Topic Options
Copyright © 1998
Cadman Enterprises Ltd
Welcome to the Graham Slee Audio Products Owners Forum  Open to all owners plus those contemplating the purchase of a Graham Slee HiFi System Components audio product and wishing to use this forum's loaner program: join here (Rules on posting can be found here) This website along with trade marks Graham Slee and HiFi System Components are owned by Cadman Enterprises Ltd |
1970s Design Indulgence |
Post Reply 
|
Page <1 183184185186187 345> |
| Author | |
Graham Slee 
Admin Group 
Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298 |
 Post Options
 Thanks(1)
 Quote Reply
 Posted: 28 Dec 2021 at 5:52pm |
|
Just trying to post a full circuit diagram of a Japanese integrated amp from the late '60s.
Hoping you can open the image in a new tab and see enough of it. 
|
|
|
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
|
|
 |
|
|
Graham Slee
Admin Group
Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298 |
Post Options
Thanks(2)
Quote Reply
Posted: 28 Dec 2021 at 6:13pm |
|
You think amplifier designers know what they're doing? Two quotes from Doug Self: "Audio power amplifier design, even if confined to solid-state amplifiers, and even if further confined to those with bipolar output stages, is already too big a field for one person to know everything. I certainly don’t think I do. The journey continues." "The Study of Amplifier Design: Although solid-state amplifiers have been around for some 40 years, it would be a great mistake to assume that everything possible is known about them." To delight your ears, the amplifier designer cannot afford to string the customer along, even if many do, because, eventually, your ears will tell you the truth (if you can overcome the manufacturers arrogance, and his gang of trolls - many cannot). The journey does continue!
|
|
|
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
|
|
|
|
|
Sylvain
Senior Member 
Joined: 18 Jan 2010 Status: Offline Points: 481 |
Post Options
Thanks(0)
Quote Reply
Posted: 28 Dec 2021 at 8:00pm |
|
Certainly have learned much and still trying to understand those 185 pages.
Manufacturers still introduce their product as the '' reference'' ''ultimate development design '' . of Transistors of the 70's. Now the claim for the digital beast starts glorifying '' Digital amp offers the mend that Class A and Class B lacks'' in one manufacturers marketing ...
|
|
|
|
|
Graham Slee
Admin Group
Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298 |
Post Options
Thanks(1)
Quote Reply
Posted: 28 Dec 2021 at 8:21pm |
|
Decoupling - such an important thing nobody talks about. Well, some guess at it. "Big capacitors make big bass" Really? Decoupling is there to combat against the inevitably long power supply wires - which are inductive - and their action with semiconductors that have parasitic capacitance - all semiconductors are accidental oscillators. The decoupling capacitor acts as a clamp, preventing higher frequencies from manifesting themselves. An old AES paper got to the understanding of how bipolar transistors want to be Colpitts oscillators, finding the resonant frequency to be around 50kHz - which is probably a good guess - but gives us a starting place. If it is 50kHz and the nominal speaker impedance is 8 ohms - or the Zobel network helps it to be so - then by the 1/2pi formula the capacitive reactance to match gives 0.4uF. To polish it off, we might choose a value 10 to 100 times greater, so 4uF to 40uF. So why a big capacitor? Yes, it stores more energy, but something has to put it there, and that something is the rectifier, which conducts into the decoupler just like it conducts into the reservoirs - in jerky movements! Put a rudimentary current probe into the supply wire (admittedly that increases its inductance) and you see the charge per cycle, but it bobs up and down. It bobs up and down 2.5mS per 10mS which can only mean that the decoupler is part of the reservoir, but on the end of a wire. So, what does that do for our charge discharge cycle output stage? It gives us 25% of something we didn't expect (good luck Googling it!) The old wives tale was always to use the largest capacitor possible, but what if that causes it to go bright? Brightness and sibilance are distortion, and if it cannot be measured, then you have to peer into the invisible. We can only do so using the laws of physics. Let's start at the big capacitor end - my 680uF capacitor. The charge current swings to the amplifier were about so big (imagine). So let's half the value, and the current swings were twice as big. Oh, perhaps the current is getting to where it's needed? The 2.5mS charging was still there. Now let's third the value and have a listen to the brightness and sibilance - much better at 100uF! It appears that the distortion is being pushed up in frequency. Shall we try the 40uF value from above? The nearest I have is 47uF. Then again, the Japanese included a 1n capacitor in their 20 watt module - so is the oscillation frequency far higher than the AES paper? As Self states, decouplers are usually of the value of 10uF to 470uF. Why? I'm waiting... no answer. "Although solid-state amplifiers have been around for some 40 years, it would be a great mistake to assume that everything possible is known about them." (actually, since it was written more about 10 years ago, we need to correct that to 50 years)
Edited by Graham Slee - 29 Dec 2021 at 9:02am |
|
|
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
|
|
|
|
|
Graham Slee
Admin Group
Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298 |
Post Options
Thanks(0)
Quote Reply
Posted: 30 Dec 2021 at 12:34pm |
|
