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 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 Post Reply Page  <1 253254255256257 345>
Author
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 2022 at 5:50pm
4 transistor Dinsdale open vs closed loop

Second test: AC

Not bad for an open-loop circuit (no negative feedback).

Sensitivity: 500uV

The gain at 200Hz is 87dB, which is 3 or 4 dB better than simulated (that's the "mountain" in the above simulation plot).

No good looking at THD and noise in a circuit that's open-loop!!!

OK then. 10% THD+N at 20Hz, falling to 1.1% at 1kHz, and then falling some more. But with mountain shaped gain, I ask you!

S/N: 64dB at 200Hz, so add gain of 87dB and EIN is -151dB.

...told ya!

Next job is to add some NFB in the form of phono preamp EQ.

That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (2) Thanks(2)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 26 Nov 2022 at 2:20pm
A tale of two big money amplifiers

In the 1800's our knowledge was increased by Volta, Ampere, Ohm, Faraday etc. And then further in the 1900's by Black, Schade, Sziklai, Darlington, Shockley, Lin etc.

But rather than be bothered, because "only fools and horses work", somebody relied on Toshiba, and then Sansui, to do the hard mental work.

But Mitsubishi decided to do the hard mental work. And they were the Japanese. And there are rules of physics that determine what can be achieved, but they're kind rules, and if sought diligently, result in something nice and we value those things.

So who is now a world leading corporation?

All I've done is to take the Mitsubishi two transistor circuit and give it a buffered output, and if the circuit doesn't work, it's my buffered output and not Mitsubishi's two transistor circuit.

The only people in the UK to do the hard mental work, are now forgotten. Nothing is made here anymore. It might say it's made here, and if it's packaged here then you can say it's made here, but otherwise it's wrong.

OK, I don't make transistors or resistors or capacitors or opamps etc., but I enjoy the hard mental work, but more than that, I enjoy sharing it with anybody. But that's not to big myself up. It's about sharing in such a way as to show truth, so others can have a go - the passing down of knowledge, how it used to be, even if limited in value.

Individuals make things happen. And that's done through learning, and the guy at Mitsubishi must have been more educated than me, because he got his a.c. theory right. But one thing we must have both known about is curves of constant noise, and also about balancing input stage current to try and obtain the best of both worlds...

Because you can obtain the lowest noise and have lots of distortion, or the lowest distortion and plenty of noise. But to obtain the best of both with just one transistor - T1 - requires understanding through doing the hard mental work.
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Sylvain View Drop Down
Senior Member
Senior Member


Joined: 18 Jan 2010
Status: Offline
Points: 481
Post Options Post Options   Thanks (1) Thanks(1)   Quote Sylvain Quote  Post ReplyReply Direct Link To This Post Posted: 26 Nov 2022 at 4:33pm
Yes Graham, Allow me my testimony
I have been researching the right kind of amplifier for the kind/genre/ time period/ instruments of Music i listen...and after Valve Triode and expense and Dragons only available Valves ...sought Mosfet to replicate the tonal of Valve tone until 2015 when i purchased Pair of  Proprius.....the sound/the concept of SpeakerAmp/ the Stereo separation achieved/ the neutral atmosphere that Proprius convey, though sometime would wish for a more forthright/ the price bracket with +++premium cannot be bettered ..well considered well built well designed and made for purpose of music .... I sought "Integrated design  " was a compromise And cannot wait for the 70'S Nostalgia concept.
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (2) Thanks(2)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 26 Nov 2022 at 5:15pm
The transistors I'm using are BC337-25's and the Mitsubishi circuit used either 2SC693 or 2SC1000-GR transistors. Each one can be considered as universal small signal transistors. The BC337 has more in common however, with the BC107. The BC107 belonged in a group labelled AF silicon planar epitaxial transistors which also included the BC108 and BC109. The description given to the BC337 is almost the same, except AF is missing. AF meaning audio frequency. They still operate at audio frequency, but they're not made to the exacting standards as the BC107-109 range. It doesn't mean they're no good, they're good enough. The reason for not using the BC107-109 range is they're no longer generally manufactured. They are in India and Farnell stock them under their own brand, but the major semiconductor manufacturers don't bother. They're also quite expensive at over £1 each. The BC337-25, a defined beta range transistor is less than 20p each.

