Hi 81bas
Good point, don´t know why I have not remembered this from some of the early designs of UcD, where this kind of resistor was included
But I think it woould be more effecient going to the negative rail (the resistor that is
), as this will create a charging path no matter the "direction" of the first pulses. ..... but very god point Have been playing with the feedback values to get a higher gain (first walues gives some 10 times .....
)But bread board and class-d is not good for performance .... which should be evident ... so raising the gain also raises the noise a lot.
This once again makes me wonder whether single sided board, and through hole components can actually be made to work sufficiently well for any measure, and whethe it is worth the effort.
Tekko good work ... and congratulatios again
Tekko, how is the noise leve of your amp, honestly, and compared to an ok commercial AB amp as an example??
Tekko could you post some more scope pictures?
- 1kHz half output or so
- Switching before the filter
- input to IR2110 ligh and low
There is a slight hiss, kindof like when an amp it turned to max volume with nothing connected to the inputs.
At startup it farts a few times.
1V is not enough to bring it up to clipping. And im not using IR2110, im using IR2010. I have no deadtime except for 22ohm gate resistors and !n4148's in parallel for instant gate discharge, heatsink stays nice and cool on 2x 15V input.
I'll take some scope shots once i find my dummy load.
At startup it farts a few times.
1V is not enough to bring it up to clipping. And im not using IR2110, im using IR2010. I have no deadtime except for 22ohm gate resistors and !n4148's in parallel for instant gate discharge, heatsink stays nice and cool on 2x 15V input.
I'll take some scope shots once i find my dummy load.
81bas, I have now tried the little trick of a resistor on the output .... must say that connected to the positive rail works, and connected to the negetive, does not produce any improvements! ..... have to think a bit more about this
But still it is not enough to make it start everytime with the big cap in the feedback loop .... guess it does need the DC gain, when strating up .... this is bummer for when it works (with the cap) there is virtually no DC offset on the output ..... so how to make this work, and eliminate other offset reducing things
But still it is not enough to make it start everytime with the big cap in the feedback loop .... guess it does need the DC gain, when strating up .... this is bummer for when it works (with the cap) there is virtually no DC offset on the output ..... so how to make this work, and eliminate other offset reducing things
I modified the circuit alittle by adding a phase lead network like in the putzeys article.
scope shots.
Output from mosfets:
1kHz full volume, those artifacts were the function generator software acting up:
And finally 20kHz, its fuzzy because i forgot to turn down the volume on a video stream, but the effect of low switching frequency is clearly visible here:
After a minute of that the amp reeked of burnt transistor package, something got hot at 20kHz but it wasent the heatsink. For what it is, it sounds incredibly good.
scope shots.
Output from mosfets:
1kHz full volume, those artifacts were the function generator software acting up:
And finally 20kHz, its fuzzy because i forgot to turn down the volume on a video stream, but the effect of low switching frequency is clearly visible here:
After a minute of that the amp reeked of burnt transistor package, something got hot at 20kHz but it wasent the heatsink. For what it is, it sounds incredibly good.
Last edited:
I've seen the IRS2092 and i may get some of those sometime in the future.
Anyways i've removed all of the hiss, now its only the noise from the computer soundcard that manifests itself as a faint hiss, with the computer turned off, the amp is dead silent. How ?
I put a 1nF cap on the audio input. The switching frequency is now at 6µS.
Anyways i've removed all of the hiss, now its only the noise from the computer soundcard that manifests itself as a faint hiss, with the computer turned off, the amp is dead silent. How ?
I put a 1nF cap on the audio input. The switching frequency is now at 6µS.
The amp is now off the 2x15v regulated psu. Currently it runs off a 2x12vac transformer and its way cleaner now than it was on the stabilized psu. Despite no stabilization of nthe rails, there is no hup and the amp starts silently with a very faint "tot" noise.
Heres a shot of the carrier resudial:
And did i mention theres no heating what so ever of the mosfets, only the gate driver get alittle hot but i guess that is normal.
And sound quality ? BRILLIANT!!
Heres a shot of the carrier resudial:
And did i mention theres no heating what so ever of the mosfets, only the gate driver get alittle hot but i guess that is normal.
And sound quality ? BRILLIANT!!
Olá,
I take the aplication note and reference design at www.irf.com
The IRS2092 is full of security systems.
There is a need to heed well this protection system to develop a design that works well.
I suffered a lot but I got a spectacular result.
I can help in every possible doubt.
Bye
Captain Carlos
I take the aplication note and reference design at www.irf.com
The IRS2092 is full of security systems.
There is a need to heed well this protection system to develop a design that works well.
I suffered a lot but I got a spectacular result.
I can help in every possible doubt.
Bye
Captain Carlos