Class D and SMPS simulation
I’ve just ordered an Abletec power supply and, in the meantime, I’m investigating about the limits of this unit, or, saying better, the limits of a generic SMPS.
For this reason, I simulated with LTSpice an ideal half bridge class D amplifier, powered by a dual voltage supply of +/-53V (see attached picture).
In the second and third picture you can see the resulting waveform of the current supplied by the positive voltage source, for a load of 4 ohm and an output power of 200W.
In the second picture the signal frequency is 200 Hz, while in the third picture the frequency is 40 Hz; at both the frequencies, I simulated three different values of the supply capacitors, that is 1000uF (green line), 10000uF (red line) and 50000uF (blue line).
Suppose to operate at 40 Hz with a 10000uF supply capacitor: a 200W power peak (frequent condition, for example, if the amplifier is a bass guitar one or if you are using it to drive a low efficiency subwoofer ) will require a peak current of about 8 A, from the power supply.
Many of the general purpose SMPS have a fast internal current limit; if this is the case, you must be sure that this current limit is over 8 A; adding a little margin, you have to choose a 10A SMPS, that is a 1000 W nominal power one; the power supplied by this SMPS to the amplifier will be however a little more than 200W.
So, if the above reasoning is correct, the choices are:
- to use a SMPS with a nominal power many times higher then the power you want to send to the load;
- to use a SMPS of moderate power, but with a very large supply capacitor (50 – 100mF); in this case, you have to use also a soft start circuit between the SMPS output and the capacitor; moreover, the energy stored in the supply capacitor will be enough to melt down everything, in case of short circuit;
- to use a SMPS without current limiting, as I hope the Abletec unit is (in the data sheet they say: Short circuit protected: Primary limiter on aux outputs, Fuses on HV), but remember the Murphy’s Law: if you use a fuse to protect a transistor, the transistor will always blow itself, in order to protect the fuse.
Every comment will be appreciated.
P.S. 1: in my simulation, I have used ideal capacitors, with no ESR; I suspect things are worst with real capacitors.
P.S. 2: the strange green line waveform in the last picture (1000uF, 40Hz), may be caused by bus pumping (in simulation, the voltage across the supply capacitor reached values considerably higher than the voltage source).
I’ve just ordered an Abletec power supply and, in the meantime, I’m investigating about the limits of this unit, or, saying better, the limits of a generic SMPS.
For this reason, I simulated with LTSpice an ideal half bridge class D amplifier, powered by a dual voltage supply of +/-53V (see attached picture).
In the second and third picture you can see the resulting waveform of the current supplied by the positive voltage source, for a load of 4 ohm and an output power of 200W.
In the second picture the signal frequency is 200 Hz, while in the third picture the frequency is 40 Hz; at both the frequencies, I simulated three different values of the supply capacitors, that is 1000uF (green line), 10000uF (red line) and 50000uF (blue line).
Suppose to operate at 40 Hz with a 10000uF supply capacitor: a 200W power peak (frequent condition, for example, if the amplifier is a bass guitar one or if you are using it to drive a low efficiency subwoofer ) will require a peak current of about 8 A, from the power supply.
Many of the general purpose SMPS have a fast internal current limit; if this is the case, you must be sure that this current limit is over 8 A; adding a little margin, you have to choose a 10A SMPS, that is a 1000 W nominal power one; the power supplied by this SMPS to the amplifier will be however a little more than 200W.
So, if the above reasoning is correct, the choices are:
- to use a SMPS with a nominal power many times higher then the power you want to send to the load;
- to use a SMPS of moderate power, but with a very large supply capacitor (50 – 100mF); in this case, you have to use also a soft start circuit between the SMPS output and the capacitor; moreover, the energy stored in the supply capacitor will be enough to melt down everything, in case of short circuit;
- to use a SMPS without current limiting, as I hope the Abletec unit is (in the data sheet they say: Short circuit protected: Primary limiter on aux outputs, Fuses on HV), but remember the Murphy’s Law: if you use a fuse to protect a transistor, the transistor will always blow itself, in order to protect the fuse.
Every comment will be appreciated.
P.S. 1: in my simulation, I have used ideal capacitors, with no ESR; I suspect things are worst with real capacitors.
P.S. 2: the strange green line waveform in the last picture (1000uF, 40Hz), may be caused by bus pumping (in simulation, the voltage across the supply capacitor reached values considerably higher than the voltage source).
Can you tell me more about this mod? What does this do exactly? How is this influencing the output voltage?
