Hello All,
I have a question about power supplies for our old-time favorite Class AB amplifiers.
I have lately measured the THD+N of several amplifiers. Many, most if not all have with Full Wave diode bridge plus Capacitors even including inductors the power supply noise, you know the 50 or 60 Hz series, 60Hz, 120Hz, 180Hz and on up, of diodes turning on and off spikes dominate the THD+N figure. The power supply spikes are higher dB’s than the 3rd, 4th and 5th harmonics sometimes.
Switch mode power supplies in Class D amplifiers are quieter than the previous generation.
The question: Has anyone or does anyone use switch mode power supplies in old school Class AB amplifiers?
Just for grins here is a sample of a well behaved Full Wave diode bridge power supply in a NAD amplifier. I replaced with new the protection relay and all of the PS capacitors. Plus a not so well behaved example.
Thanks DT
I have a question about power supplies for our old-time favorite Class AB amplifiers.
I have lately measured the THD+N of several amplifiers. Many, most if not all have with Full Wave diode bridge plus Capacitors even including inductors the power supply noise, you know the 50 or 60 Hz series, 60Hz, 120Hz, 180Hz and on up, of diodes turning on and off spikes dominate the THD+N figure. The power supply spikes are higher dB’s than the 3rd, 4th and 5th harmonics sometimes.
Switch mode power supplies in Class D amplifiers are quieter than the previous generation.
The question: Has anyone or does anyone use switch mode power supplies in old school Class AB amplifiers?
Just for grins here is a sample of a well behaved Full Wave diode bridge power supply in a NAD amplifier. I replaced with new the protection relay and all of the PS capacitors. Plus a not so well behaved example.
Thanks DT
Attachments
Chord was among the first to use SMPSUs in their classAB power amps. From photographs it looks as though Soulution also uses them - off the shelf ones at that : https://www.soundstageaustralia.com/images/stories/equipment1/201806_soulution5.jpg https://soulution-audio.com/wp-content/uploads/2017/03/soulution_701_1410_011.jpg
Last edited:
The Benchmark AHB2 uses a SMPS and seems to be a very well regarded class AB amplifier.
Benchmark AHB2 Power Amplifier - Benchmark Media Systems
Benchmark AHB2 Power Amplifier - Benchmark Media Systems
Attachments
For me, the simplicity, reliability (often see 50-60 year old working power supplies with original capacitors) of a classic PSU and a short-term overload reserve is an important factor.
Of course, there is a downside: cost, weight, size , voltage drop at heavy loads and correct installation, earthing, grounding.
but I don't see a reason why smps doesn't work - in the 80's sony did something like that for AB power amps.
Of course, there is a downside: cost, weight, size , voltage drop at heavy loads and correct installation, earthing, grounding.
but I don't see a reason why smps doesn't work - in the 80's sony did something like that for AB power amps.
For class AB avoid SMPS , it is going to fail with in 10 yrs for sure.#
Most medical equipment still uses transformer for some reasons.
Most medical equipment still uses transformer for some reasons.
Linn products was the first manufacturer to have switch mode Power supplies in all their chip-amp class A/B amplifiers, long before Chord started with that.
Its about 24 years ago, and they still do that.
Linn 5125 was their first amplifier with switch mode supply, outperforming the former Linn Klout.
Linn AV5125 5-channel Amplifier Review | Audio Science Review (ASR) Forum
Last edited:
Hello,
I do not have much interest in medical equipment anymore. I worked in hospitals too long.
I want to be rid of the 60Hz, 120Hz 180Hz 240Hz ............. series of noise in amplifiers.
In terms of amplifier THD+N, power line noise is the worst offender.
Thanks DT
I do not have much interest in medical equipment anymore. I worked in hospitals too long.
I want to be rid of the 60Hz, 120Hz 180Hz 240Hz ............. series of noise in amplifiers.
In terms of amplifier THD+N, power line noise is the worst offender.
Thanks DT
Many use smps these days, I had an older Onkyo and now the Emotiva XPA gen 3 DR series. They sport a 3kw smps!
It works very well and I even decided to re build an old class AB amp using a smps as the power supply.
Not sure it will help you that much with the mains frequency stuff though. I was thinking the same but then decided to try and measure it.
Measurements were made using a simple audio usb ADC, line in and connected to the rails via a resistor and a 100uF cap. The rails were 45 and 55 volts so the scale is sort of off. I am surprised how similar they are so maybe my measurement is flawed.
First up is a traditional supply:
Second the smps:
It works very well and I even decided to re build an old class AB amp using a smps as the power supply.
Not sure it will help you that much with the mains frequency stuff though. I was thinking the same but then decided to try and measure it.
Measurements were made using a simple audio usb ADC, line in and connected to the rails via a resistor and a 100uF cap. The rails were 45 and 55 volts so the scale is sort of off. I am surprised how similar they are so maybe my measurement is flawed.
