Hi!
According to the name of topic, here is my first Class D amp just for fun:
http://www.hszk.bme.hu/~sp215/elektro/PWMideas/switching_amp_by_Pafi.gif
And a later design, wich is almost usable:
http://www.hszk.bme.hu/~sp215/elektro/PWMideas/Atti's_PWM_Amp_(or_sthing_like_that).gif
I build an active subwoofer with this. I'm not absolutely shure if it was the exact schematic, but more or less.
My latest will be available soon, if you ask for. This is 500W/4ohm full band, with 0,3% THD at 800Hz, full power, but sadly much more THD at 5 kHz, and I don't understand why! I hope sbody will have an idea! It's self oscillating.
This is how it looks: http://www.hszk.bme.hu/~sp215/PWM_500W/311.jpg
http://www.hszk.bme.hu/~sp215/PWM_500W/316.jpg
And some waveforms:
http://www.hszk.bme.hu/~sp215/elektro/500W_measure.htm
Measured with 1/10 divider, and insufficient power supply.
According to the name of topic, here is my first Class D amp just for fun:
http://www.hszk.bme.hu/~sp215/elektro/PWMideas/switching_amp_by_Pafi.gif
And a later design, wich is almost usable:
http://www.hszk.bme.hu/~sp215/elektro/PWMideas/Atti's_PWM_Amp_(or_sthing_like_that).gif
I build an active subwoofer with this. I'm not absolutely shure if it was the exact schematic, but more or less.
My latest will be available soon, if you ask for. This is 500W/4ohm full band, with 0,3% THD at 800Hz, full power, but sadly much more THD at 5 kHz, and I don't understand why! I hope sbody will have an idea! It's self oscillating.
This is how it looks: http://www.hszk.bme.hu/~sp215/PWM_500W/311.jpg
http://www.hszk.bme.hu/~sp215/PWM_500W/316.jpg
And some waveforms:
http://www.hszk.bme.hu/~sp215/elektro/500W_measure.htm
Measured with 1/10 divider, and insufficient power supply.
Hi!
"Maybe you could show us the schematic. "
Just a minute!
The actual version's schemacic is allways in pieces, but here is the previous. This is almost the same, just has different current-sensor, and output-filter:
http://www.hszk.bme.hu/~sp215/PWM_500W/PWM23.gif
There is a ground plane on bottom layer, but only under the modulator, because if it was overlap the power-stage, it could conduct the impulses to the comparator! The low-power and hi-power parts connected in only one point.
I thought too that the feedback is insufficient, but... I will describe later!
"Maybe you could show us the schematic. "
Just a minute!
The actual version's schemacic is allways in pieces, but here is the previous. This is almost the same, just has different current-sensor, and output-filter:
http://www.hszk.bme.hu/~sp215/PWM_500W/PWM23.gif
There is a ground plane on bottom layer, but only under the modulator, because if it was overlap the power-stage, it could conduct the impulses to the comparator! The low-power and hi-power parts connected in only one point.
I thought too that the feedback is insufficient, but... I will describe later!
Since the two integrators have a 90 deg phase-shift across a wide frequency band and the feedback of the amp is lowpass-filtered, I assume that the switching frequency is quite low. Am I right ?
Furthermore I'd like to know why you use such a complicated topology fo the lower half of the output stage ?
Regards
Charles
Furthermore I'd like to know why you use such a complicated topology fo the lower half of the output stage ?
Regards
Charles
Hi!
It oscillate at 150 kHz with these values, but after this I incresed it to 210 kHz. Yes, it's not too high indeed.
(But the 2nd integrator only take effect under 8 kHz. Hmmm. This could be higher!)
"complicated topology"? Most of these parts (74HC00, MOSFETs, npn bjt-s) belongs to the overcurrent-, overvoltage-protection, and auto-reset! The lower half is as complicated as upper half!
It oscillate at 150 kHz with these values, but after this I incresed it to 210 kHz. Yes, it's not too high indeed.
(But the 2nd integrator only take effect under 8 kHz. Hmmm. This could be higher!)
"complicated topology"? Most of these parts (74HC00, MOSFETs, npn bjt-s) belongs to the overcurrent-, overvoltage-protection, and auto-reset! The lower half is as complicated as upper half!
hi.
>http://www.hszk.bme.hu/~sp215/PWM_500W/PWM23.gif>
your link doesnt seem to work , could you post it here?
rgds - karsten madsen - cadaudio.dk
>http://www.hszk.bme.hu/~sp215/PWM_500W/PWM23.gif>
your link doesnt seem to work , could you post it here?
rgds - karsten madsen - cadaudio.dk
Ok, it cleared my mind (I think).though I cant get the simulator www.transim.com/champion/ any one got it? Can I use the tiangle wave generator for more amp´s or do I have to build one for each amp?
Hi, sfx !
I plan to built same amp, but full range (up 25kHz). There is some minor bugs in your drawing, i.e. diodes in flying supplies of IR2112. Anyway, concept is great! For higher frequency responce I will use IR2011 and much faster MOSFETs. Is it good idea use very fast MOSFETs like IR3715Z? They only 20V/15mOhm and could be located in VRM on some motherboards (switching at 300-400kHz).
