UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
 
Thread Tools
Old 28th Apr 2017, 3:54 pm   #21
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Quote:
Originally Posted by Skywave View Post
Quote:
Originally Posted by mhennessy View Post
. . . hardly any amplifiers use regulated power supplies and there are compelling reasons to stick to the simple approach.
Where, presumably, the "simple approach" entails not using a regulated PSU. So, just out of interest, what are those "compelling reasons" for not using a regulated PSU - apart from cost?

Al.
Sorry for the delay - it's really busy at the moment. Still, that gave time for some excellent responses, which makes this easier for me

I'll add this: I wouldn't dismiss cost - that is the most compelling of all the reasons. For clarity, that's not about being "cheap". Adding regulation has a number of technical drawbacks - even if done well - and no real advantages. The cost-benefit analysis says: "bad idea".

Notwithstanding that, Marketing is a big "pull". Especially as regulated supplies allows you to get closer to the perceived ideal of an exact doubling in power for a halving of load impedance. It never happens in practice as there are conductive losses in output stages that can't be overcome, but an amplifier that comes close to this will somehow be seen as "stiff" and "stoic". Actually, Douglas Self relates a tale in the later versions of his power amp handbook (see below), where he was asked to design an amplifier that did exactly that - for some reason the client was hung up on this magic doubling - and the only practical way to achieve it (because of the conductive losses) was to put a voltage limiter prior to the power amp. This means that the amplifier would never achieve its full potential when working into 8 ohms, so it was fundamentally a bad idea - Self talked him out of the idea...

It's very easy to design a power amplifier with more than generous PSRR, so there's no reason to worry about ripple from unregulated supplies. Of course, a power supply is a power amplifier that has a voltage reference as its input, and power supplies can be designed to have very low levels of ripple at their output, so why shouldn't a power amplifier be equally good?

Regulated supplies require a lot more components which - BOM and labour costs aside - will inevitably reduce reliability.

Don't forget the loss of dynamic headroom, which is typically 1 to 3dB, depending on design, and is important for real signals that aren't sine waves.

Some argue that a regulated power supply can include current limiting, thus making the current limiting arrangements in the power amplifier more straightforward (or omit them completely). However, such an argument ignores the fact that current limiting in the amplifier can be slope based, meaning the SOA of the output devices can be more fully exploited. The PSU current limiting can't do that unless it knows what the output voltage is - and would that work for stereo?

I could go on, but instead, have a look at Chapter 9 (page 266) of this book: http://files.books.elebda3.net/elebda3.net-gh-238.pdf - essential reading for anyone interested in audio. Many of the points made here and in earlier responses are covered, and see page 270 for more about the anecdote I mentioned above.

Hope that helps,

Mark
mhennessy is offline  
Old 28th Apr 2017, 9:31 pm   #22
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
Default Re: Define amplifier power rating from DC rails + speaker impedance?

It's a moot point, really - what would you call the thing if you ran everything from a stabilised rail except the output transistor collectors?

Assuming that the output stage is the generic complementary NPN/PNP emitter follower pair, if you took the output NPN collector to a slightly higher, unregulated rail but fed everything else from a regulated supply, you'd have inherently excellent PSRR at the output anyway. And the regulator need then only be relatively low power. Viewed like this, there is absolutely no reason to take the output collector to the regulated rail - excepting that you can have a bit more protection if the output transistor fails short-circuit.
kalee20 is offline  
Old 28th Apr 2017, 10:24 pm   #23
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Though it's extra complexity for no benefit. Especially given how easy it is to get excellent PSRR without regulation.

Regulated (or separate but not regulated) rails for the small-signal stages has been done occasionally over the years - we can only speculate at the reasons, but perhaps an inability to get adequate PSRR was a factor (maybe because the designer was using a particular topology that made that a problem?). Or marketing? Or, the whim of the designer, who perhaps had a strong belief in the concept for whatever reason - it's not unusual for ideas like that to fail to stand up to scrutiny today...

In fairness, while it's easy to judge from afar now, we should remember that prior to the advent of the Audio Precision System One in the mid-'80s, designers really were working in the dark with THD+N meters at best (e.g. HP8903) - as such, a lot of luck was involved. Of course Douglas Self did a lot to demystify the subject, most significantly with his series in Electronics World ("Distortion in power amplifiers") in '93 which ultimately evolved into the book. Today, it's a mature science.
mhennessy is offline  
Old 28th Apr 2017, 11:15 pm   #24
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Define amplifier power rating from DC rails + speaker impedance?

