This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

UCC28251: Designing SMPS using ucc28251 PWM controller

Part Number: UCC28251
Other Parts Discussed in Thread: UCC28950, UCC28250, TINA-TI

Hi Team,

          We are developing 300V DC output switched mode power supply using UCC28251 PWM controller.I have connected 0.1uf capacitor in SS pin. While testing IC alone by giving 5V VDD,  we found that voltage across the SS pin is 3.5V and it is not varying if i increased the output. But i saw in your datasheet that it should be 0V intilally and varies with respect to output voltage.But in this case it shows constant 3.5V. Please help me out. Thankyou

  • Hi Subash


    VSS(max) in the datasheet is given at 3.6V - which is consistent with what you are seeing. Fig 21 shows a linear ramp of Vss from 0V to 3.6V which takes place over the initial soft start interval you have programmed with the 0.1uF capacitor. This votlage is not a function of Vout.

    The Soft Start process terminates when the voltage at the SS pin reaches that at the REF/EA pin - at this time the control loop obeys the REF/EA input and ignores the SS pin input which continues to rise to its final target of 3.6V.

    Regards
    Colin

  • Thanks Colin. Since my output voltage is 300V, soft start is not possible for my design, right? will it be a problem to operate the device without using soft start?
  • Hi Subash

    No - you can get soft start with any output voltage - What the soft start does is to control the initial rise of Vout - essentially Vout will be proportional to the voltage at the SS pin. As the output voltage approaches its target the SS pin will have reached and then exceeded the voltage at the REF/EA pin and when this happens control is 'handed over' from the SS ramp to the error amplifier.

    The speed of the SS ramp dictates the currents in the output of the PSU - you need to make sure that these currents are not any higher than those seen under current limit or max output current conditions so it's a trade off between having a fast turn-on time with high currents or a slower turn-on time with lower currents but the designer is able to make this tradeoff by selecting the correct SS capacitor.

    Regards
    Colin
  • Thanks Colin. Im Clear now and is there any software tool for UCC28251. I searched in your website but couldn't get it. Excel design file or software to design UCC28251.
  • Hi Subash

    I'm afraid we don't have a calculator for this device but there is one for the similar UCC28950 at www.ti.com/.../toolssoftware - double check the differences between these devices - they are given in the UCC28251 DS - but you should be able to get 95% of what you need from it.

    There is a 'typical Applications' section in the datasheet that goes through a typical design - although I'm afraid it doesn't give too much detail about the power stage design.

    you should also look at the application notes for this part and for the very similar UCC28250
    There are some reference designs for the UCC28251 that should be of use to you at www.ti.com/.../technicaldocuments

    Finally, I would reccomend that you run the software models - TINA-TI is a free simulator and the average models will allow you to look at control loop compensation.

    Regards
    Colin
  • Thanks for your help colin.
  • Hello colin, Could you provide us the design file to design Type III compensator? or provide us a sample design.
  • Hi Subash

    I don't have a calculator as such but you may find useful information in the articles at:

    http://www.ti.com/lit/an/slva662/slva662.pdf 

    https://www.ti.com/seclit/ml/slup113/slup113.pdf  

    https://www.ti.com/seclit/ml/slup173/slup173.pdf

    Regards
    Colin

  • The internal error amplifier in UCC28251 i.e 8,9,10 th pins. Here i have connected Vref to REF/EA+ and FB/EA- connected to output voltage divider and the output COMP is fedback to FB/EA-.Now my question is the COMP voltage is not varying with respect to FB/EA- voltage. it is constantly showing 3.24V. I checked it by manually giving DC voltage upto 100V to voltage divider circuit but VCOMP remains constant.Help us please.
  • Hi Subash

    Can you send me your circuit diagram please - post it here if you can - otherwise send it to me at colingillmor@ti.com

    To answer your question - COMP is 'locked high because the voltage at the EA+ pin is too high - the feedback signal at the EA+ pin should reduce as the OUTPUT voltage goes up. there is no direct direct relationship between the voltage at COMP and that at VSENSE. I'd suggest that you simply reduce the EA+ pin to less than the max possible COMP pin voltage (3.24V) and then the error ammplifier should behave as a voltage follower.

