HB to 11.89 and no improvement.

Submitted by PI-Tucker on 03/20/2014

It is hard to make a measurement without a good differential probe. Using 2 regular probes means you are looking at a small difference between 2 large voltages, which is error prone. You can test your setup by hooking up both probes to look at the voltage from the S pin to the HB pin. When you do a subtraction, the voltage should be zero.

 

You increase VDDH by increasing the voltage that supplies VCC. Increase it by a volt, see if there's a change.

 

 

Submitted by gobooth on 03/20/2014

I used a meter to measure VCCH-HB besides the scope traces . So the numbers I quoted for VCCH-HB are from a islated differential reading from a meter.

I did raise VCC as I described with no improvement.

Pls. retake the same scope photo as teh first attachment, but add the following voltages:

 

HB pin to Source

IS pin to GND 

 

Submitted by gobooth on 03/24/2014

Attached

Submitted by gobooth on 03/24/2014

Added a scan of :

Yellow Primary current 

Blue HB-Gnd

Magenta VCCH-GND

Red VCCH-HB 

I am looking again at the very first scope shot you posted, "LLC current.jpg".
It has a repeating pulse train of that lasts about 5 ms, and repeats every 8  ms.

Pls. could you tell me the frequency of the first several and the last several cycles of the pulse train? It may be easier to measure frequency by looking at the HB pin voltage (instead of current)

It would be even better if you could zoom into the waveform, of the first several and last several cycles, of the HB pin voltage.

Submitted by gobooth on 03/27/2014

I can do that.

But that is a working 705 unit.

The 8 ms rate is because i anm doing an inverter to 60 hz.

So the llc circuit is outputting for 5 ms. during the 60 hz peaks then the output caps supply the current to the PWM inverter around zero crossing.

 

OK can you then instead take the HB pin voltage, and the primary current, of the 708, but take the very first startup attempt. I need to see the startup until the point where it stops switching, and if it repeats. 

Submitted by gobooth on 03/27/2014

Attached

Submitted by PI-Tucker on 03/27/2014

Can you zoom out to the point where the switching stops?

Submitted by gobooth on 03/28/2014

The switching never stops. The  HB voltage goes down to 2v but you can still see switching.

The switching starts at about 15 V hb at about 400 KHZ then drops to about 2V and frequency decreases to about 250 kHZ over about 1 ms.

Then the HB increases to about 40 V at 100 KHZ and stays at that level.  

Submitted by gobooth on 03/28/2014

The switching never stops. The  HB voltage goes down to 2v but you can still see switching.

The switching starts at about 15 V hb at about 400 KHZ then drops to about 2V and frequency decreases to about 250 kHZ over about 1 ms.

Then the HB increases to about 40 V at 100 KHZ and stays at that level.  

Submitted by PI-Tucker on 03/28/2014

It looks like the high side driver is not turning on.
Pls. try using a 12V battery (u can use 9V battery + 2 x 1.5V), to power the high side driver. Connect it to HB and VCCHB pins. Use a twisted wire pair, and keep those wires, and the battery, away from the low-level signal traces and components. Keep them also away from any ground planes (an inch is enough). The battery will provide the isolation needed. You can disconnect D100 (the VCCH bootstrap diode) during this test.

Submitted by gobooth on 03/28/2014

I will try that. I replaced the 708 with a different new  708 part.

With no  load it came up OK.

But whan I loaded it like I do the 705 it failed similar to the one I have sent you scans of.

Is it possible the inverter load  of 7.5 A peaks at 100 KHZ for 5 ms every 8 msec is damaging the 708?

Or the turn on transient as it seems to fail right at turn on. 

Maybe the 705 is more forgiving than th 708? 

Maybe my design needs a more power capable LLC driver?

I think the 708 the largest though. 

 

Submitted by gobooth on 03/28/2014

I put a 705 back in to make the unit work.

Here are two scans of  llc turn on and load step.

The transients are pretty high upwards of 50 A for useconds.

Is there a way and can reduce these? 

Submitted by gobooth on 03/31/2014

In reply to by maleManGuy

Another wave form. The out V is rising exponentially, can I do something tso the  rise is a cap charging votage shape?

Pls. try increasing C245 (soft start capaitor) from 0.22uF to 1 uF. This will slow down the output voltage rate of rise at startup, and thus the peak startup currents.  The peak startup current shouldn't be more than about 1.5-2x the normal full load currents.

 

How much load capacitance do you have? 

 

If the IS pin protection circuitry is working right, it should be pretty hard to damage a HiperLCS device.