Any suggestion about opamps and comparators? OPA4134? LF357? LM339 (is it doubles LM319 or it much slowly)?
About Shottky diodes: 1N5822 - is it enought for 16V supply/4 Ohm load or I must looking for 5A devices?
Pay attention, that I live in Ukraine and haven`t access to latest IC. So there is no HIP* or ISL* drivers, no fast LT* comparators :-(
Btw, can anyone suggest radiocomponents dealer in London?
I plan to built same amp, but full range (up 25kHz). There is some minor bugs in your drawing, i.e. diodes in flying supplies of IR2112. Anyway, concept is great! For higher frequency responce I will use IR2011 and much faster MOSFETs. Is it good idea use very fast MOSFETs like IR3715Z? They only 20V/15mOhm and could be located in VRM on some motherboards (switching at 300-400kHz).
Any suggestion about opamps and comparators? OPA4134? LF357? LM339 (is it doubles LM319 or it much slowly)?
About Shottky diodes: 1N5822 - is it enought for 16V supply/4 Ohm load or I must looking for 5A devices?
Pay attention, that I live in Ukraine and haven`t access to latest IC. So there is no HIP* or ISL* drivers, no fast LT* comparators :-(
Btw, can anyone suggest radiocomponents dealer in London?
Hi,
thanks for pointing out the errors. I really should update the schematic. I also think it might be better to use a center-tapped coil instead of the two seperate coils in the output filter of the half-bridges.
In general, for class D, faster is better, especially for full range amps. In case of the MOSFETs, slow ones induce high losses, since they will be no fully switched on while rising for a considerable time. For the triangle wave generator, I got better results using a function generator IC.
Let me know if you get it to work - I had some trouble with mine, which eventually killed it. Hadn't had time to fix it yet... and the next project will probably be a UcD.
Good luck!
thanks for pointing out the errors. I really should update the schematic. I also think it might be better to use a center-tapped coil instead of the two seperate coils in the output filter of the half-bridges.
In general, for class D, faster is better, especially for full range amps. In case of the MOSFETs, slow ones induce high losses, since they will be no fully switched on while rising for a considerable time. For the triangle wave generator, I got better results using a function generator IC.
Let me know if you get it to work - I had some trouble with mine, which eventually killed it. Hadn't had time to fix it yet... and the next project will probably be a UcD.
Good luck!
sfx is it possible to do feedback in bridged BCA topology and take feedback signal BEFORE output inductors?
What THD you have achieve with your amp? How about frequency responce?
About dual windings on same core- can you draw it? I think, it possible, but needs very symmetrical winding. And I don`t get about winding direction- are they synphase or not?
What THD you have achieve with your amp? How about frequency responce?
About dual windings on same core- can you draw it? I think, it possible, but needs very symmetrical winding. And I don`t get about winding direction- are they synphase or not?
hey congaratulations on building your amplifier i am actually trying to build one myself but i was just wondering if you used a comparator for you ur pulse width modulation of if you did why is the convention used a traingular wave why not a saw tooth or something else and if you dont mind posting a report on what you did that would be very helpful as soon as mine is fully built ill do the same
Hi Kuzmenko,
I don't see a way to get feedback in BCA easily before the output inductor. The inductor is not just a lowpass filter, but also "calculates the sum" between the high and low side (the center tap has zero potential when they are equal pulse width). Maybe you can generate the feedback signal with a summing amp, connecting the inputs to either side between the mosfet and the inductor. I don't think it'll work, though, since there are phases where both mosfets are switched off, which might result in a strange feedback signal.
With feedback after the inductor, a clocked design might run into oscillations - but it should be possible to some extent. A friend of mine is currently simulating a self-oscillating BCA amp, but it's not quite there yet. I also haven't completely understood how it works...
For the center tapped inductor, both windings have to be in the same direction. If you wind them in opposite directions, the magnetic flux will cancel, and your inductor will have no effect (=no resistance for high frequencies => MOSFETs will burn). Imagine a single winding with a center tap. Maybe you can make a single winding in two layers on the toroid, so that after half the turns, you reach the start again, and then solder a tap exactly there, in the middle. Or you can wind two identical windings, and connect them in series (watch the winding direction, has to be the same).
My subwoofer amp had a frequency response from 15 - 5000 Hz into a resistor. Into my speaker, the high frequency response went higher, almost up to 20.000 Hz (haven't really understood why...). I didn't measure THD - since I didn't use feedback, and my triangle wave looked rather bad (drawback of veroboard without ground plane), it wasn't that great. It sounded "not too bad" even into a full range speaker, so I guess it was around 1% or so. It was enough for my purposes, since as most subwoofers, the speaker had a distortion well beyond 1%, and the room response was even worse - so the amplifier didn't matter that much. I just needed power
Elisabeth,
triangle waves are usually preferred over sawtooth, because they are easier to produce, and the PWM signal is centered, or symmetric, which has less distortion. In the case of BCA, with two centered PWM on the output stage, you'll also get an effective doubling of the switching frequency, since the output is the difference between the two PWM signals. The MOSFETs can switch with only 300kHz, while the output filter sees 600kHz and thus can reduce ripple very effectively.
