Hi Charlie,
I've always been under the assumption that the power supply only saw the amp as a load, not the speaker. (as long as the amplifier could handle the speaker's load the supply didn't care). I thought that the power supply's voltage and current potential would be more or less static (54v x 4.13A) + (54v x 4.13A) therefor it wouldn't matter to the power supply what the speaker's impedance was.
I woke up this morning thinking about the formula that you posted:
Irms=22/2.4=9.16A (rms)
Ipk=1.41*Irms
Ipk=12.8A
Wouldn't this formula only apply to the amplifier which can provide over 22A if the power supply is capable) ?
Shouldn't the Abletec have a static voltage output of 54v+54v and provide whatever current the amp draws up to it's 4.13A max (approx 450W)?
I will be picking up a pair of these supplies in the coming weeks, once I get mine I will be able to give them a full test using a variable DC load to see what their continuous power is, as well as some transients.
Until then it does seem rather strange that you can only get around 200W out of it. Perhaps a larger input capacitor on the amp would help absorb some of the current spikes?
Until then it does seem rather strange that you can only get around 200W out of it. Perhaps a larger input capacitor on the amp would help absorb some of the current spikes?
Yes, exactly. Think of it this way:
The load is drawing some amount of current that depends on the voltage imposed across the load by the amplifier, which in turn is doing so based on the input voltage. Ignoring any consumption by the amplifier, transistor voltage drop, etc. you can think of the amplifier as a voltage controlled "current valve". The amplifier, and the PS (which itself is very much like an amplifier) can only safely deliver as much current into the load as the output devices can handle or you will let out the 'magic smoke'. If the demands of the load exceed the capacity of either the amp or the supply you risk damage.
In this case, the power supply is reaching its rated current capability. The amplifier wants to maintain the desired voltage across the load but to do so it needs to supply more current than the PS can (safely) deliver. The amplifier will dutifully attempt to draw the necessary current, but the PS (I assume) has built in overcurrent protection and when the current becomes high enough to present a danger it shuts off. At least this is my assumption about what is happening.
You mentioned that your driver can cause the PS to shut down when you attempt to deliver a modest amount of power to the driver at 20Hz. In this case I assume that the driver's reactive load is actually drawing more current at this frequency than you would expect for a resistive load so even though the apparent power is low, the current draw may be quite high and the PS again shuts down. It all depends on the phase angle.
I can't say for certain what is going on in your system but under some assumptions I am making based on your info it seems to me that you are reaching the current limit of the SMPS.
For your linear supply, when the current draw becomes large the caps are drained to the extent that the rail voltages sag under the load. This is the nature of the design, but you probably don't notice it. On the other hand, an SMPS attempts to keep its output voltage(s) stable much like a regulated supply, so current limiting becomes important for protecting the circuitry.
Siegfried Linkwitz has a spreadsheet where you can calculate the voltage and current limited power output of an amplifier into a driver in a closed box (given the TS parameters, etc.). You can find it here:
http://www.linkwitzlab.com/closed-box1.xls
This can shed some light on these kind of problems.
Last edited:
FWIW, you could move up to one of these:
1000 Watts 110V Input Switching Power Supply Audio Power Amplifier Board | eBay
This has approximately the same rail voltages but has MUCH higher current capability - enough to supply the amplifier at its full rated power into a low impedance load.
1000 Watts 110V Input Switching Power Supply Audio Power Amplifier Board | eBay
This has approximately the same rail voltages but has MUCH higher current capability - enough to supply the amplifier at its full rated power into a low impedance load.
lol im probably just displaying the fact that how ignorant i am, but if it provides a DC volatage and enough current why wouldnt it be used with a class A amp like pass F6?!
I don't think that adding a large cap to each rail will help much. In this case we are doing continuous sine wave testing. This is a much more demanding signal compared to music signals.
There is nothing wrong with the PS. It seems to me that it is working as advertised. The user did not realize the implications of the PS specifications despite the advertised 400W/900W rating. Because of the impedance of the load, the current limit is reached before the voltage limit is reached.
Let's talk about your variable resistance test load, at least a hypothetical one that can adsorb any amount of power and can vary over any range of resistance.