In the topology of most power amplifier circuits, in the very first stage, the signal is moved into the proximity of the power rails, away from ground.So it makes sense to consider these areas as rail-referenced not ground-referenced.

I wanted higher voltages to drive my output devices and I wanted to use low voltage transistors in the bottom of my cascode stages, so it made sense to add 15v regulated supplies, a positive one floating on top of the +ve heavy current rail and a negative one floating below the -ve heavy current rail. The regulated nature of these supplies give stable voltages to bias the long tail current sources and cascode stages. Of course the whole added supply sits on top of any ripple on the heavy rails, but it isn't hard to get lots of PSRR. With cascoded balanced stages repeated in NPN and PNP forms, it wasn't dificult at all. It just works.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 29th Apr 2017, 11:29 am   #25
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Yes, if you're using MOS-FETs, then additional rails are useful for voltage efficiency, and it doesn't hurt to regulate them (relative to the unreg rails). That's a 7815/17915 carrying a few 10s of milliamps - no real issues there. For prototypes, I've generated those rails from flying capacitors, but for production, it's usually cheaper to just specify the extra low-current windings.

Of course, with bipolar output devices, things are somewhat simpler in that regard. And there's something to be said for letting an earlier stage saturate...

For the standard "Lin" topology, the rail with the input mirror and VAS is the most sensitive to ripple. Cascoding the input pair helps (Early effect), as does R-C filtering.


Quote:
Originally Posted by Radio Wrangler View Post
The Quad 303 was regulated. It seems they did this because they were going close to the voltage ratings of the 2N3055 family, the beefiest devices of the period it was designed in, and the regulator saved them having to lower the power rail voltage and reduce the amplifier's power spec to allow for mains voltage variations. Clearly the right decision at the time.
This rang a bell with me, and when I was trying to get back to sleep in the early hours I eventually remembered the source: http://douglas-self.com/ampins/libra...nts1.htm#quadb

Quote:
Originally Posted by Douglas Self
Peter Baxandall once told me that the stabilised power supply in the Quad 303, which must have added significantly to the cost, was incorporated because he and the Quad team were worried about exploiting the full collector voltage capability (Vcbo= 100V, Vceo= 70V) of the relatively new 2N3055 output transistor, and that he later concluded it was unnecessary.
Better safe than sorry
mhennessy is offline  
Old 29th Apr 2017, 8:39 pm   #26
stevehertz
Dekatron
 
stevehertz's Avatar
 
Join Date: Sep 2011
Location: Rugeley, Staffordshire, UK.
Posts: 8,809
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Isn't obtaining good, unregulated, power rails just a matter of having a high quality, hefty mains transformer, beefy diodes and as much capacitance as you can buy or fit into the space allotted? eg tens of thousands of uF per rail? Each rail on my Pioneer SX-1250 is smoothed by 2 x 22,000uf caps the size of baked bean tins. That's 88,000uF for the stereo receiver.
__________________
A digital radio is the latest thing, but a vintage wireless is forever..
stevehertz is online now  
Old 29th Apr 2017, 9:11 pm   #27
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Sort of. But then some more recent regs about what you can do to the good old mains supply come in.

If I take a trad power supply and start improving it, by fitting a heftier transformer with lower copper resistance, fatter diodes and huge reservoir capacitor, then the regulation does indeed improve to a point set by the resistance of the mains wiring in the house and street.

The ripple voltage decreases. The monster C discharges by less voltage between cycles driving the same old load. This is good.

But having more voltage in the C when the next cycle comes up means that the diodes turn on later, closer to the peak, and there is less time left to recharge the C before the mains cycle goes over the top and its voltage falls, the diode turns off. Now both the beefy and the weedy power supplies are to feed the same load taking the same current, so they need to put back the same amount of charge into their respective reservoirs. The beefy PSU has to do it in a shorter time. The current from the mains is taken as a short, big pulse close to the peak of the mains voltage waveform. Ouch. The weedy power supply spreads it out a bit more. The current pulse is longer and hits a lower peak current.

