Power amplifier tests with respect to FTC: 16 CFR Part 432 (July 5, 2024) requirements on output power claims

[amirm]

BTW, it is funny enough that AP in their app notes show 50W resistors as a test load. Maybe Amir uses the same ones?

Don't be cute if you don't know the answer to something. I use 300 watt Ohmite 8 ohm resistors. Two are in parallel to create 4 ohm (so 600 watts). They weigh 700 grams or 1.5 pounds each. I have four for stereo testing. It is mounted in a rackmount case to provide heat sinking. This allows burst testing well above their rating (I think they can handle 10X their rating for short period).

They are cross checked against Dale which Benchmark and JA use: https://www.audiosciencereview.com/...sr-dummy-load-configuration.8026/#post-196116

 

[amirm]

If the amp shorts internally there’s nothing the load can do about it, and the load will see zero volts, or have I misunderstood your point?

Yes. A "short" in this parlance means the output stage in the amplifier has shorted out, dumping the full power supply into the dummy load. If the power supply is rated at say, 1,200 watts, that is what is going to go into the dummy load. The massive 300 watt resistors I use could get damaged if I let the condition continue but likely I can save them by pulling the plug. And at any rate, I am not worried about them blowing up given the massive armor around them.

 

[amirm]

If the amp under test goes open rail at e.g. 60V and the load starts heating up, that can be detected and a relay opened by the microcontroller, but it would need to be a big relay.

A relay won't be fast enough -- certainly not at these current levels. A fuse will be fast but have to make sure it doesn't self modulate and cause distortion.

Anyway, have you tested your smaller load with over current to see what the failure mode is for instantaneous fault? Kind of like that video I post.

 

[mcdn]

Yes. A "short" in this parlance means the output stage in the amplifier has shorted out, dumping the full power supply into the dummy load. If the power supply is rated at say, 1,200 watts, that is what is going to go into the dummy load. The massive 300 watt resistors I use could get damaged if I let the condition continue but likely I can save them by pulling the plug. And at any rate, I am not worried about them blowing up given the massive armor around them.

That’s what I thought you meant. This dummy load can sink 1200W for at least 5 minutes, 1000W for at least 15 minutes, and 800W indefinitely. The smell of the amp burning is going to be the bigger problem.

 

[amirm]

That’s what I thought you meant. This dummy load can sink 1200W for at least 5 minutes, 1000W for at least 15 minutes, and 800W indefinitely. The smell of the amp burning is going to be the bigger problem.

Ah OK. I just remember seeing your original thread and the low wattage prototype. What is the cost of one already built?

 

[mcdn]

A relay won't be fast enough -- certainly not at these current levels. A fuse will be fast but have to make sure it doesn't self modulate and cause distortion.

Anyway, have you tested your smaller load with over current to see what the failure mode is for instantaneous fault? Kind of like that video I post.

Now you’ve lost me, but I am here to learn! The load is happy taking 30A for a minute or more, why would we need a super fast relay?

 

[amirm]

Now you’ve lost me, but I am here to learn! The load is happy taking 30A for a minute or more, why would we need a super fast relay?

For anything above its maximum safe rating. But also see my last reply. I didn't realize it could handle so much current as is.

What is the max isolation voltage? These power supplies are clock to 80+ volts these days.

 

[mcdn]

Ah OK. I just remember seeing your original thread and the low wattage prototype. What is the cost of one already built?

About $200 if you already have a 3d printer and suitable cables. Mainly the board costs from JLC, heat sinks, 2 fans, and the case. PM me if you’d like me to build a couple for you.

 

[mcdn]

For anything above its maximum safe rating. But also see my last reply. I didn't realize it could handle so much current as is.

What is the max isolation voltage? These power supplies are clock to 80+ volts these days.

1oz copper on aluminium boards, so at least 1kV

 

[Old_School_Brad]

Now you’ve lost me, but I am here to learn! The load is happy taking 30A for a minute or more, why would we need a super fast relay?

You don’t. If something shorts momentarily and catches fire, it happens quickly. The situation can only be addressed afterward, typically with a thermostat or similar device that cuts off power, smoke alarm.

Under load temperature guarding would be a thermostat with an internal relay that opens when the temperature rises above an adjustable threshold can safeguard the load -a straightforward safety cut-off. These components are readily available off-the-shelf and can be had with manual or automatic reset.
If you’re inclined, you could implement two safety measures: one for the load and another for the amplifier.

As for protecting electrical components quickly, I’d opt for a resettable circuit breaker or an overcurrent / motor protection relay instead of a fuse.
Overcurrent relays can be precisely set to the maximum current the amplifier is expected to handle and adjusted for each test scenario on the fly so the threshold can be set just above what the max load is.

 

[restorer-john]

A short in an amplifier can vaporize and set load resistors on fire in a blink of an eye.

And it's statements like this that make sensible discussion or debate with you, impossible.

An average blink of an eye is around 100-150mS. You say your load resistors will take 600W continuous (4R) and 10x that short term. That's over 6kW. You won't set your load resistors on fire in 150mS with any amplifier you're likely to test even if the entire output stage went CE short or MOSFETs all go source-drain short.

You could even connect them directly into your 120V wall socket and they'd survive long enough to trip your breaker before they "caught fire".

 

[mcdn]

And it's statements like this that make sensible discussion or debate with you, impossible.

