New Posts New Posts RSS Feed - 1970s Design Indulgence
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Welcome to the Graham Slee Audio Products Owners Club

 

Open to all owners plus those contemplating the purchase of a Graham Slee audio product wishing to use our loaner program: join here
Subscribe to our newsletter here (Rules on posting can be found here)


1970s Design Indulgence

 Post Reply Post Reply Page  <1 174175176177178 190>
Author
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 11997
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 29 Nov 2021 at 8:31pm
I remember it came in handy when there were two of us at the bottom of a half mile shaft, having to turn a bulldozer on its side to hook it to the underside of a cage (the type of lift used in a coalmine). We had two "Tirfors" and two pull-lifts. We didn't want it to swing as it lifted free, but it did anyway, but without bringing any girders down. It just caused a bit of dust!

The law of conservation of energy comes in three parts written by our old friend Kirchhoff, and it helps us understand things like vectors - the unseen that must exist for these things to behave within physical laws.

All we tend to see is voltages, because voltage is easy to measure. Current likes to hide and because it isn't easily seen, when it is seen, we need to express it in understandable ways. From that, we can base analogies.

The objects can be a bulldozer or the load represented by a rudimentary loudspeaker. The force required to move it a distance might be considered like a voltage, but the bulldozer has mass due to gravity, and that mass must be overcome to make it start to move.

Here it helps us see the energy required on the positive half cycle, which has to be more in one direction, if there is less in the other direction. The vectors must sum to the same as the symmetrical for a complete sinusoidal wave. So the asymmetry helps us see the "mass" increase in the positive half cycle. The vectors represent power drawn from the supply, so "mass" here is current.

We then need to imagine how this asymmetry affects the power supply. All the positive half cycles of any frequency comprise loudspeaker current plus output capacitor charging current, and as that is given up to the loudspeaker load too, via the output capacitor "shuttle", all the positive half cycles draw the total load current.

There is unity gain in an emitter follower, but due to the asymmetry and the positive half cycle current being roughly twice what you'd expect, there must be gain in the upper emitter follower. Where there is gain, there has to be poorer power supply rejection. This demonstrates that the power supply "sees" an asymmetrical current demand.
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 11997
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2021 at 12:02pm
The 2nd set of diagrams infer asymmetrical loss. If the frequency being amplified were to synchronise with the current pulse of the conduction angle, perhaps we could see the effect on power supply current.

Synchronised, we should see one transformer current pulse smaller than the other. It would take some serious dexterity with my 'scope, but I'll have to try it.

It's funny, but some people buy high quality audio to listen to weird groups like ELP, Atomic Rooster, even The Crazy World of Arthur Brown. Most amplifiers will do a half decent job of reproducing natural instruments, but when orchestrated to a crescendo, might sound odd, and playing weird stuff, can produce an ear shattering din. This is what we want to avoid.

Class AB of the truly differential type, plus push-pull valves, as far as I can see, need no regulation, because what happens to one side, also happens to the other side, and mathematically, the result is zero because both sides of their equations balance.

I think we may be safe to say that anything which doesn't result in a balanced equation, needs something to balance it, and in this difficult situation, it is best to go back to basic engineering science, and use "what we know".

Where there is waste, then to comply with the laws of thermodynamics, the result must be heat. So knowing that heat allows the equation to balance, we need to find the point at which heat is generated, and then allow it to happen.

An example is class-A: it produces vast amounts of heat simply because it wastes power. How can we waste power in this class AB?

One solution is to apply voltage regulation, and the series pass transistor generates the heat.

Here we have voltage regulation to the voltage amplification stage, and it could be argued that the unity gain current amplifier stage equates to the series pass transistor of a voltage regulator.

It would do if it had unity gain, but it doesn't. From a purely voltage approach, it must be unity gain, but in analysing the current, it cannot be.

This is due to the upper EF "threading" current into the output capacitor, to be used on the negative excursion. The gm of the upper EF cannot be made to the value of one. Therefore, because its gm is greater than one, infers gain, and so the output stage cannot carry forward the regulation given to the voltage amplifier stage (well, not all of it).

Without resorting to an overall voltage regulator, what could we use? Do we really need voltage regulation to do the "big watts!? We mostly listen at one watt or less, and the tallest transients might be 25 times that. Can we honestly say that the tip of a transient is long enough to register that we can assess its quality?

Thinking about it, an overall voltage regulator does to the rectifier exactly what a class-A amplifier does. But it's so wastefull. The reservoir capacitors would not last very long in both cases.

Each of the above widens the conduction angle, so the transformer is loaded for a longer period. Perhaps we can make the transformer current larger? It makes sense from the heat result, in that heat is the product that balances the equation.

The proof of a pudding is in the eating, so perhaps we could give a transformer with a high magnetising current a try? The only type capable is the EI laminated. A toroidal transformer run at high magnetising current can be easily tripped into saturation, which does result in heat, but destructive heat.

The EI laminated transformer is gapped which regulates increasing magnetic flux.

Well, surprise - surprise, the cap coupled amplifier does sound much better using a high magnetising current EI laminated transformer than a toroidal one, and this is a lasting thing.

