classd2008
New member
norazmi post your PCB of Discret ClassD. Thank-you.
Class D 200 Wrms with 2 mosfet cheap
- Thread starter norazmi
- Start date
if u do modification to the gate driver with incorrect value, those transistor will be saturate and will make the amp unstable, lots of spikes.
there is some explanation to this circuit must take a note.
For the UCD design, change the 2n5xx1 by mpsAx2 (save that the legs are backwards) and mosfets for more voltage mosfets ... should not take more modifications - (well, maybe the feedback resistor to have the maximum power with the same input signal).
Now, for all unbelievers who think that this amp can give 200Wrms or more ... just depends on the mosfets. Some quick calculations:
For a power W (rms) on a speaker of R ohms, you need a supply voltage of V = sqrt (W * 2 * R), where sqrt = square root.
The maximum current that will ask the power supply is R = V / R. As the mosfets or totally turned on (fully saturated) or totalmenteapagados (Class D is!) then the dissipation in each MOSFET is pmos = I * I * RDSON / 2 (s / 2 is because alternate mosfets are never turned on both at once, and on average, is half the time on each). Now, specific calculations: For the mosfets used, the RDSON is 0.3 ohms. We take 200Wrms about 4 ohms. little formula According to the above, we need a supply voltage of 40 volts. The maximum current that is going peak to ask the source is from 40/4 = 10A. maximum dissipation of 15W each mosfet will. For the mosfets do not burn overtemperature, MUST go on a heatsink ... Now calculate this sink> assume a temperature the air surrounding the heatsink 50C, the maximum junction temperature of 150C is mosfet.
That will give a thermal resistance junction-ambient (150-50) / 15W = 6.6C / W. The mosfet and has a thermal resistance -joint casing 1.5C / W, so that the thermal resistance remains is 6.6 - 1.5 = 5.1C / W. Assume an extra heat resistance mica and dissipating grease 0.5C / W. This causes the heatsink required for each MOSFET has to be 5.1C / W - 0.5C / W, 4.6C / W. This cooler there, and it is perfectly possible to use. All these calculations are for the worst .. In reality, the audio is not at full power ever, so there is a safety margin far greater than you can believe initially. If only we wanted to get 100W, the heatsink is considerably reduced. If instead of using mosfets specified mosfets with lower RDSON could either increased or reduced power dissipation in the same further. In relation to the coil to the amplifier propose, suggest them instead of a coil, a coil and a capacitor: 10uH coil and capacitor of 1.5 uF (poly) ... The facts are simple ... The cutoff frequency of the filter must be higher than 20Khz, but below the minimum switching frequency of the amplifier (near 500Khz) ... Fcorte = 1 / (2 * PI * sqrt (L * C)), L = Inductanciade coil in henries, C = capacity in farads, Fcorte = cutoff
frequency in Hz, PI = 3.1416. will see that there is an enormous range of values that with that relationship. So I suggest that C = 1 / (2 * PI * R * Fcorte), where R is the maximum resistance of the speaker to use ... I mean, if they are to use it with 2 different speakers, and an 8 ohms, the calculation made ??for 8 ohms. The filter capacitor that is parallel with the speaker, and the coil is in series with the same (ie, the coil conexta the amplifier output with a speaker terminals). For cdo UCD amplifier, coil and capacitor are already part of the amplifier. Finally, the issue of performance: d = 100 * Pparlante / PMOS (d = percent yield, Pparlante, power in the speaker, PMOS = power loss
in the mosfets).
say drivers get hot because the switching frequency is high (1 MHz) and have to deliver large current spikes ... There are 2 possible solutions: Either decrease the frequency of switching (switch C9 by 2n2 - This slightly affects the fidelity at high frequency, but Try it), or less able to use mosfets gate-source (Sliconix any). However, warming did not consider it dangerous. The other is that they hit a disipation on each transistor ... I think that is not worth the trouble.
On the subject of the feedback resistor, you can change if the input signal does not exceed 2v p = p. That is, the 2nd operational output should not exceed in any case 2Vpp. If you have the assurance that the input signal is not going to happen never, 1 Vpp, then make the change, but if not, no. not change the supply voltage of the TL074: This makes the output stage may not go well .. To be exact, The bases of Q1 and Q2 must not exceed the 3 volts, otherwise, the transistors go into saturation, and that's just what we DO NOT have to happen. The saturation dissipate very much power. Everything is
calculated for the issuer in relation to mass up 2.64 volts, which generates a voltage drop on R5 and R6 12V approx.
in relation to the supply voltage. The 12 V zener are a precaution and not because they operate as limiters. Avoid at all costs the saturation of any of these transistors smaller, because they cease to operate at the required speed. The only people who should saturate are the mosfets. If you need to give more amplification, modification are needed for the R16 or R14 and nothing else.
regards.
..been studying these for quite sometime...
Thanks for the details Norazmi.
regsixpence
New member
thanks a lot for the projects guys,will do some simulation,compare the practical results and simulated.
regsixpence
New member
Thanks
Thanks for providing this detailed explanation lol
Thanks for providing this detailed explanation lol
ahh, im forgot, i already mention with same value at driver component it will become unstable with overheat at mosfet and will be failed to operate at +/- 75 vdc even with single pair mosfet, double it u make it more worst of this amp. The better solution is go with ucd style ir2110 or discrete ucd with stewin design, believe me with discrete u will not disappointed at all, or go with NON-ucd ir2110 + TL074, yes u can double the mosfet or triple it with totem pole driver, calculation must be made for the driver depends on what type mosfet u will use, more faster will be better and u can get crystal clear sound at output.