An average blink of an eye is around 100-150mS. You say your load resistors will take 600W continuous (4R) and 10x that short term. That's over 6kW. You won't set your load resistors on fire in 150mS with any amplifier you're likely to test even if the entire output stage went CE short or MOSFETs all go source-drain short.

You could even connect them directly into your 120V wall socket and they'd survive long enough to trip your breaker before they "caught fire".

It’s Amir’s site John. Yes there is some hyperbole, doesn’t mean you need to take it as bait.

 

[restorer-john]

These power supplies are clock to 80+ volts these days

80 volts? That's nothing.

I've got amplifiers with +/-112V (224V swings). Your MLs will have internal rails equal or greater than that.

Sometimes I have to daisy chain lab PSUs because I need more than +/-120V DC. Even the 1985 NAD 2200 (like you tested) runs on +/-110V rails.

A large receiver on my bench right now needs +/-99V and +/-106V and it's always best to supply current limited external power when repairing.

 

[amirm]

And it's statements like this that make sensible discussion or debate with you, impossible.

An average blink of an eye is around 100-150mS. You say your load resistors will take 600W continuous (4R) and 10x that short term. That's over 6kW. You won't set your load resistors on fire in 150mS with any amplifier you're likely to test even if the entire output stage went CE short or MOSFETs all go source-drain short.

You could even connect them directly into your 120V wall socket and they'd survive long enough to trip your breaker before they "caught fire".

I suggest following the discussion before commenting. We are NOT discussing my load resistors. I have no worry about them catching on fire and I said so earlier. We are discussing @mcdn's load which is made from large array of surface mount resistors with no structure around them to stop them from catastrophic failure.

 

[mcdn]

I suggest following the discussion before commenting. We are NOT discussing my load resistors. I have no worry about them catching on fire and I said so earlier. We are discussing @mcdn's load which is made from large array of surface mount resistors with no structure around them to stop them from catastrophic failure.

Well that is also a bit hyperbolic, there is no risk of catastrophic failure!

The load is made of 4 boards mounted in a tunnel with a fan at each end. Each board has 152 x 1% tolerance 2W rated 4R7 resistors, so even at 300W per board everything is within bounds. Each board is made of single sided aluminium and each resistor has a thermal via under it. There are 400g of heatsinks per module x 4 modules = 1.6kg. The case has two 140mm fans, each capable of 150cfm. All cables are 12AWG. Temperatures are monitored by a series of ds18b20 thermometers attached to the boards.

The design was researched well to achieve solder joint and board temps of 80C for long life and to remain within the design guide of the components. The fans can shift a thermal load up to 1.2kW continuously, although 800W is preferred for a 50C uplift from ambient. Use two loads if necessary for very big amps. Load accuracy is within 0.1%.

 

[restorer-john]

I suggest following the discussion before commenting. We are NOT discussing my load resistors. I have no worry about them catching on fire and I said so earlier. We are discussing @mcdn's load which is made from large array of surface mount resistors with no structure around them to stop them from catastrophic failure.

Here we go again.

Read you own post. You make no mention of MCDN's load resistors when you made your ridiculous claim.

And in that post you link a stupid Youtube video of a guy with some 1/4W carbon composition 5% resistors he's blowing up one at a time. I mean seriously, is that all you got?

I cannot think of any amplifier made in the last 50 years that will not shut itself off/disconnect the load in mere milliseconds in the face of a short circuit or output stage failure. Modern Class Ds are incredible as they can shut the power supply down too.

And when did you last vaporize and set fire to a set of load resistors? I've smoked up a few, but never 'vaporized' wirewound loads of any capacity.

 

[amirm]

And when did you last vaporize and set fire to a set of load resistors? I've smoked up a few, but never 'vaporized' wirewound loads of any capacity.

No load resistor looks like this that you and i have used:

So once more, please stop jumping in the middle of a discussion without understanding the context.

 

[amirm]

Well that is also a bit hyperbolic, there is no risk of catastrophic failure!

Once again, my assumption was that we were talking about the naked boards like above. What you have built and are proposing now is much more elaborate. I am unclear of its value with all the extra heatsinking you have added to it. I appreciate your effort there but it costing as much as using high power resistors.

As I have mentioned, whatever I use needs to be readily available to others to replicate the test conditions. I know how to do that with an array of high power resistors. Using your load complicates matters there unless you want to get in the business of producing commercial versions of it.

 

[mcdn]

No load resistor looks like this that you and i have used:

I know, fun isn't it! Amazing what we can do so cheaply with great quality and tolerances. Sure it needs some heatsinks adding for long term usage, but I posted those pics above.

 

[mcdn]

Once again, my assumption was that we were talking about the naked boards like above. What you have built and are proposing now is much more elaborate. I am unclear of its value with all the extra heatsinking you have added to it. I appreciate your effort there but it is now it costs as much as using high power resistors as I am now.

As I have mentioned, whatever I use needs to be readily available to others to replicate the test conditions. I know how to do that with an array of high power resistors. Using your load complicates matters there unless you want to get in the business of producing commercial versions of it.

Ah, that makes sense, sorry for the confusion. The naked boards were never meant to be used in high power applications, they are just very accurate. Once mounted in the tunnel the differentiator over high power wire-wound resistors is the continuous power available due to the large heatsink area. I've run these at 800W for 2 hours at a stable temp of 70C.