Ironocally, losses are proportional to I squared times R. By cutting the time and scaling up the peak I, the losses have actully been increased. The resistive effects would really have to be decreased disproportionately by a lot to come out even!

So it isn't all plain sailing in the amplifier cabinet.

Worse happens elsewhere.

THe amount of electricity going into electronics has been increasing. The amount going into light bulbs and motors has been falling a lot. A bigger proportion of the load on a power station turboalternator is going to rectifiers and reservoir capacitors. The force on the windings wire in the alternator and the torque demanded by its rotor are proportional to the current taken from it. With nice resistive loads taking sinusoidal currents and three phases well balances, the torque is constant and smooth.

The pulsey current taken by the rectifiers acts as a repetitive shock load on the generator and creates forces like a hammer drill. Not good. To prevent failures due to metal fatigue etc there are now legal requirements on the harmonic current demands of higher power appliances. Unfortunately the people who did the sums underestimated the growth of small stuff that went under their radar.

This a a very serious problem in the power industry. The current surges scale up the losses in all of the network, not just the bits inside someone's amplifier.

Big amplifiers, to get approvals and to be legal for sale have had to do something about this, several years ago Linn had to go to switch mode supplies in their larger amplifiers.

We're going to see a lot more things with fancy power-factor corrected switchers in them. They're OK if done well, a disaster if corners are cut.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 29th Apr 2017, 9:50 pm   #28
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Just to help illustrate the point about charging currents, attached is a Spice simulation of a rectifier circuit (a simulation because I didn't have access to a DSO when I needed to produce it years ago).

It shows the AC from the transformer (light blue), the voltage across the smoothing capacitors (green), and the current taken from the transformer (red).

The load was a 10 ohm resistor, so the average DC current is around 650-700mA. Note that the charging currents peak at 3A! The smoothing capacitor was 2000uF, which is pretty generous for the supplied current.

Notice how the AC sine wave is visibly distorted just before it reaches the peaks. This is the voltage being dragged down by the current "gulps". This causes all sorts of problems upstream, as David mentions.

Power-factor correction involves trying to make the appliance look resistive - there are some clever techniques about, but as mentioned, they are almost always switched-mode power supplies of some form. Not an easy thing to design and manufacture if you're a small hi-fi company.

Personally, I feel that over-sized capacitors is rather brute-force and inelegant. When done, it's often for marketing reasons. It's simply not necessary.

One amplifier I built many years ago - and still use today - can give 4 times 50 watts, or 2 times 150+ (similar to your Pioneer). The smoothing capacitors are only 10,000uF, and there's only 2 of them. This is quite small for 300+ watts, but actually, there are no issues whatsoever. I could have fitted larger ones - it was a home job, so cost wasn't an issue - but there was really no reason to do so.
Attached Thumbnails
Click image for larger version

Name:	Charging Currents.png
Views:	71
Size:	47.9 KB
ID:	141757  
mhennessy is offline  
Old 29th Apr 2017, 10:14 pm   #29
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Thanks, Mark, that's a good illustration, though with it allowing 25% ripple, it's quite mild compared to some audio amplifiers out there.

Saleable hifi amps have to have wolloping huge reservoir capacitors because they're mostly sold on what they look like. Get this wrong and you go bust. Engineering never gets a look-in.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

Last edited by Radio Wrangler; 29th Apr 2017 at 10:26 pm. Reason: oops, it was Mark, not Jeremy. Do Currys still sell brains?
Radio Wrangler is offline  
Old 30th Apr 2017, 10:22 am   #30
stevehertz
Dekatron
 
stevehertz's Avatar
 
Join Date: Sep 2011
Location: Rugeley, Staffordshire, UK.
Posts: 8,809
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Ah, now we're into audiophool territory. The non-technical, 'know-it-all' audiophools who know everything there is to know about audio, audio electronics and lately, vibration in speaker cables that affects the sound enough to be heard. I mostly listen to my hifi when I'm getting on with other things, walking around the house even, and certainly not in that perfect stereo, seated position. I suspect a lot of other people listen like that too.
__________________
A digital radio is the latest thing, but a vintage wireless is forever..
stevehertz is online now  
Old 30th Apr 2017, 10:57 am   #31
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

There's no escaping it. Even the "sane" manufacturers have to play the game to an extent if they want the sales. For example, Arcam stick little bits of damping material to some of their electrolytic capacitors. They know full well it makes no difference to anything, but it costs very little, and makes great copy. They also do the same to crystals, and there's a grain of truth there, so that thinking naturally gets extended to other components - but the skill of the marketeer is to let the customer make those links. But, you've only got to read their service manuals to realise that the design team is a decent bunch of grounded engineers who know what they're doing.