Cheers,
Felix
I don't see a way to get feedback in BCA easily before the output inductor. The inductor is not just a lowpass filter, but also "calculates the sum" between the high and low side (the center tap has zero potential when they are equal pulse width). Maybe you can generate the feedback signal with a summing amp, connecting the inputs to either side between the mosfet and the inductor. I don't think it'll work, though, since there are phases where both mosfets are switched off, which might result in a strange feedback signal.
With feedback after the inductor, a clocked design might run into oscillations - but it should be possible to some extent. A friend of mine is currently simulating a self-oscillating BCA amp, but it's not quite there yet. I also haven't completely understood how it works...
For the center tapped inductor, both windings have to be in the same direction. If you wind them in opposite directions, the magnetic flux will cancel, and your inductor will have no effect (=no resistance for high frequencies => MOSFETs will burn). Imagine a single winding with a center tap. Maybe you can make a single winding in two layers on the toroid, so that after half the turns, you reach the start again, and then solder a tap exactly there, in the middle. Or you can wind two identical windings, and connect them in series (watch the winding direction, has to be the same).
My subwoofer amp had a frequency response from 15 - 5000 Hz into a resistor. Into my speaker, the high frequency response went higher, almost up to 20.000 Hz (haven't really understood why...). I didn't measure THD - since I didn't use feedback, and my triangle wave looked rather bad (drawback of veroboard without ground plane), it wasn't that great. It sounded "not too bad" even into a full range speaker, so I guess it was around 1% or so. It was enough for my purposes, since as most subwoofers, the speaker had a distortion well beyond 1%, and the room response was even worse - so the amplifier didn't matter that much. I just needed power
Elisabeth,
triangle waves are usually preferred over sawtooth, because they are easier to produce, and the PWM signal is centered, or symmetric, which has less distortion. In the case of BCA, with two centered PWM on the output stage, you'll also get an effective doubling of the switching frequency, since the output is the difference between the two PWM signals. The MOSFETs can switch with only 300kHz, while the output filter sees 600kHz and thus can reduce ripple very effectively.
Cheers,
Felix
Hello all,
I made my own class-d amp from those schematic what I've found here.
But I have a problem with it, when the output reaches the power supply (35V in my case), it starts to crackle.
I can't figure out where is the problem.
I tried with and without feedback, the problem remained.
When it starts to crackle it seems that both mostfets are conducting. (I cant view it because my scope triggering is started to go wrong)
I think dead time is enough, the mostfets stayed cool after 2 hour play at low output power.
I tested it with sine wave, below few kHz, everything is OK, it is clipping like a class AB amp, but at 5khz (and above) it doesn't clipping as good as at lower frequency.
For short moments the output to disappears.
The triangle wave is about 250-290kHz (I tried between 150kHz and 350khz, so the problem is not there) ~5.6Vpp.
I tried different mostfets ( STW34NB20, IRF530 ), different level shifter (6N137), different values for the error amp and for feedback too.
I also tried with bigger DT but no change at all.
I changed the TL072 to LM833, tried another LM319 and IR2110 but they were not defective.
The output inductor seems to be good, it was barely warm (the load was 2ohm, and the the output power was just below crackle)
One more question: what values would be good for R4,R13,R16,C6,C7 ? Later I want to use it with higher supply, about +/-65V.
Any suggestions would be really great.
Sorry my english is poor...
Regards
fehrudi
I made my own class-d amp from those schematic what I've found here.
But I have a problem with it, when the output reaches the power supply (35V in my case), it starts to crackle.
I can't figure out where is the problem.
I tried with and without feedback, the problem remained.
When it starts to crackle it seems that both mostfets are conducting. (I cant view it because my scope triggering is started to go wrong)
I think dead time is enough, the mostfets stayed cool after 2 hour play at low output power.
I tested it with sine wave, below few kHz, everything is OK, it is clipping like a class AB amp, but at 5khz (and above) it doesn't clipping as good as at lower frequency.
For short moments the output to disappears.
The triangle wave is about 250-290kHz (I tried between 150kHz and 350khz, so the problem is not there) ~5.6Vpp.
I tried different mostfets ( STW34NB20, IRF530 ), different level shifter (6N137), different values for the error amp and for feedback too.
I also tried with bigger DT but no change at all.
I changed the TL072 to LM833, tried another LM319 and IR2110 but they were not defective.
The output inductor seems to be good, it was barely warm (the load was 2ohm, and the the output power was just below crackle)
One more question: what values would be good for R4,R13,R16,C6,C7 ? Later I want to use it with higher supply, about +/-65V.
Any suggestions would be really great.
Sorry my english is poor...
Regards
fehrudi
Perhaps your bootstrap capacitor is not large enough and when you clip (hence no switching during that time) it discharges and stops driving the upper side mosfet.
But that effect should be more noticeable with lower frequencies (as the clip time is longer), and as you say your case is contrary...
Have you tested for weak connections?
But that effect should be more noticeable with lower frequencies (as the clip time is longer), and as you say your case is contrary...
Have you tested for weak connections?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.