Who cares about the Early Effect in a LTP concerning PSRR ?
How well two Bjts from the same batch have matching Early voltage ?
How well two Bjts from the same batch have matching Early voltage ?
Last edited:
Just exclude resistive LTP load via any kind of current mirror and, of course, put a high-input resistance buffer before next common emitter stage.
First - you will pick at least 80 dB gain from that three stages, second - due to high feedback depth voltage deviation at LTPs collectors will be neglible.
Feedback is a cure for Early effects in an input low-current, low-voltage deviation stages.
The main distortion source in the amplifier is Early effect in the last VAS stage where huge voltage deviation are really couldn’t be excluded.
I am aware of what is usualy done around an LTP used in a power amplifier.
In my design, using a pair of transistors, I have good reasons to keep low resistor collector values.
My point here, is to learn about the statistics on the Early Voltage of BJTs from a same production batch.
In my design, using a pair of transistors, I have good reasons to keep low resistor collector values.
My point here, is to learn about the statistics on the Early Voltage of BJTs from a same production batch.
Buy a batch and measure them yourself; depending upon which devices you choose, the cost could be as low as $9.92 for 100 parts. example.
To maximize the probability of getting parts from the same manufacturing "lot", I suggest you buy components in "Ammo Pack" and/or Tape And Reel packaging, rather than loose devices a/k/a "bulk".
To maximize the probability of getting parts from the same manufacturing "lot", I suggest you buy components in "Ammo Pack" and/or Tape And Reel packaging, rather than loose devices a/k/a "bulk".
Indeed, this will be the only way to figure this out, if nobody has gone before or dare to publish it. This asks for a test jig dedicated to extract the Early voltage.
Beta/Early voltage product is roughly a process related parameter, i.e. matching beta will tend to get you matched early voltage.
The main source of distortion in modern amplifiers is the output stage. Amplifier theory is advanced that Input and VAS distortions are negligible when compared to the output stage distortion in any modern design. It's the output stage crossover distortion that is the major contributor to any competently designed modern amplifier.
Very interesting.
OTOH What about Vbe. Will matching beta tend to get matched Vbe voltage ?
No, too weak of a relationship. Beta relies on recombination in the base which depends on base width so modulation of base width (Early effect) and beta are tied together. Vbe is more related to doping levels and current density (area).
Wow, more interesting.
The relashionships of BJT parameters and the reasons why. Thanks, this is right on what I am looking for.
Is it possible to go a step further about Beta versus Va ( Early voltage ) with a quantitative relashionship. Stating for instance: A +5% Beta goes with a +5%Va ?
The relashionships of BJT parameters and the reasons why. Thanks, this is right on what I am looking for.
Is it possible to go a step further about Beta versus Va ( Early voltage ) with a quantitative relashionship. Stating for instance: A +5% Beta goes with a +5%Va ?
On my amps I used extra front end decoupling to get around PSRR problems in LTP abd VAS. Usually 100r with 220uF works well on each rail.
Sorry it's not that specific, you would need to consult the standard texts on this like Sze.
I got interested in Early voltage a while back, and published some, not all, of my measurements in this thread:
Measuring the Early voltage "VA" of some PNPs at 8mA
Lots of people made the same bold assertion in follow up postings: (Beta times EarlyVoltage) is a constant!! They are inversely proportional!! It's all directly related to VceMax. Duh!!
So I created a scatterplot of my measured data. See below. I only measured one transistor of each part#; if that's not what you were hoping for, I am sorry.
For the transistors that I measured, Vearly varied over about a 12-to-1 range, while (Beta*Vearly) "only" varied over a 6-to-1 range. Which is not a very good approximation of a constant, in my opinion.
_
Measuring the Early voltage "VA" of some PNPs at 8mA
Lots of people made the same bold assertion in follow up postings: (Beta times EarlyVoltage) is a constant!! They are inversely proportional!! It's all directly related to VceMax. Duh!!
So I created a scatterplot of my measured data. See below. I only measured one transistor of each part#; if that's not what you were hoping for, I am sorry.
For the transistors that I measured, Vearly varied over about a 12-to-1 range, while (Beta*Vearly) "only" varied over a 6-to-1 range. Which is not a very good approximation of a constant, in my opinion.
_
Attachments
Same manu same fab same process or there are just too many variables. Never said anything about VceMax.
Last edited:
In late 2016 the folks making bold assertions, did not include you, Scott. I suspect you would be unlikely to propose a numerical rule of thumb which applied to all devices made by Toshiba, Central Semiconductor, Siemens, NEC, and Motorola.
Early Voltage distortion?? Just Cascode and forget about it.
😉
And you’ll find Early-related distortion in emitter-drived cascode device.
😀
There are different crystal spec’s responding for Beta and Va.
Really severity of Early effect could be obtained for specific trannie model by looking at its static characteristic.
Low-early:
Moderate to high Early:
Yes, and your graphs can even be used easily to obtain the early voltage. Just extend a Vce/Ic line to the left and see where it crosses the neg voltage axis. That's your early voltage.
(The graph is from Peter Baxandall in Linear Audio | your tech audio resource and illustrates the early voltage distortion effect)
Jan
(The graph is from Peter Baxandall in Linear Audio | your tech audio resource and illustrates the early voltage distortion effect)
Jan
Attachments
Last edited:
- Status
- Not open for further replies.