Not the first, they usually go one step further and use the
typical "cheap" (but superior) IC/darlington combo.
I work on "cheap" all the time. The little compact LP/CD/Tape
unit in my garage ("cheap- CHEAP")uses discretes but actually has a CCS. (lennox sound
).I worked in RS's service shop, but never saw this particular
unit , most of RS's discrete receiver amplifiers were
typical japanese class b's with CCS's , undersized power supplies's .. But at least biased to give acceptable sound for
the systems they were sold with.
OS
You guys, think too much. RS didn't think about it as much. Of course it could have been improved and sounded better. It also would be more expensive, IF the trick input stage was left in, and marginal if it had been left out. IT is a 40 year old design, please understand this. What were you designing 40 years ago?
What you must understand, is that the RS designers are essentially 40 years behind in experience with a comp differential input stage. IF and when they have MORE experience, the resistors will be either filtered or current sources will be added. It isn't as big an error as all that, and I made circuits like that in the 1968-1969 era at AMPEX with resistors as the current sources on comp diff. input stages. They worked OK, BUT I started to use jfet current source diodes in 1970, with designs for the Grateful Dead, and maybe, even back in late 1969. It is a natural progression in design. It is also more expensive.
This is a ST&S amp inside Accrurain sold by Radio Shack.
I have almost completed mods (change 2.7k bias to 2.0 bias which upped the
bias from 0 to 5mV to 10mV. . .depending on air temp). Also will change input
resistor and bypass resistor. . .changed 10 microF to 1 microF already.
Will probably tone down the 20kHz by a few dB.
I have almost completed mods (change 2.7k bias to 2.0 bias which upped the
bias from 0 to 5mV to 10mV. . .depending on air temp). Also will change input
resistor and bypass resistor. . .changed 10 microF to 1 microF already.
Will probably tone down the 20kHz by a few dB.
A symmetrical input stage is nice if you already have a good amp, but when you start with something like this, it makes sense to go for the things that make a big difference, like current sources, rather than what looks pretty on the schematic.john curl said:
Bootstrap current sources. Not as good as active current sources, but a massive improvement on resistors, and simple to implement too.MJL21193 said:
Mr Evil said:
A strong case was made in the BT thread that a symmetrical LTP would be measurably worse than a single ended LTP. In the symmetrical case, the difference currents necessary to balance the two halves can unbalance the Vas stage, where it all comes together. That problem doesn't exist in the unsymmetrical LTP.
Jan Didden
I may have fewer decades of experience than you, but I've never seen any case where a symmetrical LTP gave an improvement anywhere near that given by using current sources. Despite the appeal, the lack of real symmetry between NPN and PNP devices prevents the promised distortion cancellation from being effective. If you have any reason why I'm wrong, then spit it out and help educate everyone who reads the thread, don't resort to ad hominem.john curl said:
This is not entirely correct. IF you actually build a circuit, and attempt to make it fast enough, the complementary drive of the second stage, compared with what the RS design would be like, if you removed the complementary differential, will be of great advantage.
What is being confused here is the use of a differential pair on the second stage and a current mirror. This does not exist in this design, and it would be expensive and more complex as well.
Of course, you are correct, that current sources would help, but perhaps improving the power supply regulation would help as well, and over a wider range of components. RS chose not to do that, as well, in order to save pennies, which is the design goal of this product. This is how consumer stuff is made, and even a small increase can exceed the budget.
Apparently the designer went for comp diff input, like the 'BIG BOYS' perhaps as a marketing tool, and ran out of pennies to throw at the design.
I have had to do designs like this in the past, and even a few extra resistors and caps can seem excessive. Heck, why use a differential at all? We didn't, back in 1968. At that time, a current source would have seemed a ridiculous luxury.
What is being confused here is the use of a differential pair on the second stage and a current mirror. This does not exist in this design, and it would be expensive and more complex as well.
Of course, you are correct, that current sources would help, but perhaps improving the power supply regulation would help as well, and over a wider range of components. RS chose not to do that, as well, in order to save pennies, which is the design goal of this product. This is how consumer stuff is made, and even a small increase can exceed the budget.
Apparently the designer went for comp diff input, like the 'BIG BOYS' perhaps as a marketing tool, and ran out of pennies to throw at the design.
I have had to do designs like this in the past, and even a few extra resistors and caps can seem excessive. Heck, why use a differential at all? We didn't, back in 1968. At that time, a current source would have seemed a ridiculous luxury.
Great discussion. . . Remember this design is ST&T not Radio Shack -- but similar
business plan.
I have two points:
1) Would R209 + R205 & R207 be replaced by 2 transistors and resistor (CCS).
2) Replacing C201 from 10 microF to 1 microF should only change the low response
by rolling it off earlier. Sorry about the typo.
10 microF at 200Hz is 0.02dB compared to 1kHz
1 microF at 200Hz is -0.44dB compared to 1kHz
10 microF at 20Hz is -0.58dB compared to 1kHz
1 microF at 20Hz is -4.64dB compared to 1kHz
business plan.
I have two points:
1) Would R209 + R205 & R207 be replaced by 2 transistors and resistor (CCS).
2) Replacing C201 from 10 microF to 1 microF should only change the low response
by rolling it off earlier. Sorry about the typo.
10 microF at 200Hz is 0.02dB compared to 1kHz
1 microF at 200Hz is -0.44dB compared to 1kHz
10 microF at 20Hz is -0.58dB compared to 1kHz
1 microF at 20Hz is -4.64dB compared to 1kHz
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.