Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Cuk Converter

scopeprobe

Full Member level 2
Joined
Sep 17, 2020
Messages
147
Helped
27
Reputation
58
Reaction score
85
Trophy points
28
Location
Uk
Activity points
1,639
Hi All,
I'm currently evaluating some concepts for a new project which requires an offline PSU with PFC and a THD <5% to generate an isolated 40V DC @ 100W. I'm trying to avoid using a boost converter due to some challenging inrush requirements so a cuk topoloty looks reasonable solution. I've been researching the cuk and am able to sucessfully simulate a non isolated topology but i need some guidance for adding the isolation transformer to get isolation. Could anyone point me to a design guide / book where i could get some clarity or equations for getting this detail, i'm searching the web but can't seem to put my finger on the detail, i'm guessing there some resonant relationship between C2 & C3 and the transformer value.

Thanks in advance

cuk.png
 
Last edited:
We have done plenty of these - and control ( which is tricky ) - what do you want to know exactly ?

also the topology should be changed slightly for improved layout & performance

EP
 
We have done plenty of these - and control ( which is tricky ) - what do you want to know exactly ?

also the topology should be changed slightly for improved layout & performance

EP
Hi Easy Peasy,
I have managed to find enough information online via google and youtube to test a non isolated converter but aside from the odd outline diagram like in my original post theres very little information on designing an isolated version at all. Specifically i'm trying to determine how best to get the isolation transformer inductance and turns ratio optimised.

Regards
--- Updated ---

Thank you for the links, it touches on some of the information i'm looking for but not quiet to the depth i need for this topology
 
Hi Easy Peasy,
I have managed to find enough information online via google and youtube to test a non isolated converter but aside from the odd outline diagram like in my original post theres very little information on designing an isolated version at all. Specifically i'm trying to determine how best to get the isolation transformer inductance and turns ratio optimised.

Regards
--- Updated ---


Thank you for the links, it touches on some of the information i'm looking for but not quiet to the depth i need for this topology
Did that appraisal include his books? and references 2, 3 on SSA for Cuk in appendix
 
Last edited:

    KlausST

    Points: 2
    Helpful Answer Positive Rating
Did that appraisal include his books?
So far yes, he covers it only briefly as a topology but doesn't provide formulas or calculation accompanying it. It could well be its to specialised an area. I have most of his books and the ones that cover it seem to do so sparingly
 
I once spoke with Dr. ´Cuk in Cali on a proposal for a Lexmark laser printer 2nd source with quad PWM controlled HV outputs with low power for $25 in high volume. I declined to accept his proposal for EMI and potential cost reasons. In the end it was a sine resonant stepup converter with feedback. No LC switching at all except rectification. Lexmark wanted 1k units /day.
 
Last edited:
I once spoke with Dr. ´Cuk in Cali on a proposal for a Lexmark laser printer 2nd source with quad PWM controlled HV outputs with low power for $25 in high volume. I declined to accept his proposal for EMI and potential cost reasons. In the end it was a sine resonant stepup converter with feedback. No LC switching at all except rectification. Lexmark wanted 1k units /day.

Thats pretty cool, its sometimes forgotten these were just (clever) people at the end of the day. I just wish he put some information online for the not so blessed :)
 
Last edited:
Thank you, unfortunately the link doesn't work, but the additional information is useful
--- Updated ---

Kudos to you sir, I was planning on the UC2854 which is a derivative of the UC3854 so i'm pleased to be on the right lines with the control IC. I re-rerun my hypothetical calculations for for the Non Isolated CUK at 70Khz operation and the operation seems sensible and in some cases not a million miles from the circuit you provided. So in the context of what you provided could you tell me how you got to the values you calculated for L3 and L4 please as i'd really like to get a spreadsheet together to calculate this myself if possible?

Meanwhile i'm going to run some simulation on your circuit. Thank you

cuk.png
 
Last edited:
when not found, use archive .org


circa 1919

1712927768934.png


Beware of those % tolerance errors on zeros with f and LC found by the non-linear negative sensitivity in my previous link on sensitivity.

and as Basso wrote, if you change the load, you must redo the state-space equations.
 
when not found, use archive .org


circa 1919

View attachment 190034

Beware of those % tolerance errors on zeros with f and LC found by the non-linear negative sensitivity in my previous link on sensitivity.

and as Basso wrote, if you change the load, you must redo the state-space equations.
Thank you, I've never seen the wayback machine before - what a resouce that seems to be.

I'm going to need to spend time digesting this. Thankfully a nice weekend for some Garden reading.

Many thanks, i'll be back hopefully better informed
 
Sorry if this isnt too great, but i did some sims on isolated cuk PFC some time back, and this was a sim from it...(LTspice).
Dont mean to be negative but if inrush is problem then go buckboost PFC, or flyback PFC.
Cuk isnt a resonant topolgy, but their is a "resonant cuk".

FWIW if i were looking to it, i would look at the way that a two transistor forward is a transformer isolated Buck...and then apply kind of same thing to get the cuk into its isolated version. (but it looks like you already are doing this from top post) If you set the values of the Ls and Cs right then you can avoid the "resonant operation" situation, and just do the plain old hard switching isolated cuk......since the resonant version may take much more dev time.
 

Attachments

  • Cuk PFC _1.zip
    1.2 KB · Views: 31
Last edited:
That is the drawback of the Cuk converter ( standard type with or without Tx isolation ). Unless you know some elegant alterations to the topology, the turn off spikes on the main mosfet ( and indeed the output diode ) can be quite deleterious in the real world ( IRL ).

As the OP is seeking only 100W - these can be cured with clever snubbing techniques and adjustment of gate drive.

Decades ago - my firm had to provide a mains input, pulsed output PSU, 180 - 256Vac input - with a switchable output: -

  • 150V 3Apk
  • 220V 2Apk
  • 310V 1.5Apk
  • 420V 1Apk
And used a Cuk converter for the purpose - this was something of a learning exercise - as the dynamics of the Cuk are quite poor - needing real understandings around the behaviour of the power stage - and how to control it in a fashion that damps the resonances inherent in the power hardware

For the Cuk: - Vpk on the mosfet = Vin + Vout & Ipk = Ipk in + Ipk out ( same for diode )

So controlling the power stage properly makes quite a difference to the peak stresses seen in the semi's

( you can see for us the peak V on the main mosfet was peak mains ( 325V) + reflected vout ( 350V ) = 675V, allowing an 800V mosfet to be used - these psu's are still running in the field due to protection ckts, quality soldering and solid control of the Cuk ckt ).

--- Updated ---

Just further - for 100W out @ 40VDC. with UPF formed by the Cuk converter, there is a formula for the ripple on Vout for perfect power factor correction ( PFC )

Vout-pk = Power / ( 2. pi. Fr. Co. Vo ) where Fr = 100Hz for 50Hz mains

so for 40 Vout, and say 1Vpk ( 707mV rms ) of 100Hz ripple on Cout, Cout must be 3,980 uF ( and half this for 2Vpk ripple ( 1.414 V rms ) )

It is worth noting that for 120Vac input ( 170V pk ) the power processed at the peak of the sine is 2x the ave power required

Thus at the 170V peak, the current in the mosfet is 1.17A, + the reflected output current, (2.5 A ) + any slope current on the chokes = about 4.5A min for the mosfet

and about 8A pk for the output diode.
 
Last edited:

LaTeX Commands Quick-Menu:

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top