There is a phenomenon called electromigration that displaces the atoms of the wires on a chip, eventually leading to broken wires. This phenomenon goes faster at higher temperatures and at higher currents. That 20 mA is probably the current at which the device lifetime and reliability targets of the manufacturer are still met at the highest specified junction temperature. If you can keep the junction temperature well below the specified maximum, you can handle somewhat more current - assuming that my educated guess that electromigration is the limiting factor is correct.
Does the transistor still work well at 35 mA, or do the hFE or the fT collapse?
Seems to me, that your hearing is somehow impaired. I've built at least 30 different amplifiers in past 20 years. Simple LTPs even with Japanese twin transistors sound inferior to properly cascoded ones. Not to mention more complex LTP circuits. Those sounds even better. Are you listening high (natural(!)) dynamic range music or only dinamically compressed "modern" songs? For the latter you'll never hear any difference between circuits because the RMS of these music is nearly constant and not excite dynamic thermal changes soo much. I think one of the reason of the so called Loudness Race is the Thermal distortion of bad circuits. Please listen to some SACD or 80's pop music from original CDs or FLAC. The new CD issues are compressed to sh@t so please search for original CDs.
Last edited:
To me, however, it seems that it has not been solved even now, and that is why many people use electron tube circuitry instead of LTP. With, say, 0.1% THD, it still sounds better than plain LTP because thermal distortion with electron tubes is inherently lower without complex circuits.
Last edited:
Thanks Nigel!
I think one of the reasons why MAPLIN 75WRMS sounds good is that the thermal distortion of lateral MOSFETs is lower and therefore LTPs need to compensate less. So an output stage amplifier with lateral mosfets probably sounds good even with plain LTP. In lateral mosfets, the current is always evenly distributed over the entire volume of silicon, and thus the piece of silicon used effectively has a larger volume than in other MOSFETs or BJTs. The thermal cutoff frequency of lateral MOSFETs is likely to be lower. And you don’t even have to install a complex heat compensated bias circuit.
Yes, it's a BJT cascoded CFP with JFET voltage reference for the cascode BJTs. The patent of the input stage is here, pdf can be downloaded. [ Amplifier stage with low thermal distorton - Aug. 10, 1993]
It would have been better to call this kind of distortion memory-type distortion. Maybe then there wouldn’t be so much opposition or doubt that it exists.
Like maxlorenz, I say from my practical experience that it clearly exists. It can be measured with suitable measuring circuits.
Measurement with constant spectrum signals, such as e.g. standard THD and IMD measurements do not show it because long-term averaging takes place, but this kind of distortion occurs when the constant-spectrum signal starts and stops. Many times, they deliberately wait for the output signal to be measured to set before starting data collection. This is the case for either automatic or manual measurements. Distortion, which is important for beautiful sound, is thus eliminated from these measurements. Therefore, there is no correlation between THD and hearing tests.
Like maxlorenz, I say from my practical experience that it clearly exists. It can be measured with suitable measuring circuits.
Measurement with constant spectrum signals, such as e.g. standard THD and IMD measurements do not show it because long-term averaging takes place, but this kind of distortion occurs when the constant-spectrum signal starts and stops. Many times, they deliberately wait for the output signal to be measured to set before starting data collection. This is the case for either automatic or manual measurements. Distortion, which is important for beautiful sound, is thus eliminated from these measurements. Therefore, there is no correlation between THD and hearing tests.
Last edited:
Using common source output looks interesting to me for a number of reasons (including lower thermal/mem. distortion in the driving circuit). Do you know any examples of drain output other than GB150D?
Here is one schematic to look at, just to see something unusual. It is a drain output and similar to seletronic profet / pass F5. There's a lot of a marketing gimmick on this one, for example they claim "no global feedback", but the design has separate preamp and power amp sections and both have global feedback.
https://www.kendeil.com/Portals/0/AM-Audio.pdf
Here is one schematic to look at, just to see something unusual. It is a drain output and similar to seletronic profet / pass F5. There's a lot of a marketing gimmick on this one, for example they claim "no global feedback", but the design has separate preamp and power amp sections and both have global feedback.
https://www.kendeil.com/Portals/0/AM-Audio.pdf
Thanks for the interesting circuit! From a cap manufacturer's marketing site
Not sure i agree with your view. Yes, the input side has its own nfb loop and so does the output but a global loop is missing, exactly as per spec.
Very curious how a common drain output with no loop nfb compares to a common source with some loop nfb soundwise.
This kind of discussions, sadly reappearing too often, remind me of:
How many angels can dance on the head of a pin? - Wikipedia
How many angels can dance on the head of a pin? - Wikipedia
Hi Alexandre,
do you mean the output is fed by the MOSFET source (or sources for push-pull)?
Likewise, it is a term I have not seen before so need to make sure what you mean.
Your question above seems strange because 99% of MOSFET power amps have the output fed by the MOSFET sources, a.k.a. source followers or usually just called "followers".
Or did you intend to ask "Any other power amps with common drain out?"
Just to clear up what you mean by "common source out"
do you mean the output is fed by the MOSFET source (or sources for push-pull)?
Likewise, it is a term I have not seen before so need to make sure what you mean.
Your question above seems strange because 99% of MOSFET power amps have the output fed by the MOSFET sources, a.k.a. source followers or usually just called "followers".
Or did you intend to ask "Any other power amps with common drain out?"
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.