Btw im still doing simulation with +/- 65 vdc, its clear wave so far, i need to check with component which maybe can have overheat situation. 2 pair of mosfet will work, but driver will suffer more to supply enough current, but u can try it with voltage +/- 35 vdc at first test.
Running at +/- 65 vdc, 100 hz wave input test. Problem with P channel driver heat lots than N channel. seems that P channel hard to past at high voltage, we can reduce it by reduce swithing freq but... will be loose at hi fidelity sound quality too . and it will be suitable for subwoofer low pass filter only sound.
Fourier components of V(out)
DC component:0.099346
Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 7.530e+00 1.000e+00 -174.30° 0.00°
2 2.000e+02 3.068e-02 4.074e-03 1.84° 176.14°
3 3.000e+02 2.054e-02 2.728e-03 0.23° 174.53°
4 4.000e+02 1.540e-02 2.045e-03 -0.96° 173.34°
5 5.000e+02 1.229e-02 1.632e-03 -1.76° 172.54°
6 6.000e+02 1.027e-02 1.364e-03 -2.69° 171.61°
7 7.000e+02 8.723e-03 1.158e-03 -3.48° 170.82°
8 8.000e+02 7.688e-03 1.021e-03 -3.95° 170.34°
9 9.000e+02 6.837e-03 9.080e-04 -4.90° 169.40°
Total Harmonic Distortion: 0.199611%
Date: Sun May 13 02:24:38 2012
Total elapsed time: 992.885 seconds.
tnom = 27
temp = 27
method = modified trap
totiter = 7774094
traniter = 7772357
tranpoints = 1653333
accept = 1174357
rejected = 483545
matrix size = 132
fillins = 1276
solver = Normal
Matrix Compiler1: 188 KB object code size 55.6/33.5/[21.3]
Matrix Compiler2: 2223 opcodes 9.5/[8.8]/295.8
. and it will be suitable for subwoofer low pass filter only sound.Fourier components of V(out)
DC component:0.099346
Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 7.530e+00 1.000e+00 -174.30° 0.00°
2 2.000e+02 3.068e-02 4.074e-03 1.84° 176.14°
3 3.000e+02 2.054e-02 2.728e-03 0.23° 174.53°
4 4.000e+02 1.540e-02 2.045e-03 -0.96° 173.34°
5 5.000e+02 1.229e-02 1.632e-03 -1.76° 172.54°
6 6.000e+02 1.027e-02 1.364e-03 -2.69° 171.61°
7 7.000e+02 8.723e-03 1.158e-03 -3.48° 170.82°
8 8.000e+02 7.688e-03 1.021e-03 -3.95° 170.34°
9 9.000e+02 6.837e-03 9.080e-04 -4.90° 169.40°
Total Harmonic Distortion: 0.199611%
Date: Sun May 13 02:24:38 2012
Total elapsed time: 992.885 seconds.
tnom = 27
temp = 27
method = modified trap
totiter = 7774094
traniter = 7772357
tranpoints = 1653333
accept = 1174357
rejected = 483545
matrix size = 132
fillins = 1276
solver = Normal
Matrix Compiler1: 188 KB object code size 55.6/33.5/[21.3]
Matrix Compiler2: 2223 opcodes 9.5/[8.8]/295.8
Hi Glo,
D1-2 is fast diode, u can use any super fast diode, such as mur120, mur160, uf4004, uf4007, 1N5819, SB160 and etc...
BZX85CL12 is 12v zener diode, u can use any 12v zener diode rated 1 w.
its safe to run at +/- 75 vdc as long as u follow the ucd-ir2110 table for component change, and remember for the mosfet use and its limited to drive load with higher rds-on, mosfet will suffer means more heats need big heatsink.
i did this amp first time long time a go , im forgot where i put the pcb design for it, and this amp not dead yet so i never look for the pcb file yet. im trying to find but suddenly cant find it, btw this circuit is not hard to design and its simple with few component only, but as many of us know, class D amp need good trace, compact and avoid long trace.
regards,
Azmi
D1-2 is fast diode, u can use any super fast diode, such as mur120, mur160, uf4004, uf4007, 1N5819, SB160 and etc...
BZX85CL12 is 12v zener diode, u can use any 12v zener diode rated 1 w.
its safe to run at +/- 75 vdc as long as u follow the ucd-ir2110 table for component change, and remember for the mosfet use and its limited to drive load with higher rds-on, mosfet will suffer means more heats need big heatsink.
i did this amp first time long time a go , im forgot where i put the pcb design for it, and this amp not dead yet so i never look for the pcb file yet. im trying to find but suddenly cant find it, btw this circuit is not hard to design and its simple with few component only, but as many of us know, class D amp need good trace, compact and avoid long trace.
regards,
Azmi
http://desmond.imageshack.us/Himg526/scaled.php?serv[IMG]
[IMG]http://http://desmond.imageshack.us/Himg59/scaled.php?server=59&filename=dfoil.jpg&res=landing
I did some mod on the PCB of this D Amp in an attempt to lower the DC offset, as you can notice the output cap was tapped to the starground.
DC offset was at 27.0mv but before it was 30.0mv. Can this D amp be wired the same as a Class AB amp? Particularly the speaker-out where the - of the speaker out is tapped at the starground of the PSU.
[IMG]http://http://desmond.imageshack.us/Himg59/scaled.php?server=59&filename=dfoil.jpg&res=landing
I did some mod on the PCB of this D Amp in an attempt to lower the DC offset, as you can notice the output cap was tapped to the starground.
DC offset was at 27.0mv but before it was 30.0mv. Can this D amp be wired the same as a Class AB amp? Particularly the speaker-out where the - of the speaker out is tapped at the starground of the PSU.