The BC109 had defined noise or noise figure (NF) expressed in dB and typically was 1.2dB at 1kHz within a narrow bandwidth of 200Hz. The BC107 was typically 2dB, so not much worse.

The curves of constant noise were elliptical plots at different noise figures of collector current plotted against source impedance. The source impedance of a moving magnet cartridge is frequency dependent because of its inductance, but we can find it is bounded at the high frequency end as 47k ohms, and often the manufacturer states its resistance which is somewhere between 1k ohm and 3k ohm.

From the curves of constant noise for the BC109, the area bounded by its lowest noise figure for 1kHz, that spans 1k ohm to 50k ohms is with a collector current of 100uA (0.1mA), and for 10kHz, somewhere between 100uA and 200uA.

We can see in the Mitsubishi diagram that the voltage across T1's collector load is about 20V and the resistance is 220k ohms, and so the collector current is just shy of 100uA.

In my diagram I have stepped up the collector voltage by using a level shifter, and so the collector resistance totals 150k ohms (2 x 75k ohms) and the collector current is somewhere in the region of 120uA.

However, as collector current falls, it gets closer to collector cut-off current, and if say it is 1uA then the distortion is going to be around 1/120, all times 100, or 0.833%.

OK, the circuit as a whole has negative feedback which reduces noise and distortion, but it is good to know where we are starting, open-loop.

The "70s nostalgia concept" is in fact, proper engineering with what we had.
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 26 Nov 2022 at 7:00pm
So, what is NF (noise figure) and how does it affect the amplifier.

There are lots of noise sources at or near the input that are going to be amplified along with the wanted signal. 

For example, there's resistor noise at room temperature which is 4nV per 1000 ohms by the sq.rt of frequency (that's 4nV x 141 for the 20kHz bandwidth = 0.564uV).

Noise figure is how much all the noises at the input to the amplifier are going to be increased by prior to gain being applied.

So a 1.2dB NF increases input noise by 1.15 times, and 10dB NF increases input noise by 3.16 times.

If the gain of the circuit is 100 (40dB) then 100 x 0.564uV = 56.4uV, but if the NF is 1.15 times, that's 64.86uV, and if 3.16 times, 178.22uV.

And if the output is 100mV, then with 64.86uV noise the S/N is 1,542:1, or 64dB

And if the output is 100mV, then with 178.22uV noise the S/N is 561:1, or 55dB
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 27 Nov 2022 at 7:08am
The test board has been equipped with the RIAA NFB network shown in the RIAA version of the four transistor "Dinsdale" preamp schematic.


(dig that hard wired network!)

The test board was afforded some far from effective shielding in the form of some folded aluminium sheet, and the input ground routed to the sheet using my meter test leads - about 1 metre in total.


I'm using 'scope probes to connect to the AP audio analyser.

There's 5 foot of side by side twin flex (zip wire) between the PSU1 regulator and the board - as some like to say, no shielding at all.

The board itself is borrowed from power amplifier duties simply because it's convenient as it carries basically the same schematic, but is no way optimised in layout for preamp signals.

That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 27 Nov 2022 at 7:30am
The first thing to check was the RIAA gain frequency response.


The output load is 50k ohms, comprising the on board 100k ohm resistor and the 100k ohm AP termination.

By switching to relative level the RIAA response can be viewed in relation to 0dB and 1kHz.


In the first image it can be seen that 100Hz is at approx. 13.6dB (delta box on right) and this is 0.5dB high compared with the RIAA replay response for 100Hz. Otherwise, the entire RIAA curve is pretty accurate. So, maximum deviation is 0.5dB, and some would call that +/- 0.25dB which doesn't look as bad... But can you hear the difference? Doubtful.

Next thing was to look at the gain.


So for 5mV input (ref 1kHz) the output to the "tape-loop" and volume control is 400mV instead of the block diagram 500mV. And that means the tape monitor input that should match, is 400mV also. This is an IEC 61938 misnomer. Having said that, it will probably be a good match for a number of stereo cassette decks, so I think I'll keep it like this.
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
 Post Reply Post Reply Page  <1 253254255256257 345>
  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 0.125 seconds.