Bias winding to power UCC28061?

hello treez,

I am thinking to use UCC28061 as flyback pfc too, but now I am not sure to start with this design because i heard about some trouble at startup time:  at startup the output capacitance appears as a virtual short circuit across the flyback transformer and there is little or no signal at t the ZCD sense inputs of the UCC28061 and it has difficult to start.

Did you have this problem ?

Did you find a solution for it ?

Please let me know if your design works fine.

Regards,

Marco

Marco,

http://focus.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=sluu316&fileType=pdf

Look at page 2, diode D1 provides a typical solution often seen in a PFC pre-regulator, to your concern. Output capacitors are charged before PFC pre-regulator is powered up and running.

Treez,

Vout = 36V and at 50W, the current is quite small, less than 2A. For this low current, it is typically no need to use interleave technique. Could you please let us know why your application needs the interleave technique for such a low current? 

My application (48V/33W) needs a pfc and an extremely low profile (max board height 12 mm), thus I am thinking to use a single stage pfc with two efd20 transformer.

regards,

marco

Hong,

on flyback topology is not possibloe to connect a diode beetwen input and output without losing galvanic isolation.
regards,
marco

Marco,

In case of flyback with single stage PFC, I believe to minimize surge current during start from initial charging the output capacitors, soft start should help.

Flyback with signal stage PFC is different from PFC in that initial power current surge is always there in PFC such that the diode is needed to avoid the PFC chokes satauration at start.

With flyback, such kind of start-up process does not exist. Soft start should solve the start up surge current issue in flyback with single stage PFC. 

Marco,

Can you use two inductors in series connection without using interleave technique while still meeting 12mm height? A power of 33W (48V out, then less than 1A) is a small power rating and usually should be ok without interleave.

Hong,
as far as I know,  the UCC28061 in flyback configuration has the opposite problem: it has difficult to start because it does not see the ZCD signal and it runs to its default low frequency with very low energy transfer to the output capacitor. In this situation  the starting phase could be really too long.
 I have no definite information but seems that T.I. has developed a prototype of interleaved flyback pfc with UCC28061 (see http://www.cadfamily.com/download/EDA/TI-Power-Supply-Design/101.pdf  pag 18-21) and it found this trouble.

regards,
marco

Hong,

I am not sure if using two series connected transformer, the ZCD is correctly recognized from UCC28061 due to unavoidable differences between the two transformers.
If there are differences on primary inductance they will have different discarge time and i don't know if in this condition the UCC28061 generates false or missing triggers.

regards,
marco

Marco,

In one of your messages, it looks your design specs are 48V and 33W but needs low profile 12mm. I thought the low profile was the reason that you want to use interleave technique from using two transformers.

My question is if your design can stay with non-interleave technique, i.e., single-phase PFC not two-phase interleaved PFC? If signle-phase PFC is acceptable, can the low profile be met with one efd20 transformer in your application?

I have been trying to understand the particular reasons why your application needs two-phase interleave technique with UCC28061 for only output power 33W at 48V.

On the UCC28061start, could you please provide some detail when you say , "starting phase could be really too long"? Could you please specify how long is too long from your perspective? How fast turn-on is required in your design?

Hong,
on my application one single efd20 transformer is not able to manage 33W, so 2 of them are neccesary (the efd25 transformer meets power requirements but does not height requirements).
They could be put in series (but I am not sure about ZCD signal) or I can use them in interleave mode with  the advantage of smaller emi filter.
I have no other reason for using the interleave topology.

Unfortunately I have just  a T.I. document fragment:
"...When the converter starts up the output capacitance appears as a virtual short circuit across the flyback transformer secondary windings. The lack of a flyback voltage results in little or no signal at
the ZCD sense inputs of the controller. This causes the controller to default to a low frequency of operation (~15 kHz), with a corresponding duty of 1% - 2%. The power transfer for this condition is
very low and the output capacitance charges slowly. Meanwhile the soft-start circuit allows the ontime period of the power switches to progressively increase.
When the output capacitance has charged to about 30% of the nominal regulation voltage sufficient signal magnitude exists to arm the controller’s ZCD circuits. At this point the controller begins operating in full transition mode, causing the switch duty to suddenly rise to say 30%. The resulting input surge current can only be controlled by limiting output current.
A very slow soft start helps prevent the on-time control voltage from rising too high before the output capacitance has become sufficiently charged.
..."

I think that 100 or 200 milliseconds of startup time is adequate

Marco,

May you please provide where the citation comes from?

At the end of your last message, 100 to 200ms startup time is mentioned. Do you mean such startup time is accepted to your application?

Hong,
at very beginning I spoke about UCC28061 in flyback mode with a T.I. engineer; he told me about a customer that built a prototype and he sent me some notes derived from that project pasted in a mail. The most relevant of wich I showed you.
I have nothing more that that.

Yes on my application 100 or 200 milliseconds are acceptable.