Requirements and selection of MOSFETs in step-down DC/DC converters

Synchronous rectification step-down DC/DC converters all use a controller and an external power MOSFET structure. The controller manufacturer will provide a complete application circuit in the data sheet, but the user's operating conditions are often different from the typical application circuit, and the parameters of the power MOSFET must be changed according to the actual situation.

Requirements for power MOSFETs

The input and output circuit of the synchronous rectifier step-down DC/DC converter is shown in Figure 1. It consists of a controller with a driving MOSFET, an external switch tube (Q1), and a synchronous rectifier tube (Q2). At present, both Q1 and Q2 use n-channel power MOSFETs because they can meet the requirements of the DC/DC converter in terms of input voltage, switching frequency, output current, and loss reduction.

Figure 1. Schematic diagram of the input and output circuit of a synchronous rectifier step-down dc/dc converter

The working conditions of the switch tube and the synchronous rectifier tube are different, and their losses are also different. The switch tube has conduction loss (or conduction loss) and gate drive loss (or switching loss), while the synchronous rectifier tube only has conduction loss.

Conduction loss is caused by the on-resistance r _{ds(on) of the mosfet, and its loss is related to I} ² d, r _ds(on) and duty cycle. To reduce conduction loss, it is necessary to select a power mosfet with a small r _{ds(on) . The r ds(on)} of a new mosfet is about 10mω at v _gs = 10v, and some new products can achieve an r _{ds(on) of about 2 to 3mω at v gs} = 10v .

Gate drive loss is the loss caused by charging (establishing VGS voltage to turn on MOSFET) and discharging (making VGS = 0 to turn off MOSFET) the inter-electrode capacitance (as shown in Figure 2) of MOSFET at a certain gate-source voltage VGS _when the switch _{is turned} on and off. This loss is proportional to the input capacitance C _iss or feedback capacitance C _rss of MOSFET , gate drive voltage V _gs and switching frequency F _sw . To reduce this loss, it is necessary to select a power MOSFET with small C _iss or C _rss and low threshold voltage V _{gs (th) .}

Figure 2: Capacitance between electrodes of MOSFET

The synchronous rectifier also works in a switching state (its switching frequency is the same as that of the switching tube), but because the synchronous rectifier works in a zero voltage (vgs≈0v ₎ state (as shown in Figure 3), its switching loss can be ignored.

Figure 3 When the synchronous rectifier is turned on, vds≈0v

In order to ensure the safe, reliable and efficient operation of the DC/DC converter, the selected power MOSFET must meet the following conditions under a certain gate drive voltage: the withstand voltage of the MOSFET must be greater than the maximum input voltage, that is, V _dss >V _in(max) ; the drain current of the MOSFET must be greater than or equal to the maximum output current, that is, I _d ≥I _out(max) ; select a switch tube with a _Ciss or C _rss as small as possible, and select a synchronous rectifier tube with an R _ds(on) as small as possible to minimize the loss of the MOSFET and meet the requirement that its loss value is less than PD ₍ PD _is the allowable dissipated power of the MOSFET under certain conditions). In addition, it is also necessary to select a SMD MOSFET with a moderate price and small package size (such as SO-8, DPAK or D2PAK package).

_{Parameters such as V dss} , I _d and r _ds(on) of mosfet can be found directly from mosfet samples or data sheets, while its loss can only be determined by calculation under certain conditions.

Calculation of losses in mosfet

DC/DC controller manufacturers provide loss calculation formulas for switching tubes and synchronous rectifier tubes in their data sheets. The calculation of switching loss is often an empirical formula, so the formulas of different DC/DC controller manufacturers are different. The calculation must be based on the loss formula provided in the data sheet for that model, otherwise there will be a large calculation error.

The method of loss calculation is to preliminarily select a power mosfet according to the known use conditions, which must meet the requirements of v _dss >v _in(max) , I _d ≥I _out(max) , small ciss or crss, and small r _ds ( _{on )} , and then calculate its loss according to the formula. If the calculated loss is less than p _d under certain conditions , the calculation is valid and the preliminarily selected power mosfet can be selected; if the calculated loss is greater than p _d , reselect or use two power mosfets in parallel to make 1/2 (the calculated loss)

d.

Before calculation, you need to know: input voltage vin (or vin _(max) and vin(min)), output voltage vout, maximum output current Iout _(max) , switching frequency fsw _{. Generally, the pd} of the selected MOSFET is often 1~1.5W, the purpose is to reduce loss and improve efficiency.

