PMA-NC252MP power amplifier

[pma]

I assume the test is on one channel only, so I assume that with both channels running the problem may be more apparent as they share same power supply.

Absolutely correct, I have only one "good enough" resistor load rated at 400W so I can test only 1 channel at such high power.

 

[pma]

We have been discussing in another thread the definition of measuring bandwidth for power amplifier distortion measurements, with conformance of practices used for decades and also with standards like IEC 60268-3.

https://webstore.iec.ch/preview/info_iec60268-3%7Bed5.0%7Db.pdf

The key is that distortion measurements with bandwidth of some 20kHz loose any technical reason and point above about 4kHz, because only few harmonic components are then taken into account which absolutely cannot reflect non-linearity of the amplifier. For decades, the common practice was to use at least 80kHz measuring BW. This has changed after the introduction of class D amplifiers, that were unable to compete under standard conditions of distortion measurements.

As a comparison, below please see THD+N vs. power measurements, into 4 ohm load, of Hypex NC252MP amplifier and of my A250W4R class AB amplifier (which has just average parameters for the class), measured at frequencies from 1kHz to 20kHz, with 90kHz measuring bandwidth. I think the plots speak for themselves. With the higher measuring BW, the class D amplifier cannot compete to an average class AB amplifier.

 

[pma]

Announcement

I bought an actual valid issue of the IEC 60268-3 standard (year 2018), the one that is linked for example below:

IEC 60268-3:2018 | IEC Webstore

IEC 60268-3:2018 Standard | Sound system equipment - Part 3: Amplifiers

webstore.iec.ch

For those who might be interested (@restorer-john , maybe others as well) - let's try to suggest a concept of power amplifier testing that would be in conformance with current valid standards. PM me if you are interested.

 

[pma]

I would like to add, from now, some measurements with respect to IEC 60268-3 standard.

Characteristics of d.c. offset protection circuits (abstract)

NOTE: These characteristics do not apply to amplifiers with only capacitive coupling to the load terminals.

Characteristics to be specified
The following characteristics should be specified:
a) rated maximum offset voltage at the load terminals, stated by the manufacturer in the specification, that does not operate the protection circuit;
b) rated response time of the protection circuit, optionally stated by the manufacturer in the specification, to an offset voltage of 30 % of the highest d.c. supply voltage of the amplifier output stage;

Methods of measurement
Methods of measurement for amplifiers which continuously sense the d.c. offset condition

The procedure is as follows:
a) The amplifier is brought under standard measuring conditions, and the input signal frequency is then changed to 20 Hz. A d.c. oscilloscope and a d.c. voltmeter are connected to the load terminals.
b) The d.c. balance of the amplifier is then slowly changed, by a suitable method which depends on the detailed design of the circuit and cannot be standardized, and the maximum reading on the d.c. voltmeter, before the protection circuit operates, is noted. This is the maximum offset voltage that does not operate the protection circuit.
--------------------------------------------------------------------------------

Measurement of DC offset protection of Hypex NC252MP

This amplifier is DC coupled and it transfers amplified DC voltage from input to output.

Input test signal - 20Hz sine + DC offset and amplitude jump after 1s

Input spectrum before the amplitude jump

NC252MP measured response to the input signal

The input signal level in the 1st second is at "standard measuring conditions" level of the output. Then comes the 20dB jump. One can see that NC252MP transfers DC average level of 10.69V to the output, without any action of the module DC protection. This is dangerous and may lead to speaker voice coil damage in case of DC error signal at the amplifier input. After 1s, when the input signal DC jumps of 20dB, the protection circuit finally takes action. It shuts down the amplifier output in 7.8ms after heavy DC clipping.

From my point of view, DC offset protection action of the Hypex NC252MP is unacceptable.

 

[DonH56]

Isn't that the same as any other DC-coupled amplifier? Not saying that is necessarily good, but I do not think it uncommon. Seems like the input buffer should be capacitively coupled to keep DC from reaching the input stages.

 

[Rick Sykora]

Isn't that the same as any other DC-coupled amplifier? Not saying that is necessarily good, but I do not think it uncommon. Seems like the input buffer should be capacitively coupled to keep DC from reaching the input stages.

I had asked Purifi about this and the response was that they expected the source device to take care of the DC. Agree it seems that it is more common for amps to pass the DC. Suspect many more mainstream amp designs take a more conservative approach and block the DC.

Given the heritage, expect Hypex takes the same stance. Most passive speakers will have a cap in series with the most susceptible drivers. Anyone who hears an ugly pop and allows it to occur repeatedly should not be surprised if the speaker fails.

 

[restorer-john]

I had asked Purifi about this and the response was that they expected the source device to take care of the DC. Agree it seems that it is more common for amps to pass the DC. Suspect many more mainstream amp designs take a more conservative approach and block the DC.

Given the heritage, expect Hypex takes the same stance. Most passive speakers will have a cap in series with the most susceptible drivers. Anyone who hears an ugly pop and allows it to occur repeatedly should not be surprised if the speaker fails.

