Can Half bridge driver be used with 385VDC bus?

[grizedale]

Hi,

I wish to use LM5039 to do a 330W half-bridge SMPS running off 385VDC.

**broken link removed**

..LM5039 has a high side driver pin that can handle only up to 105V.

-if i want to drive a high side fet with an external high/low side driver rated for the 385VDC, then i presume i can just use LM5039's hi-side output by grounding the HS pin and driving like that.?

I am surprised that they put an internal high-side driver in it thats only rated to 105V.....nobody's mains voltage is that low.

 

[atripathi]

I think it is targeted for 48VDC bus based power supplies. I think for telecom application 48V DC bus is common input source.

 

[grizedale]

thanks Atripathi, I am wondering if it is possible though to use LM5039 with the mains, because it has in-built average current control method, which is superb with half-bridges...the half-bridge is not a realistic choice of topology otherwise due to the overload situation where the capacitor divider voltage drifts to a rail.

 

[FvM]

I am wondering if it is possible though to use LM5039 with the mains

By supplementing an external bootstrap HS driver.

 

[mtwieg]

Sure, you should be able to use the high side driver as a low side output by grounding HS and connecting HB to your supply rail, then using it to control the high side of an external bootstrap driver.

I remember orson cart mentioning a few times the advantages of ACM control for half bridge supplies, but I never got an explanation why... ACM control isn't even discussed in most of my reference books (strangely). And the more I read that datasheet the more dubious it sounds. First of all, it's not even real average current mode control. I don't even see where any current averaging or filtering takes place. Also they make this statement:

A capacitor connected between the ACL pin and GND operates
as an integrator in the average current limit circuitry. The ACL
capacitor is charged during current limit condition. As the ACL pin
voltage rises, it terminates the cycle through the PWM
comparator by pulling down the input of the comparator that is
normally controlled through the COMP pin. This maintains equal
pulse-widths in both the phases of the half-bridge and thereby
maintains balance of the half-bridge capacitor voltages.

This is pretty dubious, because even if the pulses from the controller are balanced, the effective volt-time product applied to the transformer can be unbalanced due to differences in the rectifier diodes and switching transistors. This doesn't sound like a very robust method.

 

[grizedale]

mtwieg.the LM5039 is a new chip....and i am wondering if this is a bug that you have found.

Also, national.com are on record as saying they "dont do offline controllers"........but if LM5039 does what it purports to do, it would literally be the best offline halfbridge controller on the planet.......since FvM also tells that it can be used for offline by using external gate driver.

It is puzzling me why they are not mentioning the offline capability of LM5039 and other PWM ic's of their stock of controllers........do they just not want to sell their stock in larger numbers?

 

[mtwieg]

mtwieg.the LM5039 is a new chip....and i am wondering if this is a bug that you have found.

Well I wouldn't call it a bug... I'm just not convinced it does what it claims to do.

Also, national.com are on record as saying they "dont do offline controllers"........but if LM5039 does what it purports to do, it would literally be the best offline halfbridge controller on the planet.......since FvM also tells that it can be used for offline by using external gate driver.

It is puzzling me why they are not mentioning the offline capability of LM5039 and other PWM ic's of their stock of controllers........do they just not want to sell their stock in larger numbers?

It also has a limited supply voltage of 105, so you couldn't use it offline without external supply biasing and startup circuitry. So it really isn't an offline controller (since handling line level voltages is pretty much what defines offline converters).

 

[grizedale]

Fig 7 of the LM5039 datasheet shows a startup circuit for when vin>105V.

....for the superb average current capability , the price of a starter circuit etc is small price to pay.....

i also just watched the youtube video on it and it seems genuine enough

LM5035C & LM5039 Controllers Enable Smarter, Higher Density Isolated Power Supplies - YouTube

UC3842 doesnt directly accept mains voltages, but its still an offline chip.

I am totally confused that National.com declare themselves as not providers of offline pwm controllers.....do they just NOT want to make money.

 

[mtwieg]

I looked at the schematic for their evaluation kit and I understand how they avoid drift during fault conditions. I forgot it's still fundamentally a voltage mode controller, so they don't give any DC path for the primary, so any imbalance applied to the transformer will just be countered by drift in the split rail voltage. The drift should be bounded to a small change (though this will result in some subharmonic output ripple, though I think that's unavoidable). So yeah, I see how this advanced fault protection works (though I wonder why they don't just use average current mode control all the time...).

 

[dick_freebird]

I am totally confused that National.com declare themselves as not providers of offline pwm controllers.....do they just NOT want to make money.

Probably more a question of how hard you have to work for the money you get.
Particularly if you don't already happen to have a 600V do-all process sitting
there in your pocket, and teams of people who love dealing with all the whining
about power factor and so on. And being front and center when some consumer
product starts blowing up in the field because somebody picked the wrong cap.
There's a reason why people like to buy the brick from somebody else with the
UL sticker already on.