Hi all,here is simple D amp l designed for rails about to +-50V.
Fets you can use can be any low gate capacity,for example IRF540,IRF640, or some specialized D class fets. Working frequency is about 170...180kHz.
Output inductor should be in range of 40 to 80uH. Mine l salvaged from old TV set, from high voltage transformer. Used is only one half, it needs about 20 to 24turns. I used standard cable wire ,0.75mm2.
Amp is working stable as its using self osc hysteretic modulation (it must oscillate). Circuit is made to be discrete comparator,no ic or anything. Input resistor is 1K. Offset resistor will be in range from 680k to 3M3,depend on your rail and input pnp.
Input pnp can be 2n5551,mpsa92,mje350 or any with Uce over 80V. There was no decoupling C on purpose. I guess it can be driven at higher rails,as this driver is tested on 950vdc for something (but on lower freq).
Pcb is drawn this way because its milled on cnc. I gues there is not lot to say more,as scheme is really simple. If you want you can add LC lpf filter (1mH small resistor like choke,and 2n2...4n7 C to gnd),if have interference issues with smps,etc. Aux 12 to 15V psu can be make with simple transistor zener linear regulator (see pcb) tied between gnd and -V.
Aux psu is referenced to -V (not GND!),and need 300mA max.
Gate diode can be 1n5819,sb160,etc.
Power schottky need to be at least 6A, voltage is not important, can be mr or mbr series,you can find it in pc psu for example. Output filter inductor value and load resistance directly dictates high frequency end response as in any D amp.
This amp is not for professional use but have very good results, no squeaking or similar artefacts as other diy d amps (l have tried lot of it before l started design my own). Clipping is little harder than AB amps but which amp is intended to be listened in clip mode?
Sine wave on picture is on about 50...60% of max power,at 3 ohm load and its linear up to ~95% of rail.
If you using higher rails you probably will get deadtime issue (fet heating) when high fet is turning off and low on.
Rails : +-15 to +-50
Gain : 47 (change with both 47k resistors to maintain same freq)
Configuration : inverting
Input impedance : 1k
Working frequency :~170k
Load : depend on fets, can be low as 0.5ohm but no point
Fets you can use can be any low gate capacity,for example IRF540,IRF640, or some specialized D class fets. Working frequency is about 170...180kHz.
Output inductor should be in range of 40 to 80uH. Mine l salvaged from old TV set, from high voltage transformer. Used is only one half, it needs about 20 to 24turns. I used standard cable wire ,0.75mm2.
Amp is working stable as its using self osc hysteretic modulation (it must oscillate). Circuit is made to be discrete comparator,no ic or anything. Input resistor is 1K. Offset resistor will be in range from 680k to 3M3,depend on your rail and input pnp.
Input pnp can be 2n5551,mpsa92,mje350 or any with Uce over 80V. There was no decoupling C on purpose. I guess it can be driven at higher rails,as this driver is tested on 950vdc for something (but on lower freq).
Pcb is drawn this way because its milled on cnc. I gues there is not lot to say more,as scheme is really simple. If you want you can add LC lpf filter (1mH small resistor like choke,and 2n2...4n7 C to gnd),if have interference issues with smps,etc. Aux 12 to 15V psu can be make with simple transistor zener linear regulator (see pcb) tied between gnd and -V.
Aux psu is referenced to -V (not GND!),and need 300mA max.
Gate diode can be 1n5819,sb160,etc.
Power schottky need to be at least 6A, voltage is not important, can be mr or mbr series,you can find it in pc psu for example. Output filter inductor value and load resistance directly dictates high frequency end response as in any D amp.
This amp is not for professional use but have very good results, no squeaking or similar artefacts as other diy d amps (l have tried lot of it before l started design my own). Clipping is little harder than AB amps but which amp is intended to be listened in clip mode?
Sine wave on picture is on about 50...60% of max power,at 3 ohm load and its linear up to ~95% of rail.
If you using higher rails you probably will get deadtime issue (fet heating) when high fet is turning off and low on.
Rails : +-15 to +-50
Gain : 47 (change with both 47k resistors to maintain same freq)
Configuration : inverting
Input impedance : 1k
Working frequency :~170k
Load : depend on fets, can be low as 0.5ohm but no point
Attachments
Last edited:
Hi!
What a fine design!
What a fine design!
- It looks like it has both positive and negative FB, probably it has complex output impedance, did you simulate/measure it?
- What surprised me, the base of an input PNP is grounded via 10uF cap, is it drawing error or this works?
- Schematic has cut off the top, would you upload whole sheet, power supply?
- Does the sound is so good like TDA8932s designs?
- Lower half of the waveform on the scope is ringing, really no squeaking sound?
Hi my slavic brother 🙂
- It has negative fb in aspect of audio. Positive fb is used for setting tresholds for comparator (range of 10nf charge discharge voltages).
- Base is not grounded,that is input.
Base is grounded with 10nF to gnd and thats Ct for frequency.
- Power supply is most basic linear regulator with npn and zener.
- I dont know how these TDA sounds,as l never have any in workshop.
- Where you see ringing?
- It has negative fb in aspect of audio. Positive fb is used for setting tresholds for comparator (range of 10nf charge discharge voltages).
- Base is not grounded,that is input.
Base is grounded with 10nF to gnd and thats Ct for frequency.
- Power supply is most basic linear regulator with npn and zener.
- I dont know how these TDA sounds,as l never have any in workshop.
- Where you see ringing?
Ahh that.This is because this simple scheme cant have big dead time when low fet is ON and high OFF. In terms of audio quality,thats better as smaller dead time mean less thd. Also thats high freq oscillation at mosfet output,not audio wave. Audio is on 4th pic.
Test at about 0.3ohm load. Whole resistor is about 4+4 ohm. Output voltage drop a little as my inductor is wound with 0.5mm2 wire...