There's absolutely nothing wrong with a pure DC coupling and no capacitor at the front end as long as there is some from of DC servo and/or very fast and very sensitive DC protection. It seems the Hypex DC protection works, but may be a bit slow to react.

Traditional amplifiers have input capacitor coupling and even if they don't, they have 100% NFB at DC via the NFB R/C network. And they have fast DC, overcurrent and short circuit protection. Not only that, sources always had capacitor coupled outputs in any case.

I have several pure DC preamplifiers and power amplifiers and they do not appreciate poorly designed D/A converters or anything with DC offsets. It's rather disconcerting to attach a source and watch a woofer cone displace more and more as you turn the volume control. And that is long before the amp's DC protection trips in. Most amplifiers will need about 10-12V DC or the same level <4-5Hz to trigger a DC event. Most protectors are based on the zero-cross period not exceeding a finite time, not actual DC itself. The assumption being if the zero cross takes too long, it's close enough to DC to cause trouble.

@pma can measure the period of time and level it takes and the LF cutoff for the 'DC protector' in the 252MP. He mentioned 12V, which is pretty typical and if the period is less than a second, hopefully a woofer won't be totally destroyed by launching the voicecoil into the rear pole plate or firing the VC out of the gap and across the room. I would be wanting 100mS for DC detection and that is what I have found to be bulletproof in my speaker protectors.

 

[DonH56]

The last graph in @pma's post #124 shows a very slow, slight drop in the average mean/offset (look at the lowest peaks along the bottom) over the 1-second burst, far too slow IMO to be useful as DC protection. I am cogitating; my processor puts out a healthy thump at turn-on and turn-off, and I am thinking maybe I should add some series caps to the various amplifier inputs. Might be a useful option for @Buckeye Amps to add.

 

[restorer-john]

I am cogitating; my processor puts out a healthy thump at turn-on and turn-off, and I am thinking maybe I should add some series caps to the various amplifier inputs.

Does the processor itself have output caps and if so, why is it sending a 'thump' out at turn on? (assuming it's a split rail, opamp coupled output with tracking rails coming up together).

Personally I find the lack of care for upstream and downstream components these days a case of buck passing. I miss the days where a component powered up utterly silently, had no effect on other components and did its job well. The rot set in when amplifiers started eschewing relay speaker protection and all these sources trying to get super low THD numbers at low frequencies while saving money omitting large caps and significant time delayed startup/shutdowns.

A modern DAC/processor should drop a set of relay contacts across the outputs immediately the power is cut and delay the connection on power up to ensure all the rails are stable and DC offset is at a minimum.

 

[DonH56]

Does the processor itself have output caps and if so, why is it sending a 'thump' out at turn on? (assuming it's a split rail, opamp coupled output with tracking rails coming up together).

Personally I find the lack of care for upstream and downstream components these days a case of buck passing. I miss the days where a component powered up utterly silently, had no effect on other components and did its job well. The rot set in when amplifiers started eschewing relay speaker protection and all these sources trying to get super low THD numbers at low frequencies while saving money omitting large caps and significant time delayed startup/shutdowns.

A modern DAC/processor should drop a set of relay contacts across the outputs immediately the power is cut and delay the connection on power up to ensure all the rails are stable and DC offset is at a minimum.

Don't know; it's insanely expensive (SDP-75/Trinnov Altitude 32) to have such issues but I have not gotten around to troubleshooting. I think it uses a pro-level PC sound card and in my limited experience those do not have common input/output protection circuits like most "hi-fi" components.

A modern DAC/processor should drop a set of relay contacts across the outputs immediately the power is cut and delay the connection on power up to ensure all the rails are stable and DC offset is at a minimum.

Agreed!

 

[pma]

Isn't that the same as any other DC-coupled amplifier? Not saying that is necessarily good, but I do not think it uncommon.

No. Any seriously designed amplifier, regardless it is DC coupled or not, has DC output protection circuit that disconnects output if the output DC voltage exceeds certain dangerous level, usually about 2V, that lasts more than fraction of a second. Even decades old commercial amplifiers had such protection circuits. My old and cheap Yamaha AX396 had such protection as well.

 

[pma]

@pma can measure the period of time and level it takes and the LF cutoff for the 'DC protector' in the 252MP. He mentioned 12V, which is pretty typical and if the period is less than a second, hopefully a woofer won't be totally destroyed by launching the voicecoil into the rear pole plate or firing the VC out of the gap and across the room. I would be wanting 100mS for DC detection and that is what I have found to be bulletproof in my speaker protectors

But NC252MP output DC offset protection threshold is just 12V - check the datasheet. It is not limited to 1s, equal and below 12V, <12V lasts at the output as long as there is equivalent DC at the input. The protection circuit acts above 12V. 12V makes 35W in the 4ohm voice coil, without cooling by the coil motion - DC does not move. It will fry the voice coil and it does.

 

[pma]

To make it absolutely clear - below see how the NC252MP "DC protection" works in case of input DC signal. The input DC is gradually increased, and the output is measured.
The protection circuit was activated at 26Vdc!! X-scale is 1s per division.

This will fry any speaker!