    Regards
    Colin
  • Hi Colin

    I couldn't send my circuit diagram here but i will send it to your mail ID. And as you said the COMP is locked because that the EA+ pin voltage is too high, Now i have connected voltage divider in between VREF and EA+ such that the voltage in EA+ pin is restricted to 1.6V. But again I'm facing same problem that is VCOMP is constant(3.24V). Help us out please.

  • And also i found that REF/EA+ pin is showing 1.6V when it is floating. how is that possible? is that input pin right? and VCOMP is showing 3.24V in this case also. help us out to proceed further.
  • Hi Subash

    Your schematic shows a 5k/25k potential divider at REF/EA+ so the pin should be at 2.75V. When the system is in regulation the error amplifier in the IC will drive the EA- pin to the same voltage ie. 2.75V. To do this will require 272V at Vout. If Vout is lower than this then COMP will saturate high (3.24V), if Vout is higher than this, then COMP will saturate low (close to 0V).

    If you changed the potential divider pin so that the REF/EA+ pin is at 1.6V then the same calculation would indicate that Vout would need to be at 158V.

    I'd suggest that you reduce the regulation setpoint until you can see the COMP pin behaving - for example reduce the 47k to 4.7k - if the VREF/EA+ pin is at 1.6V then Vout regulation setpoint would be at 16V. Once you can see COMP pin coming out of saturation as VOUT goes from less than the setpoint to greater than the setpoint you can then look at increasing the resistor to increase the setpoint - eventually up to the 300V final value.

    Be careful too that the 47k resistor in your schematic has an appropriate voltage rating (at least 300V - 400V would be better ) - if you have a 300V output then your are likely to need 2 or 3 resistors in series to make up the 47k total. If the resistor does breakdown then you are likely to destroy the IC - and other components.

    Why is REF/EA+ pin at 1.6V when it is floating ? - First off this pin should never be allowed to float because that will certainly prevent the controller from controlling the system.  This pin is a very high impedance input - it's the input to some Op Amps - and the measurement probe (DVM or 'scope probe) is likely to load it down from whatever voltage it was originally sitting at. It's impossible to analyse this accurately without detailed knowledge of the measurement system and of the IC internal circuits.

    Regards

    Colin

  • Hi Colin
    Thank you so much for helping us. And i tried by reducing 47k to 4.7k with the same REF/EA+ voltage i.e 2,7V and initially Vcomp voltage was at 3.24V and i increased the output voltage such that voltage at FB/EA- exceed the reference voltage(2.7V) and i found that the COMP Voltage was saturated to zero once the FB/EA- exceeds 2.7V. So is IC working fine? and i saw the PWM pulses in Oscilloscope there were no deadtime between OUTA and OUTB. Could you tell us why?. Why the Vcomp voltage was not vary with respect to the feedback voltage at EA- pin instantaneously? why it is 3.24V or Zero? why not in between values?.
    And thank you so much once again colin.
  • And one more question Mr. Colin. If the Vcomp saturated to zero when EA- voltage exceeds EA+ voltage, then there won't be PWM pulses since VCOMP is zero and i checked the same. If there is no pwm pulse then how will i get output voltage? Because H bridge will not conduct since there is no pwm pulse .
  • I somehow understand this colin. If the output voltage exceeds 300v then it will not generate pwn until it reach below set point..right? but why Vcomp is not changing instantaneously?
  • Hi colin,

                I have been waiting for your reply. Though i have connected resistors at PS and SP pins, Why there is no dead time between OUTA and OUTB and also between OUTA/OUTB and SRA/SRB?. Help us please.

  • Hi Subash
    I was travelling on business all last week and wasn't able to look at this. I will try to look at it today.
    Regards
    Colin
  • Hi Colin,

             I hope you can  understand the problem that we are facing clearly.If is not, then kindly ask me.

    Regards,

    Subash

  • Hi Colin,

                    Despite connecting resistors at PS and SP pins, why there is no deadtime???? I measured voltages at PS and RS pins there were some voltages like 2V.Please help us.

  • Hi Subash

    1/ the dead time is between the SRx and OUTx signals. Can you send me a 'screen shot (a camera picture would do) of OUTA and OUTB - they should be at < 50% duty cycle and at 180° to each other. can you also send a screen shot of OUTA and SRA and of OUTB and SRB. Ideally 3 or 4 switching cycles visible in the image. Use Rich formatting option and post here or email to me as before.


