1970s Design Indulgence

Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic

   Anybody who ever made a small signal amplifier with non-existent PSRR, with an unregulated power supply, will have heard what smoothed DC sounds like. Valve amplifiers had poor PSRR, and hum, so they used large chokes to filter the noise. Thus began the practice of taking the DC from across the last smoothing capacitor - after the choke!
The practice has continued into solid state, except that with current sources, PSSR improved, and with lots of NFB, the output is "hum-free."

Also, the choke is no more.

So, wires or PCB track connect the rectifier to the smoothing capacitor, and from there to the amp. Each wire or track has inductance. Imposed on the ripple voltage is the current burst (ripple current) which is 6.2 times the load current which lasts 10/62nds per half cycle. This is pulsed through each wire or track - which have inductance - which by ohms law results in a voltage, and will appear at the amplifiers output as noise.

The amount of noise depends on the amplifiers PSSR, and might measure sufficiently low (and definitely so A-weighted), but put your ear up to the woofer and you might hear it.

Therefore, everything you play includes the pulses no matter how "quiet".

By seeing the capacitors with their wiring as "the smoothers", you see where they are connected, which is straight across the bridge, and so that is the lowest impedance point. Taking the HT from that point, there can be no series wiring pulses.

Actually, there is some residual, but a heavy duty bridge has large metal tabs, and with take-off wires soldered directly to them, are very low inductance - possibly less than 5nF between the diode junctions to the take-off point. Compared to the wires/tracks to the capacitors - which are 20nH x 2 per inch, the series inductance is reduced greatly.

Put your ear up to the woofer and you will not hear it.

However, when it comes to split supplies, the ground cannot be taken to the transformer centre tap, so you must put up with the charging pulses being part of the music. The "solution" is to add more decoupling, more NFB requiring more transistors to obtain the large open-loop gain, but yet the sound still contains the pulses, but at a reduced level due to all that's thrown at it.

Anything than use an output capacitor.

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic
Graham Slee Members Profile Send Private Message Find Members Posts Add to Buddy List Admin Group Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298	Post Options Post Reply Quote Graham Slee Report Post Thanks(2) Quote Reply Posted: 25 Aug 2021 at 12:31pm
	Anybody who ever made a small signal amplifier with non-existent PSRR, with an unregulated power supply, will have heard what smoothed DC sounds like. Valve amplifiers had poor PSRR, and hum, so they used large chokes to filter the noise. Thus began the practice of taking the DC from across the last smoothing capacitor - after the choke! The practice has continued into solid state, except that with current sources, PSSR improved, and with lots of NFB, the output is "hum-free." Also, the choke is no more. So, wires or PCB track connect the rectifier to the smoothing capacitor, and from there to the amp. Each wire or track has inductance. Imposed on the ripple voltage is the current burst (ripple current) which is 6.2 times the load current which lasts 10/62nds per half cycle. This is pulsed through each wire or track - which have inductance - which by ohms law results in a voltage, and will appear at the amplifiers output as noise. The amount of noise depends on the amplifiers PSSR, and might measure sufficiently low (and definitely so A-weighted), but put your ear up to the woofer and you might hear it. Therefore, everything you play includes the pulses no matter how "quiet". By seeing the capacitors with their wiring as "the smoothers", you see where they are connected, which is straight across the bridge, and so that is the lowest impedance point. Taking the HT from that point, there can be no series wiring pulses. Actually, there is some residual, but a heavy duty bridge has large metal tabs, and with take-off wires soldered directly to them, are very low inductance - possibly less than 5nF between the diode junctions to the take-off point. Compared to the wires/tracks to the capacitors - which are 20nH x 2 per inch, the series inductance is reduced greatly. Put your ear up to the woofer and you will not hear it. However, when it comes to split supplies, the ground cannot be taken to the transformer centre tap, so you must put up with the charging pulses being part of the music. The "solution" is to add more decoupling, more NFB requiring more transistors to obtain the large open-loop gain, but yet the sound still contains the pulses, but at a reduced level due to all that's thrown at it. Anything than use an output capacitor.
	That none should be able to buy or sell without a smartphone and the knowledge in how to use apps

BAK Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 14 Mar 2010 Location: Kentucky, USA Status: Offline Points: 1744	Post Options Post Reply Quote BAK Report Post Thanks(2) Quote Reply Posted: 27 Aug 2021 at 7:42pm
	"Anything than use an output capacitor." Yes, the "hifi" industry today is convinced that all capacitors should be eliminated from the audio signal path. And to achieve this they will do everything they can to design circuits using split or dual supplies, ie, + and - voltages. They will add more decoupling and add more NFB requiring more transistors to obtain the large open-loop gain. What many newer designers are forgetting is: 1. it is almost impossible to make + and - supply volts exactly equal to keep the signal balanced; 2. each added active component (transistor, FET, or valve) in the signal path adds to the distortion and the influence the power supply noise has on the signal (the active part is just modulating the power supply); 3. a good capacitor used in the signal path at key locations allows use of a single supply and avoids #1 & #2. Even when designing valve circuits, it was known to use the fewest active components in the signal path to preserve the original signal integrity. It wasn't just for doing more with less. It was for doing the BEST with less. Edited by BAK - 27 Aug 2021 at 7:56pm
	Bruce AT-14SA, Pickering XV-15, Hana EL, Technics SL-1600MK2, Lautus, Majestic DAC, Technics SH-8055 spectrum analyzer, Eminence Beta8A custom cabs; Proprius & Reflex M or C, Enjoy Life your way!