Who cares? Frankly, I feel like Ralph Nader. All the industry insiders and assemblers do not care and will explain that such behaviour is OK. Maybe @Hypexsales should care? Or ask his designers, Niels?

 

[Sokel]

There's a very interesting conversation with Bruno Putzeys in diyaudio about this which he states:

I would never consider putting an amplifier in use (apart from experimental testing when you're always present) that doesn't have DC protection on board.

DC protection is sometimes omitted in low-power systems where amps and speakers are co-packaged and inseparable, and where safety testing has demonstrated that there is no fire hazard.

I am certain that no UL approval is granted to separate amplifiers without dc protection. If a house burns down due to a failed amplifier, and the amp has no UL listing, the manufacturer is required to foot the bill all by himself.
(note to non-US readers: UL means Underwriters Laboratory and an underwriter is an insurance company. No UL approval means no insurance).

It is legal to sell non-UL approved equipment (at your own risk!) in most of the US except California. Similar safety requirements hold in Europe, and are squarely mandatory (no no personal risk taking is allowed).

All here:

SMPS design for power amps (Class D in particular)

Hi Bruno, I am still trying to figure if you are intending to use a clamped forward or flyback converter in your design. In any case I think you will need a series diode in line with your clamp switch. Best regards, Jaka Racman

www.diyaudio.com

 

[pma]

There's a very interesting conversation with Bruno Putzeys in diyaudio about this which he states:

Thank you.

They write in the datasheet:
6.3 DC-Error
If a DC component is present at either of the speaker outputs, the nFatal bus will be pulled down. This triggers
the main SMPS to switch off and the DC Error to be activated. To reset the DC Error the module must be
disconnected from mains for at least 10 minutes to allow the primary capacitors tor drain. A DC Error does
not affect the standby SMPS.

and also

Signal Coupling
To achieve optimal signal coupling, the audio signal inputs are all DC coupled. One must make sure that the
connected application is free of DC offset.

Isn't it funny? The real behaviour of their "DC protection" shown in my measurements.
Not sure how much Bruno was involved in this.

 

[Sokel]

Isn't it funny? The real behaviour of their "DC protection" shown in my measurements.
Not sure how much Bruno was involved in this.

Not sure he is,to quote him again:
(from the same thread)

The hypex modules do not incorporate DC protection because usually you have just one per box, instead of 1 per channel. So it's left to the user to make such a centralised protection. Maybe it would be a good idea to incorporate a detection on the module (status pin).
I hear that JP is making a psu tailored for the modules that incorporates a dc protection.
Simply turning off the amplifier will no longer do when it's blown. Molten silicon is hard to turn off

I think he differentiate himself from that design.
(the part about molten silicone is the horror)

 

[pma]

Not sure he is,to quote him again:
(from the same thread)



I think he differentiate himself from that design.
(the part about molten silicone is the horror)

The hypex modules do not incorporate DC protection because usually you have just one per box, instead of 1 per channel. So it's left to the user to make such a centralised protection. Maybe it would be a good idea to incorporate a detection on the module (status pin).
I hear that JP is making a psu tailored for the modules that incorporates a dc protection.

Understood. All the responsibility is put to OEM assemblers, who are often as qualified as to be able to make wiring in the box. Understood.

 

[Sokel]

The hypex modules do not incorporate DC protection because usually you have just one per box, instead of 1 per channel. So it's left to the user to make such a centralised protection. Maybe it would be a good idea to incorporate a detection on the module (status pin).
I hear that JP is making a psu tailored for the modules that incorporates a dc protection.

Understood. All the responsibility is put to OEM assemblers, who are often as qualified as to be able to make wiring in the box. Understood.

Of course that is a very old thread but is useful to know the thoughts and practice.
And as your measurements show not many things changed,even after decades.

 

[pma]

I have contacted Hypex support and got quickly this reply:

Hi Pavel,
Yes, output should be DC protected. Once there is more than 12V DC on the output for a longer period of time (RC time of a couple ms), the amplifier should turn off.
Please send me the serialnumber of the NC252MP so I can check the schematic why your module fails to go into DC protect at this level.

So, the output DC above 12V should trigger the DC protection. However, not <=12V and this is too high value. All in all - OEM assemblers should add separate DC protection board or at least make capacitive coupling at the input to prevent damage from possible input DC error voltage. It does not make any sense to speculate if DAC should or should not send DC, the amplifier must have appropriate circuit protection for error situations.

 

[Rick Sykora]

I have contacted Hypex support and got quickly this reply:



So, the output DC above 12V should trigger the DC protection. However, not <=12V and this is too high value. All in all - OEM assemblers should add separate DC protection board or at least make capacitive coupling at the input to prevent damage from possible input DC error voltage. It does not make any sense to speculate if DAC should or should not send DC, the amplifier must have appropriate circuit protection for error situations.

Agree that adding some extra protection against DC is a good idea.

However, Hypex seems to be stating that your amp is not in spec. So is not like they do nothing about DC protection. Are there ASR owners claiming speaker damage from a Hypex (or any other amp) for that matter?