Case 1: infinite resistance
If you start out with an infinitely high resistance and connect the PS to that across both rails what do you have in terms of power consumption? Being a resistor we can use Ohm's Law and calculate power as I*V (current times voltage). The voltage from rail to rail is 2*53V or 106V. What's the current? Re-arranging we have I=V/R, but since R is infinite I is zero. What's the power consumed? Well, the IV product works out to ZERO, that is no power is consumed.
Case 2: a load that consumes the rated current for the given rail voltages.
Again we have V=106V across the rails, however, this is the peak voltage and rms voltage is 0.707 times this, or 75Vrms. What load do we need to draw the rated load current of 4.13 Ampere? We calculate R=Vrms/Irms, and this comes out to R=75/4.18, or about R=18 Ohms. Lower resistance loads will draw more than the current limit of the PS when connected across the rails. In the case of a loudspeaker, the load appears on each rail, so for a driver the effective "resistance" would be half of this, or 9 ohms because the amp can only swing to one rail of the other, so the imposed voltage on the driver is not 106V but 53V, the rail voltage.
So we know that a load of infinite resistance draws no power, and a load of about 9 ohms draws the maximum power. Between 9 ohms and infinity, the amount of power that can be delivered to the load decreases as you increase the load resistance. This is no fault of the PS - you are simply not drawing the max rated current when you use the full rail voltage. Conversely, for loads below 9 ohms the amount of current that the load demands keeps increasing as the load resistance drops. For a given load, there will be some maximum voltage that will be reached when the maximum rated current is reached. This is the maximum power that can be delivered into the load while remaining within the design limits of the PS and it will be LESS than the product of the rail voltages and the max average rated current (2*53*4.13=438 Watts). The lower the load resistance, the lower the maximum power that can be sourced from the PS. It's current limited.
The PS does have some "current headroom" for brief peaks at high(er) frequencies but with sine wave testing you are placing a continuous demand on the PS and so the continuous ratings should be used.
Getting back to where this post started, if you have a dynamic load (e.g. with music signals having a non-insignificant crest factor) you could try to even out the peak demand with a large cap on each rail between the PS and the amp. The cap will adsorb (or supply if you like) transient power demands. The average power demand must still fit under the spec for average current of the PS, however, since the time averaged current supply cannot exceed that level.
Last edited:
The F5 Turbo requires about 2 amps of current steady (regardless if music is played) per channel. It also can't run any higher than 32v. So Abletec has too much voltage and not enough steady state current (1.38A max steady state). It won't work. You are better off with a pair of 24v 5A SMPS brick for LED lights which can power both channels of the original F5 (+/-24v rails) just fine.
A class A amplifier needs a power supply with VERY quiet rails, that is with very low noise and ripple on the rails. I would not use this PS unless you add some CRC filtering in between the PS and amplifier to remove the unwanted, non-DC junk. With that circuitry added, it might be fine in that application.
Thanks, I'll have a look at the spreadsheet today.
Originally I was simply sending a 25Hz test tones into my sub. I wasn't planning on putting too much voltage into my sub. I just wanted to see if the power supply was stable under a significant load, but it gave out before the sub really started to do any work. That's when I built the 2.4 Ohm dummy load to test the voltage out of the amp. I swithced to a 60Hz tone and I was thinking that the amp should generate at least 30v from that that power supply, but not even close
My linear supply only dropped 5v (from 42v to 37v per rail) when running the 60Hz test tone and 30v into the 2.4-ohm dummy load. The smps shuts off at about 60W with the 60Hz tone and 200W with the 120Hz tone.
I bought the SMPS's hoping to replace the large linear supplies in my 2 subwoofers, just to shed a few pounds.
Guys, please forget everything that I posted earlier with regards to measurements. There's something strange going on with the JLAmp / SMPS combo.
I connected the L15 Pro amp to the smps and it put 38v into the 4Ω load and 30v into 2.4Ω and the power supply never shut down... Apart from the horrible vibrating noise that the SMPS makes under load it seems fine (It sounds like an electric razor).
Sorry for the false alarm and thanks for your input.
I connected the L15 Pro amp to the smps and it put 38v into the 4Ω load and 30v into 2.4Ω and the power supply never shut down... Apart from the horrible vibrating noise that the SMPS makes under load it seems fine (It sounds like an electric razor).
Sorry for the false alarm and thanks for your input.