How did you come to figure this out? Do you have the schematics and some knowledge of the design of the unit?
I tried to find the schematic of the Abletec SMPS, to no avail.
I figured out by means of measurement and common sense what the feedback netwerk and sensing part had to be. R442 is part of resistornetwork and does the sensing/regulation of the main rails. The feedback part is changed and so the outputvoltage.
I figured out by means of measurement and common sense what the feedback netwerk and sensing part had to be. R442 is part of resistornetwork and does the sensing/regulation of the main rails. The feedback part is changed and so the outputvoltage.
Thanks for this info Piersma. Do you know what the range of mains voltage is?
Sent from my SM-G900W8 using Tapatalk
I'm sorry, Wasn't paying attention to what I typed. I'm already running two of them. I'm talking about the output voltage
Sent from my SM-G900W8 using Tapatalk
I agree with you; in my previous post, I tried to point out the difference between an SMPS that is specifically designed to power one or more class D amplifiers (like Abletec) and a generic SPMS equipped with current limiter.
In the first case, the SPMS will be able to provide a current much higher than the nominal one for a short time (according to Abletec datasheet, it supplies a nominal current of 1.4A without time limitation, 4.13A current for 5’ and 16.50A current for 10ms).
A generic SMPS equipped with current limiter is not able to supply a current higher than its nominal current value; equipment of this kind are convenient if the loading is constant.
For example, a peak of 10A will be tollerated by the Abletec (or similar) SMPS, provided the peak has a limited duration; the same peak will probably cause the shut off a generic SMPS equipped with a 9A current limiter; the behaviour of the latter can be greatly improved increasing the supply capacitors.
I re-run my LTSpice simulation with 200 W delivered to a 4 ohm load, at 40 Hz frequency; I noticed that the voltage at the V+ rail periodically exceeds 53V, probably because bus pumping; the maximum value of the voltage across the supply capacitor in this simulation has been:
- 68V with a supply capacitor of 1000uF
- 53.6 with a supply capacitor of 10000uF
If this simulation is correct, it is advisable to put always a capacitor of not less than 10000uF on positive and negative rails.
If somebody cares:
- output fuses are T5AH250 (slow blow, 5A);
- to mate the board connectors, you can use the power connector for the old AT motherboard (you can find them in the harness of the old AT power supply).
For very heavy loads, I am considering the possibility to add an external heatsink to the existing aluminum bar, but to do this, I have to remove the bar itself, so I have to loosen the screws fixing the spring clips that press the power components; this operation is impossible because of the components on the board (mainly the electrolytics); so, what can I do?
Looking at one I have here, it seems you could just mount the PS "bar" directly to additional heatsinking. On my unit a piece of some kind of tape is wrapped around the bottom of the bar and up about 2mm on the "outside" surface. You would need to cut this off.
I should caution that I have received a shock when I touched the bar with the PS operating to see how hot it was. Seems like electrical isolation is not reliable. You might want to add some or keep the unit physically isolated from users.
I would like to fasten the heatsink to the bar using some screws, so I need to drill the bar, preferably after the bar has been separated from the board.
Concerning the elecrical shock, in my board the bar is electrically connected (0 ohm) to pin 5 and 6 of the mains connector; these pins should be connected to the mains earth, also for proper working of the EMI input filter.
What I would really like is to be able to adjust the 5.6V auxiliary output down to 5.0-5.2V. Then I could use it to power my small ARM boards directly. That would be REALLY USEFUL (for me and others, too).
Any chance that you have some sense of how to adjust that output???
Thanks for your reply Bullistang. I'm (temporarily) using a SB 5A, glass fuse in one of my supplies right now as well but the thing about fuses is you don't really know that they work until you run unto some sort of fault and it does it's what it's supposed to do... I'm wondering if there's some safety consideration that I should be concerned with using glass. I've read that ceramics were more reliable under high temperature operation. Also that they have a higher "breaking capacity" even if both are slow blow with the same ampere rating. I'm still a bit fuzzy on the difference between breaking capacity and current rating.
Last edited:
Hey CharlieLaub, Is the 7.5v, regulated down to 5.2v not providing enough current for your setup? I'm running the Pi2 off of mine and I haven't had any issues so far. Mind you, I'm not powering any other peripherals with my Pi and I've only tested Moode.
Last edited:
The voltage drop across a diode is current dependent. I need the 5V to power an ARM board and provide USB power without further regulation. A diode is not a good way to provide a steady, regulated voltage output in this application.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.