LT1248 PFC Controller

Do you think the Voltage error amplifier in LT1248 can be used as a buffer by shorting pins 7 and 11?
It looks like there is a diode in the way?

LT1248
https://www.analog.com/media/en/technical-documentation/data-sheets/1248fd.pdf

...block diagram on page 1

you are forgetting the fixed 7v5 internal ref ....

Thanks yes, ....what you say is what i had been thinking....the thing is, the vref is the ss pin voltage until ss pin voltage equals 7v5.

Oh - OK - you then modulate the target Vout by pulling down on the SS pin - which can reduce the 7v5 all the way to zero

I think I actually did this once on the UC3854 - which is what this chip is a copy of - yes it can be done - but at 12uA there is potential for noise injection - and for way too much gain in controlling the pin

Thanks, though if we inject the voltage into the SS pin via an opamp buffer, then its low output Z should free us from noise issues we hope.

I think I actually did this once on the UC3854 - which is what this chip is a copy of -

Click to expand...

Thanks very much.
Its very interesting that UC3854 and LT1248 do indeed appear to be “twins”

The UC3854 looks to be less than half the price of the LT1248.

The differences look very minor...

A…though the UC3854 has no overvoltage pin…so that would need an external comparator…though in UC3854’s favour, it does have line voltage feed-forward via the VRMS pin, which the LT1248 doesn’t have.
B….The LT1248 is sync-able, but UC3854 doesn’t seem to be.
C….Bias current is very slightly higher in UC3854.
D……The UC3854 offers to reduce line current dead zone by connecting a resistor from VREF to IAC…..presumably this can also be done with the LT1248, though its datasheet makes no mention.
E……The LT1248 also appears to have biasing which prevents its Vout rising above the regulation level during light-load or no-load….UC3854 doesn’t appear to have this.
F……The LT1248 MOUT pin is clamped internally at -0.6V, whereas the UC3854 needs an external Schottky to be placed on its equivalent pin………I am not sure I like this, Schottkys have high leakage current at high temperature and this could mess up the MOUT pin (called “MULTOUT” in UC3854). The multiplier output current is only 200uA maximum so I don’t like the idea of having a Schottky on the MULTOUT pin.
G…..The LT1248 datasheet says it needs a Schottky from Gate drive pin to ground, whereas UC3854 datasheet makes no mention of this.
H…..strangely the UC3854 datasheet says it mustn’t be placed within 1 inch of the boost inductor, and musnt be placed under the boost inductor….but LT1248 makes no mention of this. It seems odd because eg the Innoswitch range of flyback controllers from power integrations actually state (on page 32 of its datasheet) that the innoswitch controller should be placed directly under the flyback transformer.

Due to A and F above, would you say the LT1248 would be the preferred choice?

LT1248 datasheet
https://www.analog.com/media/en/technical-documentation/data-sheets/1248fd.pdf

UC3854 datasheet
https://www.ti.com/lit/ds/symlink/uc3854.pdf

Inno switch datasheet
**broken link removed**

Thanks, though if we inject the voltage into the SS pin via an opamp buffer, then its low output Z should free us from noise issues we hope

Click to expand...

dream on, how much noise on 12uA do you think will affect things, you do not inject a voltage onto the ss pin, rather you pull it down, 12uA

the lt chip is later, hence some improvements, made under license ...

dream on, how much noise on 12uA do you think will affect things, you do not inject a voltage onto the ss pin, rather you pull it down, 12uA

Click to expand...

Thanks, I agree there is a 12uA current source in there…but the opamp output will act as the pull down resistor in effect…..the opamp buffer output is low impedance…and if we connect it to the SS pin, then its also directly connected to the non-inverting input of the Error amplifier….so I don’t see how we can suffer noise issues?

Depending on what size of cap you also put on the pin - you might reduce the potential for interference / noise

as standard ( with no cap ) 12uA will pull the pin to zero, no uA lets it rise to 7v5, 7v5 = 385VDC say, therefore +/-5V = +/-155nA

how low noise is your op-amp ...? 155nA on 10k = 1.55mV, thus even if you linearise the opamp o/p by using a 10k driving resistor, +/- 1.55mV on the op-amp o/p causes +/- 5V on the 385VDC - sorry - how low noise is your ckt ...?

Putting say 100nF on the pin will alleviate this to some extent - as long as you design an op-amp ckt to drive it - and assuming it won't cause delay issues in the overall control loop ...

Thanks, but this is just to drive LEDs on the output. The micro will simply regularly read the led current...and then increment or decrement the voltage on the V+ pin of the Voltage error amplifier accordingly.
The V+ input of the voltage error amplifier will be directly connected to the output of the op amp buffer.
In this way we wont suffer the noise issues.
In LT1248, the 12uA current source is off on a side connection, and therefore won't be relevant to us here.

So what we are trying to do is just be able to set the voltage on the VAOUT pin of LT1248.
The only other way we could do it is to allow the V+ input to stay at 7v5, and then set the error amp up as an inverting opamp, and then apply voltages to that in order to set the VAOUT voltage…but this is more components than just using a buffer into the softstart pin in conjunction with setting the error amp up as a buffer by connecting V- and VAOUT.

I must admit, the diode in line with the soft start pin is a cause of some concern, and makes me think that maybe its not OK to inject a voltage into this pin. However, if this is a problem, then I believe we could simply add our own diode, and put it cathode to buffer output, and effect a voltage onto the V+ pin of the voltage error amplifier like that. -And still avoid noise issues.
Page 11 of the UC3854 datasheet seems to suggest that these diodes aren’t actually real diodes.

You can also add your own current source - externally - say 1mA, and a larger SS cap and then have an opamp control a xtor to pull this down, or use an opto from the o/p side to pull it down such that the control is now not on the mains ...

Thanks yes, that wil be useful if we do it with an isolated topology...in this case we are just doing it with a Boost PFC.
So for us, it seems the easiest way is just to feed the SS pin with an opamp buffered voltage?

Which of the attached methods of manually setting the VAOUT pin voltage of LT1248 do you think is the best for the case here?
There are 5 ways shown.

"5" won't work, a bjt with emitter resistor for degeneration would be a better approach - remember 7v5 = 400V say

so 2 V ripple on o/p = only 37.5mV ripple on the SS pin - a cap on this pin is advised ...

Thanks, but in our (admittedly very unusual) use case we don’t mind. The micro will just repetitively read the LED current and simply repeatedly increment or decrement the reference voltage until the LED current is at the demanded level. I think circuit 1 of post #13 is fine for this?
We would prefer a cap on SS...but the extra circuitry needed to get it there doesnt seem worth it.

"5" won't work

Click to expand...

Thanks, but surely all thats doing is setting up the internal opamp as an inverting opamp?....and as described above, the particular use case makes this fine?

you are forgetting the internal 7v5 reference again ...

"5" is working with the internal 7V5 reference. The input and corresponding VAOUT voltages given in the schem of post #13 are in accordance with the 7V5 reference being there.

Hi
We wish to affect voltages on the VAOUT pin of LT1248 PFC controller. Will the circuit number 5 of the attachment in post #13 above be ok to do this?

(The Boost PFC will be operated open loop here, with a long led string on the output)

LT1248 datasheet

pulling the SS pin around is one of the few ways to control the Vout, remember this is a very noise sensitive pin.

the other way is to vary the V that is being measured on the Vsense pin, lifting it by some factor for a lesser Vout required, and attenuating it by some other factor for greater Vout required.

e.g. variable gain op-amp/stage driving the Vsense pin ...