This article introduces the calculation of MOSFET loss in the circuit composed of Maxim's MAX8720 single-phase buck DC/DC controller and Fairchild's multi-phase buck DC/DC controller FAN5019B. The loss calculation formula is very simple, the key is how to correctly select the relevant parameters from the MOSFET sample or data.

Selection of main parameters of mosfet

• Selection of I _d and PD _values ​​In
  
  the data of MOSFET, the values ​​of drain current I _d and allowable dissipation power PD _are different under different conditions, and their values ​​vary greatly. For example, the limit parameters of n-channel power MOSFET IRF6617 are shown in Table 1.
  
  Table 1 Limit values ​​under continuous working state
  
  
  Maximum drain current IDM = 120A (pulse state operation limited by maximum junction temperature).
  
  Different MOSFET manufacturers have different ways of expressing I _d and PD _{. For example, the I d} and _PD parameters of ON Semiconductor's NTMFS4108N are shown in Table 2.
  
  Table 2 Maximum limit values ​​(tj=25℃, otherwise specified)
  
  Note: * Installation condition 1 is that the MOSFET is installed on a pad with an area of ​​6.5cm2 on the copper plate (see Figure 4)
  ** Installation condition 2 is that the MOSFET is installed on a pad with an area of ​​2.7cm2 on the copper plate (see Figure 4)
  
  Maximum leakage current IDM=106A (pulse state, tp=10μs).
  
  In the DC/DC converter, the MOSFET works in a pulse state with a changing duty cycle, but not in a narrow pulse state; the operating temperature range is -40～85℃. There are no Id and Pd values ​​under such working conditions in Tables 1 and 2. _Id can _be selected _within the following range: (Id at ta=70～85℃ ₎ d≤continuous or short-term maximum value. For example, Id in Table 1 _{can be 11～55A, and Id} in Table 2 can be 16～35A. _Pd is generally selected as the minimum value.
  
• Selection of r _ds(on) value The typical r _ds(on) value and the maximum r _ds(on) value when
  
  the junction temperature tj=25℃, v _gs =10v and v _{gs =4.5v are given in the MOSFET data. In addition, r ds(on)} also increases with the increase of junction temperature. Generally, r _{ds(on) is calculated by taking the maximum value of r ds(on)} under the condition of known v _gs (determined by the vcc of the driver or controller) .
  
• The selection of Ciss and _Crss When calculating the loss of the switch tube, the input capacitance Ciss (Ciss = Cgd + Cgs) or the feedback capacitance Crss (Crss = Cgs) value is used _{. In} order _to reduce _the switching _loss , _a MOSFET _with a _small Ciss _or Crss _should be selected. _{Ciss is} generally thousands to thousands of pf, while Crss _is generally tens to hundreds of pf. There is often no _Ciss or _Crss parameter in the "MOSFET Selection Guide" or "Brief Table" , but there is a total gate capacitance Qg value. Since the MOSFET with a small Qg has a small _Ciss or _{Crss , it can first find the MOSFET model with a small Qg, and then find the Ciss} or _Crss value in the data sheet . Some data sheets do not have Ciss _or C _RSS parameters in their parameter tables, but do have characteristic curves of Ciss _{, C RSS} and VDS . The _Ciss or C _RSS value when VDS = 15V can be taken as the calculated value, as shown in Figure 5.
  
  
  

Figure 4 MOSFET pad (copper clad board) size

Figure 5 Characteristic curves of _{CISS , CRSS} and VDS

Applications

• The MOSFET selection in the MAX8720 circuit
  
  is shown in Figure 6. The step-down DC/DC converter circuit composed of MAX8720 is shown in Figure 6. The current conditions are Vin = 7 ~ 24V, Vout = 1.25V, I _{out (max)} = 15A, F _sw = 300kHz, the operating voltage (bias voltage) of the controller is VCC = 5V, and the appropriate switch tube (NH) and synchronous rectifier tube (NL) are selected.
  
  

  
  Figure 6 Step-down DC/DC circuit composed of MAX8720
  

  
  The SI7390DP of Vishay was initially selected as the NH (its Qg is only 10nC); the SI7356DP was selected as the NL (R _DS(ON) = 4mω). The packages are all 8-pin SO-8 packages with heatsinks. The main parameters are shown in Table 3. Notes to
  
  Table 3
  : * Obtained from the characteristic curve; ** The minimum value of the printed board pad area.
  
