Good, I had a good experience with this ucd, the sound is good at low volume, but when increasing the volume it distorts the sound is inaudible too, since I reviewed everything and do not think the problem, using esou irf 540N and 50 +50 volts. Anyone ever had this problem? Thanks a lot!
Scale able class D 1250W UCD style and Non-UCD style - PCB Ready. Tested.
- Thread starter norazmi
- Start date
pergo
New member
i was thinking at the LT1016 or MAX1016.
It work at maximum +-7V (typical +-5V), has 2 output (in phase, and out of phase), good PSRR, propagation delay of 10ns.
It need maximum care in the circuit layout if you want very high switching speed, but it can do it pretty well.
http://cds.linear.com/docs/Datasheet/1016fc.pdf
Price is 6$ but is one of the best comparator in the market.
The 0-180° output semplify the level shifters for IR2010. --> less components
So, it will be simpler but better.
Comparator is critical in the UCD-style performances.
It work at maximum +-7V (typical +-5V), has 2 output (in phase, and out of phase), good PSRR, propagation delay of 10ns.
It need maximum care in the circuit layout if you want very high switching speed, but it can do it pretty well.
http://cds.linear.com/docs/Datasheet/1016fc.pdf
Price is 6$ but is one of the best comparator in the market.
The 0-180° output semplify the level shifters for IR2010. --> less components
So, it will be simpler but better.
Comparator is critical in the UCD-style performances.
Yep i have with me couple LT1016 and others, i`ve order it last time with small value to make test, but i dont finish yet, maybe im forgot to try it. Well i can do it again with this comparator, the problem with LM311 is produce lots of noise coming at output and its lots different between non-ucd style by using TL074 and ir2110 almost no noise coming out even u did ugly wiring inside the amp case. But fully discrete ucd style is good and less noise coming out, easy part and cheap driver but with many components 
.
.
pergo
New member
HF prefilter is a "must" in this application circuit. It help to provide a ground node for the noise.
May be not the solution, but is a "good advice" in this kind of project.
I use it also in classAB amplifier.
You see noise at the comparator output when the audio signal is idle? (but the oscillation is started...so 50% duty, without modulation..so you can see the squarewave clearly).
What about LM311 supply's hard decoupling? I mean: bypass near power supply pin + power supply highly filtered (for example, by a capacitance multiplier + zener reference)
May be not the solution, but is a "good advice" in this kind of project.
I use it also in classAB amplifier.
You see noise at the comparator output when the audio signal is idle? (but the oscillation is started...so 50% duty, without modulation..so you can see the squarewave clearly).
What about LM311 supply's hard decoupling? I mean: bypass near power supply pin + power supply highly filtered (for example, by a capacitance multiplier + zener reference)
idk, but so far i`ve build simplest class D amp using lm311 always will face noise problem, but if we drive woofer using low pass there is no problem. But for full range there will be lots of adjustment needed to minimize the noise. Yet so far i did not face noise problem with TL074 non-ucd and discrete ucd amp, they all passed. its not about LM311 supply as i already did separate it with external supply so i can see the different but its same LM311 still generate noise. i can pass the noise problem but there will be little modification at bal/strobe , input section etc... hahahaha i think better try with LT1016 or other comparator.
i`ll do another experiment with new comparator, wanna see the different, i`ll look into my inventory which one i have, i think i have LT1011 and LT1016, let see. But i still like more with non-ucd style TL074 + ir2110, silent, powerfull and efficient. But no one did the totem pole driver yet with IR2110 hmmmm...
pergo
New member
for the LT1016, try that:
http://s9.postimage.org/g6c77dvq7/Immagine.jpg
p.s: in the TL074+IR schematic, you take feedback at the mosfet output (squarewave), not at the audio-output. Is it correct?! Seems not.
http://s9.postimage.org/g6c77dvq7/Immagine.jpg
p.s: in the TL074+IR schematic, you take feedback at the mosfet output (squarewave), not at the audio-output. Is it correct?! Seems not.
ucd style, the TL074 is working good already. But ucd still got problem with noise, and ir become more worst when u put 2 modules inside case. so need to find solution for it. for the non-ucd the FB is taken before the output filter, its working exactly like irs2092 chips but irs2092 chips have more build in such as protection and non-ucd is not.
but i receive 2 pcs class H modules, i need to prepare to get this monster work hehe i think i will start new thread with it, then i will continue ucd mods to get better performance.
regards.
but i receive 2 pcs class H modules, i need to prepare to get this monster work hehe
i think i will start new thread with it, then i will continue ucd mods to get better performance.regards.
pergo
New member
maybe the layout is the problem.
There are two fundamental "ground plane" to make
The first is the signal/output ground (GND): you've to use a plane to connect comparator, input signal, and level shifter. Al least, the output capacitor.
Then, the second, there's is a "power ground" that is "ground" for the power stage but not a real ground (i mean, 0 volt): it's -HV. You've to connect the IR2010, the phase shifter pull-down, the output mosfet
Final: hard decoupling. +HV to -HV, +HV to GND, -HV to GND with electrolitic+film, and the op-amp voltages
Can you post some waveforms?
I need:
- pwm before LC filter (if you can, at 50% duty, so no signal)
- output signal after the LC.
If your noise goes up when you put more IR modules in the same chassis, i think in a layout problem so your pcb is a little antenna.
There are two fundamental "ground plane" to make
The first is the signal/output ground (GND): you've to use a plane to connect comparator, input signal, and level shifter. Al least, the output capacitor.
Then, the second, there's is a "power ground" that is "ground" for the power stage but not a real ground (i mean, 0 volt): it's -HV. You've to connect the IR2010, the phase shifter pull-down, the output mosfet
Final: hard decoupling. +HV to -HV, +HV to GND, -HV to GND with electrolitic+film, and the op-amp voltages
Can you post some waveforms?
I need:
- pwm before LC filter (if you can, at 50% duty, so no signal)
- output signal after the LC.
If your noise goes up when you put more IR modules in the same chassis, i think in a layout problem so your pcb is a little antenna.
pergo at the meantime i`ll hold this project for a while since i need to start with earlier project with 2kw class H amp that i have it at diysmps. I`m now prepared with component list to start with stereo module 2 x 2kwrms class H amp. This will be my first class H amp with high power version.
So this project class D will put on hold until i`m finish class H amp. Btw i would like to invite u and others to join with class H builder thread i`ll soon open it.
regards,
Azmi.
So this project class D will put on hold until i`m finish class H amp. Btw i would like to invite u and others to join with class H builder thread i`ll soon open it.
regards,
Azmi.