Graham Slee Members Profile Send Private Message Find Members Posts Add to Buddy List Admin Group Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298	Post Options Post Reply Quote Graham Slee Report Post Thanks(0) Quote Reply Posted: 30 Aug 2021 at 8:49am
	Interestingly, the magnetising current waveform looks like the lower trace here. As load is applied, saturation is reduced, and it must follow that the spikey current waveform should be less sharp. By using a transformer that can keep the amp at full output for hours, at low level there isn't much to tame the spikes. Take a look under the lids of several hi-fi amps and you find transformers that look under-rated. Why should this be? For economy? No! Each quiescent and low level ripple current pulse is quite narrow but quite current hungry pro-rata (see last post page 124). The diode-on period is just prior to the end of the half-cycle. As the load increases, the period extends to the left. Should the load match the transformer power, the saturation will be at its least. But our (music) load is only a watt or less excepting transients. We can either listen louder, or reduce transformer power. It depends on what we want - disco or hi-fi - we can't have both. How it sounds depends on your tastes. If you want loud, then a 225VA transformer is good for 50WPC. Go for a 300VA and it can reach 60WPC, but don't expect the low level delicacy. My next trick is to install a 160VA transformer, and trust I can obtain my target 40WPC. I'm after the low level delicacy that makes it as near to holographic as possible. The 225VA Amplimo almost got there. And as for transformer quality? Well, that seems a difficult and fraught game which I might lose. Terry designed me a 225VA transformer which might work, if somebody decides to help by making me one. And if they don't, well, I shall work with off the shelf transformers, just like every other DIYer.
	That none should be able to buy or sell without a smartphone and the knowledge in how to use apps

Sylvain Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 18 Jan 2010 Status: Offline Points: 481	Post Options Post Reply Quote Sylvain Report Post Thanks(0) Quote Reply Posted: 30 Aug 2021 at 11:45am
	Thank you for the last post.. '''transformers seem to the issue'''' to generate the low level transient we aspire .....but checking the recent Hifi amplifiers design ....Power and big DEDICATED transformers seems to be the DESIGN quest ..''

Graham Slee Members Profile Send Private Message Find Members Posts Add to Buddy List Admin Group Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298	Post Options Post Reply Quote Graham Slee Report Post Thanks(1) Quote Reply Posted: 30 Aug 2021 at 4:51pm
	It was Otto Schade who made it possible to comprehend what was really going on in rectifiers. By the mid-30s the world had them, and knew the basics, but insufficient to be able to model outcomes predictably. Schade was the first to explain rectifiers from the point of the mains having impedance and diodes having resistance, otherwise much of what an oscilloscope tells us would be laughable. Still, mains impedance is not massive, it has to be within bounds to prevent heating of wiring which could otherwise cause fires. And because the mains has impedance, interference can travel along it, interfering with other items which react by adding their own interference. Earlier I fitted one of the 100VA laminated transformers that I tried some pages ago. Telegraph Road off YouTube and Decades by Joe Walsh could have made me believe I had floor standers, except they're the little M20's. Small transformer - big bass - but only 25-29 WPC. Big transformer and 50 WPC on maximum output test (set for 0.1%THD) It is in part trial and error. I started out at 400VA, then it was 300VA, then 200VA (2 paralleled 100VA EI xfmrs), then 225VA. All capable of near or over 50 WPC. Perhaps we can draw up some rules on transformer sizes? Right now it looks like 4 x power, so 4 x 40W gives 160VA - maybe. It would appear that quality is in inverse proportion to quantity. Fast food? No, cook for three years.
	That none should be able to buy or sell without a smartphone and the knowledge in how to use apps

Sylvain Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 18 Jan 2010 Status: Offline Points: 481	Post Options Post Reply Quote Sylvain Report Post Thanks(0) Quote Reply Posted: 30 Aug 2021 at 6:38pm
	4 x 5 metres listening room. My 300B offers 9 watts and more than adequate for good warm bass as only valve can and allows good lower mids and Proprius offers 25W and operate well with very good balance to 6-8ohms Speakers 91db ... I followed the 2years± forum of this project in the hope to retrieve more of the subtlety of '' transience''music info for my close-mike live recordings..... to feel the breath of the saxophonist or Sarah Vaughn live acoustic recordings of teh 60's......that digital cannot reproduce

Graham Slee Members Profile Send Private Message Find Members Posts Add to Buddy List Admin Group Retired Joined: 11 Jan 2008 Location: South Yorkshire Status: Offline Points: 16298	Post Options Post Reply Quote Graham Slee Report Post Thanks(0) Quote Reply Posted: 31 Aug 2021 at 12:23pm
	And now... DC on the signal! Nobody ever explained that one either, did they? Instead, I remember the arguments about capacitor sizes making a difference to the sound, so experiments are taking place. With an unregulated power supply, which are all the rage these past 40 plus years, the HT falls on load, and rises off load. The output capacitor retains its charge... Got it yet?
	That none should be able to buy or sell without a smartphone and the knowledge in how to use apps

1970s Design Indulgence

Cadman Enterprises Limited
7 Kirkstall Close
South Anston
Sheffield S25 5BA
England

+44 (0) 1909 568739
+44 (0) 7947 107188

john at gspaudio.co.uk

1970s Design Indulgence

Cadman Enterprises Limited7 Kirkstall CloseSouth AnstonSheffield S25 5BAEngland

+44 (0) 1909 568739+44 (0) 7947 107188

john at gspaudio.co.uk

Cadman Enterprises Limited
7 Kirkstall Close
South Anston
Sheffield S25 5BA
England

+44 (0) 1909 568739
+44 (0) 7947 107188