  
  1. Switching tube conduction loss pd(nhr ₎ calculation
     
     pd( _nhr )=(vout/vin(min))(Iout _(max) )2× _rds(on)
     =(1.25v/7v)×15a2×13.5mω
     =0.54w
     
  2. The gate drive loss PD(n _h s) of the switch tube is calculated as
     
     PD(n _h s)=[(vin _(max) )2×c _rss ×f _sw ×iout]/igate
     =[24v2×130pf×300khz×15a]/2a=0.168w,
     where
     
     the data of the gate current igate is given in the MAX8720 data sheet. The total loss PD(n _h )
     of the switch tube is PD(n _h )=PD(n _h r)+PD(n _h s) =0.54w+0.168w =0.708w<1.1w
     
     
     
     
     
  3. The conduction loss pd(n _l r) of the synchronous rectifier is calculated as follows:
     
     pd(n _l r)=[1-(vout/vin _(max) )]×( _Iout(max) )2× _rds(on)
     =[1-(1.25v/24v)]×15a2×4mω
     =0.85w<1.9w
     
     According to the above calculation, the calculated loss value is satisfied

     d, you can choose si7390dp and si7356dp.
     

• The MOSFET selection in the fan5019b circuit
  
  is shown in Figure 7. The three-phase synchronous rectifier step-down DC/DC converter circuit composed of the controller fan5019b and three drivers fan5009 is shown in Figure 7. The current conditions are: vin = vcc = 12v (vcc is the working voltage for the controller and driver), vout = 1.5v, Io ₌ 65a (Io _is Iout _(max) ), fsw ₌ 228khz, and the switch tube and synchronous rectifier tube (two parallel connections are used).
  
  
  Figure 7 The step-down DC/DC circuit composed of fan5019b
  
  initially selects the fast switch tube fdd6696 as the switch tube (its qg is 17nc), and the synchronous rectifier tube selects fdd6682 (its rds _(on) = 11.9mω). Its main parameters are shown in Table 4.
  
  Table 4
  
  Note: *p _d is related to the copper clad area of ​​the pcb, which is the minimum area.
  
  1. The calculation of the switch tube conduction loss pc _(mf)
     
     pc _(mf) =d[(Io _/ nmf ₎ 2+1/12(n×Ir _/ nmf ₎ 2]× _rds(on)
     
     where d is the duty cycle (d=vin/vout); Ir _is the ripple current (Ir ₌ 1/n×Io _× 40%); nmf _is the total number of switch tubes; n is the number of phases; Io ₌ Iout _(max) . _Ir is calculated to get _Ir =8.66a, which is substituted into the formula:
     
     pc _(mf) =15v/12v[(65a/3)2+1/12(3×8.66a/2)2]×15mω=0.89w
     
  2. The gate drive loss of the switch tube PS _(MF) is calculated as
     
     PS _(MF) = 2f _sw × Vcc × (I _o / n _mf ) × RG × (n _mf / n) × _CISS
     = 2 × 228 khz × 12 v (65 a / 3) × 3 ω × (3 / 3) × 2058 pf
     = 0.73 w,
     
     where RG is the gate resistance (this includes the gate resistance of the mosfet and the internal resistance of the driver). The typical value of RG for high-speed switch tubes is 1 ω, and the internal resistance of the driver fan5009 is about 2 ω, so RG is 3 ω. The total loss of the switch tube = 0.89 w + 0.73 w = 1.62 w, which is slightly larger than the allowable power dissipation of fdd6696. Since the allowable power dissipation is the p _d
     
     value under the minimum area of ​​the copper clad board , as long as the pcb has a larger pad, this 0.02 w can be ignored.
     
  3. Synchronous rectifier conduction loss psf calculation
     
     psf = (1-d) [(I _o / nsf) 2 + 1/12 (nir / nsf) 2] r _{ds (on)}
     = (1-1.5v / 12v) [(65a / 6) 2 + 1/12 (3 × 8.66a / 6) 2] × 11.9mω
     = 1.24w < 1.6w
     
     where nsf is the total number of synchronous rectifiers.
     
     Through the above calculation, the circuit composed of 3 fdd6696 as switch tubes and 6 fdd6682 as synchronous rectifiers can meet the requirements of vin = vcc = 12v, vout = 1.5v, I _{out (max)} = 65a, f _sw = 228khz.

Conclusion

Since the above loss calculation is rough, after the MOSFET is selected, it is necessary to verify whether the selection is appropriate through experiments. In practice, for safety reasons, the pad area of ​​the MOSFET can be larger (such as 6.5cm2).

references

1. Maxim's max8720 data sheet
2. 2Fairchild's fan5019b and fan5009 data
3. 3. MOSFET data from Vishay, Fairchild, ON Semiconductor, and IR