But will any EI laminated transformer do? Having had a low magnetising current EI transformer made, and it not resulting in good sound by a long chalk, then perhaps we could be converging on an answer.

The problem we then have is in obtaining an EI laminated transformer of the right magnetising current range. EI transformers have been going out of fashion for the past 20 years or so. They are labour and process intensive. Far easier to wind toroid's and C-cores where the cores can be wound by the steel supplier.

Just how do we make a toroidal transformer emulate an EI transformer? This is the point in development of this amplifier, which we must now discover.
Not simple enough for Google-Bot to understand...
Back to Top
Sylvain View Drop Down
Senior Member
Senior Member


Joined: 18 Jan 2010
Status: Offline
Points: 257
Post Options Post Options   Thanks (0) Thanks(0)   Quote Sylvain Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2021 at 12:34pm
Thank You for the elaboration in semi-technical but I say able to understand the Consideration at play and your very competent analysis to problems solution.

I read that the US has geared it's industry to develope Toroidal as an ''Absolute'' whereas the Orientals Japan and other will go the extra mile to a R-Core or EI transformer for the better sound. 

I am confident that you are on the right track but audio TRANSFORMERS and people to construct at reasonable honest cost and margin are available also outside the Yorkshire boundaries. 
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 11997
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2021 at 2:01pm
Another thought is why Cordell's unregulated power supplies have 2k capacitor discharge resistors, when 10k will do?

2k over 35V uses 17.5mA and consumes nearly 2/3rds of a watt. It adds little to the quiescent load, but will extend the capacitor conduction angle some. Cordell gives no reason apart from it quickly discharging the capacitors.

The more current used, the better things get, but at domestic listening levels, little current is used. Self insists on 0.1R emitter resistors and non-gm doubling quiescent voltage. That uses 250mA. But here we have dual rail amplifiers. Perhaps, because the charge pulses are the same as the single rail, but being symmetrical, should cancel, might still result in similar sonic problems?

The end result, however, is heat. 
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 11997
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2021 at 2:14pm
And now for "DC on the mains".

If the electricity company put DC on the mains, then how much would they need to put in so you got the 1.2V DC everybody bangs on about?

I'm sorry to disappoint but DC cannot pass a transformer, and substations transform the ultra high voltage to usable 240V ac. And even if the transformer miraculously passed DC, how would the miles of overhead cables deal with it?

With DC you get voltage loss, but the so called DC is always the same reported voltage.

What you do get is voltage offset, which is due to half wave rectified loads on the local mains. The domestic washing machine is mainly responsible, and it's often your own. It could be your neighbour's if the impedance of your localities supply cable allows.

It causes one side of the current waveform to collapse, and puts it on the other side. The increase in one-sided magnetising current causes the transformer to saturate on those half cycles, due to the doubling of current, and increasing magnetising current causes a transformer to emit noise.

Trying to take both sides down with a so-called DC blocker, simply worsens the sound.


Edited by Graham Slee - 30 Nov 2021 at 2:14pm
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 11997
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 03 Dec 2021 at 8:45pm
An hypothesis of sorts.

Using a low magnetising current transformer, at the end of the smoothing capacitor conduction angle there looks to be negative current. This with a low volume signal.

negative magnetising current

Using a high magnetising current transformer, albeit at idling with no signal, the magnetising current remains positive.

positive magnetising current

This transition just happens to be in the amplifier's sensitive crossover area.

The high magnetising current transformer also lent a more balanced sound quality.
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 11997
Post Options Post Options   Thanks (2) Thanks(2)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 04 Dec 2021 at 12:11pm
It is very difficult to display the relationships between primary current and smoothing capacitor conduction angle (except what is seen at the transformer primary), but soft bass is due to the lack of energy, and soft bass places the emphasis on the treble. If anything is removed from the musical signal it is heard as distortion (what else can it be?)

I would have expected to see zero current or some small amount of positive current, but anything below 0 with respect to the cyclic direction, is negative. Would this represent some unwanted capacitor discharge?

That discharge must be via the rectifier, so is the transformer pulling down the capacitor charge? How long does this take to manifest itself? Could it be 3 - 7 days?

We have to bear in mind the dielectric absorption of the capacitor - is it fast? OK, let's do an experiment: charge a large electrolytic for a goodly amount of time. Now, discharge the capacitor safely, via a power resistor of suitable value, and monitor the voltage until it reaches zero, and leave it discharging for another goodly amount of time.

Place the capacitor on a shelf and leave it a few weeks, then gradually short its terminals. At some point you should see it arc.

A good capacitor might have superb DA behaviour for 90% of its charge range, but in its last 10% or less, it is obviously poor.

It looks like making transformers to be so absolutely mechanically silent, might have caused a problem few seem to be aware of.

However, in my opinion, some amplifier manufacturers are aware of it. Just that they choose not to mention it - a trade secret?

But I'm still developing my hypothesis.
Not simple enough for Google-Bot to understand...
Back to Top
 Post Reply Post Reply Page  <1 174175176177178 190>
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.01
Copyright ©2001-2018 Web Wiz Ltd.

This page was generated in 0.092 seconds.