Thoughts on decouplers The early power amplifier circuits I've seen didn't look to have any thought for decouplers, save for perhaps a small value capacitor (<0.1uF). The first time I saw local decoupling was the recommended 1,000uF around 1976, which became a "rule of thumb" which lasted for all the period of the one issue of Practical Wireless it occurred in. Decouplers appeared again in the dual HT rail DC coupled amplifiers seen in the 1980's. Any effect they had would be symmetrical. However, could they have had a sound quality issue? Looking at Cherry's amplifier, two 470uF decouplers (one on the pos rail, the other on the neg rail) were considered obligatory. The other thing that amplifier required was bass boost - why? Cherry's design also needed all kinds of fixes for stability, which were presented as being essential to extend power bandwidth. Sometimes I wonder. Reducing the decoupler here to 47uF has improved the bass and lower mids, and has taken the sting out of sibilant singers, but I still feel that some high frequency rock guitar sounds more artificial than the player intended. The above results in similar fatigue to that which has prevented me releasing the power amp as a product. Since regulating the voltage stage supply, the decoupler serves an additional purpose, that being in providing the voltage regulator's input terminal capacitance. So, the decoupler cannot be dispensed with unless the design reverts to a completely unregulated design. In doing that, the output capacitor is then exposed to fluctuating DC level, and considering we have learned that the capacitor's charge is totally responsible for the "quality" of energy required for the negative half cycle, removing the voltage regulation looks to be a retrograde step. So, what is the absolute minimum voltage regulator input capacitance we can get away with? The datasheet says 1uF tantalum. Think about that - can you find a 1uF 125V tantalum capacitor? Therefore, the datasheet is more academic than pragmatic. The 1uF tantalum provides a "stoic" charge - its high dielectric absorption prevents it changing charge easily. But, it isn't as close to the regulator as the datasheet requires, simply because the ground return is at some length from the star ground. Then again, the datasheet shows an image suggesting the source ground is separate from the regulated ground - so it can "bob up and down". This is where datasheets can become extremely frustrating. However, by experiment, it was found by providing the capacitor with its own ground return, that it oscillated. It appears to me that the regulator's designer hasn't a clue! (those without experience might be shocked by this) When replacing a tantalum with an electrolytic (which can be obtained at the right voltage and is far more reliable), the capacitance needs to be increased by the proportion of the difference in the dielectric absorption. A factor of ten seems to be reliable - and often suggested in similar datasheets. Therefore, the smallest value of decoupler is 10uF. If 47uF still allows some audible high frequency distortion, where will we be at 10uF? It should be 4.7 times less obtrusive. Using the 1/2pi formula the suggestion is that it comes into play around 2kHz instead of just over 400Hz using the 47uF decoupler. Allegedly, in reading between the lines of Self, the effect must come into play at some 10 times that frequency. At 20kHz amplifier distortion tends to have risen somewhat from the oft quoted 1kHz, and so may not be a problem - and who will be able to discern a microsecond subtle difference on the harmonics of a cymbal stroke? As for any of this being measurable, I can assure you it isn't! It is only manifest to the ears, and so makes this fine tuning a pain in the ...
Edited by Graham Slee - 30 Dec 2021 at 2:38pm |
|
|
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
|
|
|
|
|
Graham Slee
Admin Group
Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298 |
Post Options
Thanks(1)
Quote Reply
Posted: 30 Dec 2021 at 5:51pm |
|
Although ripple rejection is well controlled, so much so that you have to press your ear to the speakers to hear any buzz or hum, and even then it's quiet - there has to be ripple current. Imagine a regulator where both voltage and current were held perfectly still. No amount of change in load could allow Ohms law to function. If only the supply voltage is held perfectly still, a change in load indicates a change in current. A voltage regulator may provide a constant voltage with a few millivolts of ripple, but as the load changes, so does the current, and if the current changes, then Ohms law tells us the resistance of the regulator must be changing. Likewise, in a bulk capacitor filtered power supply, a small load will not allow the ripple voltage to increase by much - it remains quiet. This assumes good power supply rejection to the voltage stage. However, the ripple current must vary every 10mS due to the rectifier action, and for 2.5mS each half cycle the reservoir capacitors receive some charge. Therefore, this should be seen as a frequency "drift" on the FFT (spectrum analyser) display - and now it is! It shows the signal is still modulated by the 100Hz rectified current, even if the voltage appears perfectly still. However, you should see both channels moving together, but here we now see one channel moving forward while the other moves back, but not by the 100Hz bandwidth, but by something much less. Unfortunately, this suggests there is still a stereo ground loop. By reducing the decoupler (to 10uF), one disappearing distortion has revealed the remnant of another. But this time it shows on the analyser. The overall distortion has hardly moved, although it is now possible to reduce the third harmonic to -90dB for a +10dB output (-100dB = 0.001%), and the second harmonic dominates at -65dB (+10dB makes -75dB = 0.017%). We now have dominant second harmonic distortion at one watt of just 0.017%, and nothing else worth talking about at domestic listening level. I think we now need to see where the power ground should really go.
|
|
|
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
|
|
|
|
|
Graham Slee
Admin Group
Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298 |
Post Options
Thanks(0)
Quote Reply
Posted: 30 Dec 2021 at 9:35pm |
|
The bottom end of the decoupler flows back to ground, but was found to be partly in the speaker 0V feed. This has been corrected such that power ground connects directly to the negative end of the decoupler. |
|
|
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
|
|
|
|
|
Post Reply
|
Page <1 183184185186187 345> |
Tweet
|
| Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You can vote in polls in this forum |
Forum Software by Web Wiz Forums® version 12.01
Copyright ©2001-2018 Web Wiz Ltd.
This page was generated in 0.125 seconds.
Copyright ©2001-2018 Web Wiz Ltd.
This page was generated in 0.125 seconds.