Also, do you have evaluation board for Half bridge converter with double ended transformer coupled gate drive?

Correction: Instead of a PWM controller such as LM5039, I would prefer something simple such as UCC28083, which offers what is needed. Thank you

To be precise - A PWM controller UCC28086D.

I forgot to mention that the operating frequency is 25kHz - 40kHZ

Hi Sean,

Thank you for your reply.

1) I will use the UCC21550 and UCC28086D. Please below the attached figure

The attached figure is the intended application. I will use the UCC28086D as the PWM controller, it has an output voltage of 4.3V, which is suitable for the UCC21550. 

1b. The UCC21550 IC has given three different examples of external isolated bias supplies for the output channels of the UCC21550 to support a 500V output. I am considering using the third example as shown in page 33- Figure 8-4 of the datasheet, since the UCC28086D will be operating at 50% duty cycle. Is this the best option?

1a.  Given my application of the UCC28086D, is it suitable to i) Pull up the CRTL to the VDD with 5k -10k resistor? ii) Pull down the CS pin with 1k resistor?

2) Are you referring to the VDD (supply voltage) of the digital potentiometer? (Part number MCP4017T-503E/LT, Microchip Technology) It has an operating supply voltage of 1.8V - 5.5V. The voltage on the A and W pin of the pot are VDD +0.3V. 

3) Just to confirm not to include those components in blue-coloured below. 

4) Figure 1 is the first figure above. 

Thank you for your recommendation. 

Kind regards,

Bright

5) Regarding isolated power supply for UCC21550, is it suitable to use auxiliary winding of a flyback transformer to bias the external isolated power supply? Instead of having a new isolated power supply. 

Hi Bright Robert,

1b. bootstrap_circuit_negative_turnoff.TSC

The problem is that the Cz has to charge. If it is too large, you will apply too high +V for a few cycles. If it's too small, then it forms a divider with the FET Cgs, and your output range goes down. Instead I recommend a small transformer and a push pull driver with the SN6501. That will give you the right supply voltage.

1c. No, you will still need to use this pin for feedback. You will need to "trick" the device into thinking it is regulating a lower voltage.

2. Great! 2V is the voltage that the IC will apply to the resistor, when it senses the DC current.

3. Confirmed, refer to my 1b image for my reasoning. This circuit either gives too much positive voltage (transient) or too little negative voltage, for high and low Cz respectively.

Hi Sean,

Thank you for your reply, it is very insightful.

I am not sure that I understood the point made about the small transformer and a push pull driver with the SN6501. Point 3 is understood, no need for those components circled in blue colour as mentioned about. 

I need  clarification  regarding the recommendation that you made on a push pull driver with the SN6501.

1) Are you referring to power supply to the isolated side of the gate driver (UCC21550)? 

2)  If the answer to 1) is no, Will 16V auxiliary winding of a flyback transformer  suitable for power the isolated side of the IC? 

3) Do I need to  use a pull driver with the SN6501 to supply those voltage rails to the Isolated side?

Kind regards,

Bright

Also, regarding Point 2) Great! 2V is the voltage that the IC will apply to the resistor, when it senses the DC current.

The digital potentiometer voltage at the pin will be VDD (2V) =0.3V. Will that be okay? On the datasheet, it says that RT voltage is 2V absolute. 

Kind regards,

Bright

Also, if the 2.1 -2.3V is not possible due to the absolute rating on that pin, can I use a zener diode between the RT pin and the digital potentiometer pin W to clamp the voltage across those pins to 1.8V - 2V? 

Hi Bright,

1. Yes. 

2. Even so, you can use a flyback, but it will take up more room.

3. you need to use the SN6501 to drive the primary of a small transformer, and you rectify the secondary to get your 16V gate driver supply rail.

4. RT pin voltage looks like 1.5V typ from last line on page 3. Looks okay to me.

You should probably order and EVM for the UCC28086 and see what you need to do with the RT pin in a lower voltage test. You can probably ask another E2E question with that device highlighted to get to the experts for that device. I don't think you want to put in a series zener, you should just use that digital potentiometer.

Best regards,

Sean

Hi Sean, 

Question: 1a.  Given my application of the UCC28086D, is it suitable to i) Pull up the CRTL to the VDD with 5k -10k resistor? ii) Pull down the CS pin with 1k resistor? 

1c. No, you will still need to use this pin for feedback. You will need to "trick" the device into thinking it is regulating a lower voltage.

-I take that you are referring to the CS pin of the UCC28086D IC? 

See the Figure below. The PWM controller is the UCC28086D IC in Red coloured rect-box. 

1) Where can I connect the CS in this configuration? 

Hence, considering connecting 1k r between the CS pin and GND. 

Kind regards

Hi Bright Robert,

I can help you with the snubber! The most important thing to remember is that the switch node is always connected to one suppply. High amplitude ringing often appears on the supply too. You can snub this ringing with a snubber across the high voltage bus, without having to cycle the snubber capacitor every cycle. This saves a lot of power vs. a snubber on Vsw itself. An electrolytic capacitor can give you a high ESR and also a high power rating and the right selection can be a very good snubber. Since they have a shorter lifetime than most electronic components, they are often replaced with Tantalum capacitors with many power resistors in parallel.

You have to measure the switch node ringing frequency, and than estimate the parasitic Csw and Ld values. The FET datasheet can give you a good estimate of the Cds of the FET, and you can add some extra capacitance for the estimated board parasitics. You can then solve for Ld with the equation (2*pi*f)^2=1/LC.

You can then plot both these L and C impedances on a Bode plot. The intersection will tell you how much resistance is needed to dampen the resonance.

Once you add this on the HV Bus, you might need to adjust it and monitor the effect on overshoot to account for imprecise estimation of Csw and Ld.

You might still have some high frequency, low power resonance on the switch node.  For this you can use a much smaller 0603 resistor snubber. You can use the same process as above, but now you have to limit your capacitor size, since it will cycle with the switch node. The CV^2 * fsw must stay below the power rating of the resistor. Then, you can use the bode plot process below to pick a resistor that makes your snubber and Csw parallel look as resistive as possible at the resonant frequency.

Best regards,

Sean

GateRing3 - autosave 23-11-29 11_36.TSC

Can you please start a seperate E2E question thread for this question pertaining to the UCC28086D?