First up is a traditional supply:
Second the smps:
Has anyone investigated the effect of snubbers across the rectifier diodes (resistor+capacitor)?
It remains to measure the harmonics of the network (in the same composition of the connected equipment) 🙂
Noise from a poor quality SMPS could be worse than the diode switching spikes.
Also, why not try an inductor (e.g. 5mH) before the reservoir capacitors?
Also, why not try an inductor (e.g. 5mH) before the reservoir capacitors?
It works and it is done extensively in pro audio. BSS did it in 1988 in the EPC780, which almost everyone will agree was a "no-compromises" amplifier that never turned much of a profit.
Now there are lots of very high performance and very reliable amplifiers using switching power supplies. MC2, Lab Gruppen, L-Acoustics, FFA and many others are using switching power supplies and have been for many years.
As for failing within 10 years... maybe if the power supply is really low quality, but the MC2 E series has been around for over 15 years, and I haven't heard of there being many power supply problems. Linear power supplies, when properly designed, do tend to live longer, but both types can last 30+ years without issue.
I do not necessarily think that a SMPS will do much of anything to solve hum issues, however. It is standard practice to put small caps across the diodes in a bridge rectifier. If you are seeing 120 Hz ripple on the output, the most common cause (in my experience, unless you're dealing with a SE class A design) is a grounding issue.
Now there are lots of very high performance and very reliable amplifiers using switching power supplies. MC2, Lab Gruppen, L-Acoustics, FFA and many others are using switching power supplies and have been for many years.
As for failing within 10 years... maybe if the power supply is really low quality, but the MC2 E series has been around for over 15 years, and I haven't heard of there being many power supply problems. Linear power supplies, when properly designed, do tend to live longer, but both types can last 30+ years without issue.
I do not necessarily think that a SMPS will do much of anything to solve hum issues, however. It is standard practice to put small caps across the diodes in a bridge rectifier. If you are seeing 120 Hz ripple on the output, the most common cause (in my experience, unless you're dealing with a SE class A design) is a grounding issue.
Very well regarded? Damn thing is one of the best amps in the world!
Good clean SMPS can work very well. Here is a VHEX+ Class AB amp powered with an Abletec +/-53v 450w PSU. It has a pair of CRC’s (one for each channel from common SMPS output) consisting of 4700uF // 0.47ohm // 4700uF between the SMPS and amp. A good SMPS will not have any mains (50/60Hz or their 2/3/4th harmonics) or the high frequency hash. The diode switch noise can be eliminated using an active bridge like LT4320.
This is 25wrms into 8ohms and passing through a SSR speaker protection circuit. The SSR added no distortion.
Distortion vs Frequency at 25wrms:
SMPS are also useful for Class A amps too - here is an M2 amp with linear PSU:
Using a pair of 24v 5A SMPS bricks in series as a PSU in M2:
This is 25wrms into 8ohms and passing through a SSR speaker protection circuit. The SSR added no distortion.
Distortion vs Frequency at 25wrms:
SMPS are also useful for Class A amps too - here is an M2 amp with linear PSU:
Using a pair of 24v 5A SMPS bricks in series as a PSU in M2:
Attachments
Last edited:
I used switchers for a preamp/headphone driver.
They aren’t a panacea. 50Hz still makes it through to the output of the supply. Unlike a linear, where adding capacitance to the output is generally straightforward (within reason), on a switcher doing so will only prevent the supply from starting up.
They aren’t a panacea. 50Hz still makes it through to the output of the supply. Unlike a linear, where adding capacitance to the output is generally straightforward (within reason), on a switcher doing so will only prevent the supply from starting up.
I bet that was in fact mains leakage.
That's why I don't like SMPS wall warts all that much, they're generally 2-pin jobs and the mains filter generally is two Y2 caps from secondary-side ground to L and N (which is fine in Norway and the few other places with "technical" mains, but not in most other places with unbalanced single phase AC, where halfway between L and N is at half mains voltage). In polarized plug territory (e.g. UK, US) you could use just one going to N, or one going to PE where that is ubiquitous (e.g. Germany).
SMPS are pretty neat for powered monitors though. A lot of these are IEC Class I anyway, and every time you are mounting a conventional transformer inside a loudspeaker, mechanical hum is a potential issue... some models are quite known for it. One of my older K+Hs does not like polluted mains and will emit a noticeable hum/buzz e.g. when operating a hairdryer at reduced power (half-wave rectifier).
The one downside of having one is that a ground loop in the audio connection will cause excessive hiss, which you may not expect to be happening in such a scenario. (Quite common e.g. in JBLs and some others.)
Last edited:
Chord hasn't ever been in the business of chip-amp classAB but their use of SMPSUs in amps dates back to about 1990.
There was a company named Carver that did that over 40 years ago, had a patent too.
US4218660A - Audio amplifier and method of operating the same
- Google Patents