    2/ Question: If the Vcomp saturated to zero when EA- voltage exceeds EA+ voltage, then there won't be PWM pulses since VCOMP is zero and i checked the same. If there is no pwm pulse then how will i get output voltage? Because H bridge will not conduct since there is no pwm pulse? . -
    Ans: The feedback loop will adjust itself so that the EA- voltage is the same as the EA+ voltage - pulling EA- above and below EA- is just a simple way to see if the controller is 'alive'

    3/ Question: I somehow understand this colin. If the output voltage exceeds 300v then it will not generate pwn until it reach below set point..right? but why Vcomp is not changing instantaneously?
    Ans: the feedback loop may be slow - I wouldn't worry about this at this stage - let's get the system switching first.

    4/ Question: If the Vcomp saturated to zero when EA- voltage exceeds EA+ voltage, then there won't be PWM pulses since VCOMP is zero and i checked the same. If there is no pwm pulse then how will i get output voltage? Because H bridge will not conduct since there is no pwm pulse .
    Ans: Correct, if EA- is greater then EA+ then the controller thinks that the Vout is above the control setpoint and it will not deliver PWM pulses. In this case. Vout will reduce at a rate depending on the load - and EA- will also reduce. PWM switching will start when EA- falls belwo EA+

    Regards
    Colin
  • Hi Colin,

                  I have sent you  screenshots of the  PWM pulses. Kindly check it.  Yes the duty cycle should be  <50% but in my case ,it showed 50% dutycyle and there were no deadtime.

  • Hi Subash

    Thanks for the screenshots - they were clear.

     The first thing to do is to get the OUTx and SRx amplitudes increased. 5V is too low to effectively drive the power MOSFETs – these devices normally need about 10V to become fully enhanced. I would increase Vcc to 12V. Let's do that first then worry about the duty cycle.

    OUTA and OUTB are operating at Dmax (50%)  duty cycle and are in antiphase. Antiphase is correct, Dmax is because the control loop is saturated at max demand (Vout is too low for example). There should be no overlap between OUTA and OUTB – in fact there should be a small dead time, but that’s not visible on this timebase. Anyhow, this looks reasonable. The ‘droop’ in the waveforms is probably caused by  the gate drive load – I don’t know how you are driving the MOSFETs – but whatever it is is drawing too much current for the output driver and it comes out of saturation and the voltage becomes distorted. Can you send me details of the driver circuits please.

     OUTA and OUTB should look like SRA and SRB  (Increase VCC to get a correct amplitude on SRA and SRB, but at least they are rectangular)

     Again the timebase is too slow to see the dead times between OUTA and SRA and between OUTB and SRB – can you look at 100ns per div

     I don’t know why OUTA and OUTB disappeared when Vcc increased to 12V – it may have been due to an excessive load from the driver – you could try to completely disconnect the load at OUTA and OUTB and see – does the wave shape improve (even at Vcc = 5V) and is it present when Vcc = 12V.

     

    Regards

    Colin

  • Hi Colin,
    I haven't connected gate driver yet but i had connected OUT A and OUT B to pulse transformer but now i disconnected them too.
  • Hi Subash

    The gate driver transformer may be saturating - check the volt second rating of its primary winding.

    Regards
    Colin
  • Hi Colin,
    I have connected 12V at VDD pin and measured voltage at REF pin. It was showing 7.12V. How was that possible? since the absolute maximum voltage at REF is 3.6V. There were no voltage at COMP pin.It was showing hardly 0.2V(thats why I'm not getting pulse at PWM pins). But IC wasn't heated when i checked. Should change the IC?
  • Ok Colin I will check .I referred TI's application note here it is the link-www.ti.com/.../sluu441a.pdf . Page number:4 Figure:1 you can see the schematic diagram. In that they have used PA0185 pulse transformer(T3). So we are using the same pulse transformer. I have a doubt in that schematic too. In that, OUT A and OUT B are the inputs to the single pulse transformer(T3).I dont know how will work? Explain that too if possible. But now I have disconnected the pulse transformer too.
  • Hi Subash

    I would definitely replace the IC - Vref should be correct - is there a short circuit somewhere that is pulling the reference pin up - solder short or something like that ?.

