SMPS with no loop compensation

[eem2am]

voltage controlled smps

Hello,

the HV9910B LED driver does offline LED driving SMPS, but has no compensation pin whatsoever

Datasheet HV9910B
**broken link removed**

all of the HV9910B application notes have no loop compensation in them...

**broken link removed**

Do you agree with me that this is outrageous?

Surely you cannot possibly do a SMPS without loop compensation?

Is this not riding roughshod over all the entire theory of SMPS?

 

[FvM]

variable output current smps control

Do you agree with me that this is outrageous?

Not at all. You may want to learn about operation principles of current mode controllers. After reading some of your posts discussing various application circuits: You should rather doubt your level of knowledge than considering all engineers to be fools.

 

[ynhe]

smps loop compensation

No
the HV9910B operates in open loop current mode control.There is no closed voltage loop,no voltage feedback,So no voltage loop compensation need.
But if HV9910B is in Peak current mode control,the inner slope-cpmpensation must add when duty>50% in CCM.
PFM SMPS need no loop compensation,such as hysteresis control, constant on time control,constant off time control.

 

[eem2am]

smps slope compensation

but my friends i beg you to consider with me.....

Please see page 6 of this datasheet (LT3474 Buck LED driver) under "Vc"

https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1003,C1094,C1768,P13452,D9377

It says

"The Vc pin is the output of the internal error
amp. The voltage on this pin controls the peak switch current. Use this pin to compensate the control loop."

.....So it is seen that these type of current sensing buck converters do in fact need loop compensation.

.....Now, HV9910B can be operated either in constant frequency mode, or constant off-time mode.....

....I believe only the constant off-time mode is the also named "hystertic" control mode.

The fixed frequency mode, i am certain needs loop compensation. You only need to pull the datasheets for any Buck LED driver you like....you will see the need for compensation of the control loop.

But the HV9910B datasheet just states that HV9910B needs no loop compensation, without stating that this only applies to the constant off-time mode.

 

[FvM]

loop compensation

LT3474 has a true current control loop, HV9910B hasn't. You can see the difference from the block diagram at first look. As a consequence, HV9910B current depends more or less on the input and output voltage.

 

[eem2am]

supertex hv9910b application

Ok Thankyou,

This LED driver is a peak current detector just like the above one that does not need loop compensation (HV9910B).

Would it be correct to say that this converter would be stable, even though it simply uses a comparator (LT1719) to command the power switch on and off....and has no loop compensation?

 

[ynhe]

control loop compensation

LED driver is essentially a constant current source.
In peak current mode controlled buck,If send a fixed error voltage to PWM comparator,we can easily get a constant averaged output current.
The averaged output current depends on the switching frequency,output voltage,Duty,output inductor,slope compensation and current sense gain.
As the whole system is in open loop status,no frequency compensation need.
I guess the HV9910B operating like this way.

 

[eem2am]

constant on-time buck

oh thankk you ynhe,

So the LT3474 (as in my post, fourth down) is not peak current mode control because the current sense comparator has a variable reference?
(if wish , please see figure 7, page 7 of LT3474 datasheet).....................

https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1003,C1094,C1768,P13452,D9377


since peak current mode is so much easier and simpler, requiring no compensation components, i cannot understand why anyone would do a buck LED driver any other way?

 

[FvM]

buck control compensation

The limitations of this operation mode have been clearly stated:

The averaged output current depends on the switching frequency,output voltage,Duty,output inductor,slope compensation
and current sense gain.

As a consequence, HV9910B current depends more or less on the input and output voltage.

They may be acceptable for applications that don't need exact current regulation.

 

[ynhe]

pfm smps

Dear eem2am
I think you do need study some basic knowledge about peak current mode controlled SMPS.I suggest two for you reading:
1.R. W. Erickson_fundamentals_of_power_electronics
2.R. B. Ridley, “A New, Continuous-Time Model for Current-Mode
Control,” IEEE Trans. on Power Electronics. Vol. 7 No. 2, April
1991. pp. 271-280.
Good Luck!
Bye Bye

ynhe

oh thankk you ynhe,

So the LT3474 (as in my post, fourth down) is not peak current mode control because the current sense comparator has a variable reference?
(if wish , please see figure 7, page 7 of LT3474 datasheet).....................

https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1003,C1094,C1768,P13452,D9377


since peak current mode is so much easier and simpler, requiring no compensation components, i cannot understand why anyone would do a buck LED driver any other way?

 

[eem2am]

Thankyou so much and the thing is , i tell it now,

the circuit above pictured gives pretty brilliant current control and even with big change in V(in) the current level is wonderfully constant....
.in fact, it works so well that i wonder if something is wrong.

maybe it is operating on the verge of a bad phase bump in the bode plot which is going to come out and smash me up when the ESR of the caps gets worse with age etc etc?

i just think that the comparator in the secondary side would seem better if it was replaced with a compensated error amplifier?