The sad thing is that reviewers make up their mind about a product long before they actually listen to it...
mhennessy is offline  
Old 30th Apr 2017, 11:11 am   #32
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Quote:
Originally Posted by mhennessy View Post
Arcam stick little bits of damping material to some of their electrolytic capacitors.
Actually, it looks like they've stopped doing this. This cheers me up

For anyone who likes a bit of "hi-fi porn", there's a lot of really good official hi-res photos of Arcam stuff here: http://www.robfollis.com/photolibrary/Arcam/index.html

And more generally, this site is great, but perhaps not all that well-known? Here's a direct link to an earlier Arcam, complete with damping and rubber rings, but do explore the rest of it (say "goodbye" to the next couple of hours!): http://www.hifishock.org/gallery/ele.../a-80-3-arcam/
mhennessy is offline  
Old 30th Apr 2017, 11:18 am   #33
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Define amplifier power rating from DC rails + speaker impedance?

The real power capability of an amplifier can be quite easily calculated from a few voltages and an impedance measurement. But it is not this figure which sells amplifiers. It is the appearance of powerfulness that does the selling. Whacking great transformers (several for preference) Monster capacitors arranged in banks, welding cable strung on high strength brackets. Huge heatsinks. They sell amplifiers.

The big Japanese companies got the psychology very wrong. They put their heatsinks on the inside and put nice woods on the outside. Everyone knows that a heatsink's job is to move heat to the outside world. If they'd covered the things in big black fins everyone would have thought "What a powerful amplifier to need so much heatsinking!"

Terry Pratchett coined the worh "Headology" he knew a thing or two.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 30th Apr 2017, 11:55 am   #34
stevehertz
Dekatron
 
stevehertz's Avatar
 
Join Date: Sep 2011
Location: Rugeley, Staffordshire, UK.
Posts: 8,809
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Quote:
Originally Posted by Radio Wrangler View Post
The big Japanese companies got the psychology very wrong. They put their heatsinks on the inside and put nice woods on the outside. Everyone knows that a heatsink's job is to move heat to the outside world. If they'd covered the things in big black fins everyone would have thought "What a powerful amplifier to need so much heatsinking!"
Well, they didn't get it wrong with the Pioneer SX-1250, large cast heatsinks on the outside at the back. But I know where you're coming from. Out of interest, it's reckoned that the the 1250 was the high water mark in the late 70s battle of the big receivers. Not the most powerful at a 'modest' 160W/ch (Technics SA-1000 was 330W/ch) but it is excellently designed, fully shielded, well laid out, well manufactured and aesthetically beautiful. I've just restored one (not quite finished actually) and it was a joy to work on; plug in boards etc etc. But not fun to carry around at a back breaking 64lb..
Attached Thumbnails
Click image for larger version

Name:	Side.jpg
Views:	56
Size:	68.7 KB
ID:	141801   Click image for larger version

Name:	Front.jpg
Views:	60
Size:	60.4 KB
ID:	141803  
__________________
A digital radio is the latest thing, but a vintage wireless is forever..
stevehertz is online now  
Old 30th Apr 2017, 12:37 pm   #35
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: Define amplifier power rating from DC rails + speaker impedance?

Sadly, increasingly tough safety rules make it harder to put the heat sinks on the outside. It you're going to to it, you need much more effective heat sinks to keep the temperature below the limits, and extruded aluminium is really expensive. Put it in the case, and it can be almost any temperature you like, provided the junctions are within their limits. It's not uncommon to see 100 degrees C in practice. Of course, then you're into problems with limits on openings (3mm max), which make convection difficult. And it's not unheard of to see plastic grilles over heat sinks because plastic is allowed to get hotter than metal (because of reduced thermal conductivity to the skin). Nothing is straightforward!

I quite likes some of the Rotel designs with big heat sinks on the sides and front (the latter were cosmetic only, but that's fine with me!). Rotel stuff is good - basic design that is easy to troubleshoot and good performance (80s onwards).
mhennessy is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 9:40 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.