    You should check the volt-seconds rating of the PA0185 transformer - are you running at a lower frequency than the converter in the app note ? if so the volt-seconds will be greater in your design.

    OUTA and OUTB alternately drive the transformer primary with either a positive voltage (OUTA HI and OUTB LO) or a negative voltage (OUTA LO and OUTB). D10 and D11 rectify the secondary of the transformer so that the driver inputs LI and HI see either a HI/LO or LO/HI combination. This is a fairly normal approach and requires only a single transformer to cross the isolation barrier (cheaper !)

    The capacitor in series with the transformer is there to prevent a DC current in the transformer. The capacitor is large enough that its voltage doesn't change significantly during a single switching cycle.

    Regards
    Colin
  • Hi Colin,

                As you said, I have replaced the IC UCC28251. But the problem is  that again Im facing same problem . Now i have connected enable pin to variable regulated suppy(0-15V) and VDD at 12V supply.Intially i haven't supplied voltage to enable pin but it was showing 2V at EN pin. How was that possible? . And after that i measured voltage at VREF pin. It was showimg 3.24V. Now I started to increase voltage at Enable pin using variable regualted voltage supply. When the voltage at enable pin is greater than 1.9V, I found that VREFvoltage was increased to 7V though there were no physical connection between VREF and EN pins.How was that possible and why was that changing?. And DEVICE WAS NOT HEATED AND IT WAS A NEW IC. 

  • Why VREF volatge changing from 3.2 to 7 V when enable pin is supplied by >1.9V. And when EN pin was floating. It was showing 2 V, How? . And It is input pin, right?
  • Hi Subash

    Vref cannot generate 7V itself -

    What I would do is to double check your PCB layout - look for shorts somewhere. I would take your board, remove the UCC28251 device completely and then see if the voltages you see when the Vcc supply is brought up slowly - do they conform to expectations ?

    Please check the voltage between the negative terminal of your external Vcc power supply and the GND pin (the actual metal of the pin as it goes into the IC body) - the voltage should of course be zero. Alternatively, check the resistance of the connection - it should be zero (or very close to zero)

    Regards

    Colin

  • Hi colin,

                  sorry i was on leave that is why i couldn't reply you. As you said, i will check today my circuit connections. And if we short VREF and EN pin, will it be a problem?

  • Welcome back !

    Shorting EN to VREF will be ok - you will of course lose the ability to use the EN pin to turn the controller on and off but otherwise doing this will be ok.

    Regards
    Colin

  • Hi Colin,
    I found the mistake. It was due to Rcs(RAMP resistance) value. When I disconnected the ramp part in VREF pin, it worked fine. So i replaced 30K resistance with a 390Kohm resistance . Now it is okay and Vramp peak value is 0.150V. And one more help I need that is how to design ramp amplitude. I'm bit confused with the formulas used. My design criteria is dmin=0.33 and dmax=0.45 so what should be the value of Rcs(Ramp resistance) ? . I have connected 470pf capacitance as Ccs.
  • Hi colin,
    I am using IR2110 MOSFET driver to drive H- bridge MOSFET. But when I connect OUTA and OUTB pin to Hin and Lin pins of IR2110, the IC UCC28251 started to Heat. I don't know why. But if OUTA and OUTB were kept to float, there were no heating issue. I have PA0185 pulse transformer but my switching frequency is 50Khz so it exceeds the volt-sec limit of the transformer and so i tried without isolation for testing purpose.
  • One important point is I didn't powered up the gate driver(IR2110). I searched for shorts in the driver but it was fine. There were no shorts.
  • And I tried with MCT2E opto coupler. Again it started to heat.
  • Do we need to connect 100pf capacitors on OUTA, OUTB, SRA and SRB?

  • Hi Subash

    No - you should not connect any capacitors to these pins -

    Capacitance at these pins will slow the rise and fall of the waveforms and won't improve system performance.

    Regards

    Colin

  • Hi Colin
    Ok. But why the IC heats when OUT A and OUTB connected to the driver or to the pulse transformer?
  • Hello Subash

    I've been out of the office a lot recently and it's been a while since this thread was active - please let me know if the issue is still open.

    Regards
    Colin
  • Hi colin,
    Actually i had replaced the pulse transformer with a new one which has more VT product than previous one. Now it is ok. I will contact you if I found any problem. And thank you so much.