Hmmm, interesting. Well, glad to hear that the L15 amp is working better for you.
You are still only getting 350W or so. That seems about right, considering the loads used.
Well, nether the amp, nor the power supply gave. I just stopped turning it up. The amp is rated at about that into 2ohms and I was reaching the limitations of my resistor load as well at four ohms.
I see, so in theory someone could use one of these switching power supplies by first going with a more powerful model that can provide the necessay current and secondly buy putting one hell of the capacitor bank between the power supply and each channel amplifier circuit, right?
Mine makes a racket under heavy load too. Try playing big woofer kick drums, the SMPS clicks with each beat. I think it is piezoelectric effect on ceramic caps, and magnetorestriction on the transformer plates. They mechanically move under high voltage or current.
I don't think the SMPS likes to charge caps - that may cause it to shutdown.
I don't think the SMPS likes to charge caps - that may cause it to shutdown.
Not a capacitor bank (meaning only capacitors). A CRC includes a low pass filter. The idea is that you use rather large capacitors and a rather small resistor but values chosen so that the corner frequency is, for instance 15Hz. Actually, if connected to an SMPS you don't need the first capacitor, and can simple employ the remaining "RC". For instance using 0.22 Ohms and 47000uF gives you that 15Hz corner frequency. Good for filtering any mains (9dB rejection) and higher frequency noise. If you are not familiar with a CRC filter just Google it.
You don't need a massive supply, just one that will do the job and have the necessary amount of rail voltage and current capability.
alright. i think i got it. i used this and now what you are saying makes perfect sense to me:
(Sample)RC Low-pass Filter Design Tool - Result -
thanks again. just a regular Nichion will do?
You could also just use the 53V SMPS rail and linear regulate down to 48V or so to get a nice clean rail, provided you can find some linear regulators to take the current. Basically just CRC if you put a capacitor on the linear regulator output (as you should), but with active "R" in the middle.
Charlie, I know all about Ohms law and how power flows through various purely resistive loads. However, do not forget that many of these amplifiers also act as pseudo DC-DC converters. What with the speaker coil being an inductor and class D working on switching and all. So you can end up with higher voltages/currents (not both at the same time) than your supply could theoretically produce on its own. The exact numbers are very application specific however, as is noted by Darp's experience with one amp working fine, and the other barely producing power before cutoff.
Also, you do not need to take sqrt(2) of a DC supply to find RMS voltage like you do with AC, unless you want the RMS of the Ripple+Noise. And, sine wave signal from an amplifier is not a continuous load, it is a sinusoidal load, the length of the peaks is dependent on frequency.
XRK: SMPS don't mind charging caps, as long as it doesn't have surge protection, if it does, you might need a resistor if the supply restarting a few times while it builds charge is unacceptable, an NTC can also help with this
Charlie, I know all about Ohms law and how power flows through various purely resistive loads. However, do not forget that many of these amplifiers also act as pseudo DC-DC converters. What with the speaker coil being an inductor and class D working on switching and all. So you can end up with higher voltages/currents (not both at the same time) than your supply could theoretically produce on its own. The exact numbers are very application specific however, as is noted by Darp's experience with one amp working fine, and the other barely producing power before cutoff.
Also, you do not need to take sqrt(2) of a DC supply to find RMS voltage like you do with AC, unless you want the RMS of the Ripple+Noise. And, sine wave signal from an amplifier is not a continuous load, it is a sinusoidal load, the length of the peaks is dependent on frequency.
XRK: SMPS don't mind charging caps, as long as it doesn't have surge protection, if it does, you might need a resistor if the supply restarting a few times while it builds charge is unacceptable, an NTC can also help with this
Last edited:
D
Deleted member 148505
I think the JLAmp + SMPS combo suffers from bus pumping. Might be possible that the supply triggers overvoltage protect. Bus pumping is worse on lower frequencies.
I also have this supply on me for months now but I haven't tried it though. Will perform tests on the combo once I set it up.
That being the case, perhaps bridging IRS2092 boards (though overkill) will be required if we want to use this PS for subwoofer duty. I was planning to use one PS and amplifier per sub anyway. Paralleling two SMPS and bridging 2 amps should yield close to a kW into a 4 or 8-ohm load and solve the bus pumping issue
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.