if there are 1 pair, 3 pair, 5 pair and 14 pair....
what are the disadvantages? what happens if they are not matched.
the reason why this question popped out is I see alot of amplifiers based on LME49810 ebay which is sold for about less than 100 USD i really dont think they are matched transistors..
now if they are not matched how does it reflects the performance and stability?...
what are the disadvantages? what happens if they are not matched.
the reason why this question popped out is I see alot of amplifiers based on LME49810 ebay which is sold for about less than 100 USD i really dont think they are matched transistors..
now if they are not matched how does it reflects the performance and stability?...
Source resistors will help share the current. In most amplifiers, unmatched transistors is usually not an issue. I have never matched my transistors.
The disadvantage of unmatched paralleled transistors is in the "coherency" of the sound, where single pair is better in this area because of no matching issue. But paralleled transistors will give lower output impedance or more ability to drive speaker or better bass.
So, parallel is most often better than a single pair. But matched paralleled is even better, but requires too many transistors to test. I only have done and experienced this with mosfet output.
Darlington transistor will have much greater issue when paralleled because it consists of 2 transistors in one package. But running high current should reduce the issue imo.
The disadvantage of unmatched paralleled transistors is in the "coherency" of the sound, where single pair is better in this area because of no matching issue. But paralleled transistors will give lower output impedance or more ability to drive speaker or better bass.
So, parallel is most often better than a single pair. But matched paralleled is even better, but requires too many transistors to test. I only have done and experienced this with mosfet output.
Darlington transistor will have much greater issue when paralleled because it consists of 2 transistors in one package. But running high current should reduce the issue imo.
People matching small signal transistors might buy 100 for a pair or a quad, its just not sensible to do that with power transistors, a commercial amp that did that would have to sell for tens of thousands. It would be a revelation if matching was the best use for an extra X hundred Pounds/Dollars/Euros in parts, normally people go for PSU and parts upgrades, Hifi brands also go for flashier cases and assorted methods of persuasion.
Maybe amp manufacturers that use outputs with multiple chip amps, like Jeff Rowland, are economically matching their output transistors by passing the cost of doing so back down to the IC supplier.
Maybe amp manufacturers that use outputs with multiple chip amps, like Jeff Rowland, are economically matching their output transistors by passing the cost of doing so back down to the IC supplier.
Hi Guys
You can look at Doug Self's example:
None of the designs in his books or articles use matched devices and the results are pretty good. When he doubles up or expands the output stage with more pairs, then he adds emitter resistors for current sharing. The fact is that for BJTs, the turn-on characteristic is very uniform from one sample to another, and even from one device to another.
When you order BJTs from a supplier, you will get ones from the same rail and thus the same batch, and this is usually matched as close as you need for 99% of circuits.
Small-signal transistors should be matched for lowest distortion, particularly in symmetric circuits. If they are not matched, then hopefully there is emitter degeneration in use to reduce the sensitivity to device parameter spreads. Leach says he never matched any devices in his circuits and had pretty good performance.
Note that for output BJTs, higher Re will enforce better current sharing but will increase output impedance. Having 14 pairs versus 2 pairs allows the multi-pairs to smooth out variations so the push-pull sides are likely to be more balanced.
With proper cooling and protection circuit design, nearly the full rating of the output device can be utilised on a peak basis.
Have fun
Kevin O'Connor
You can look at Doug Self's example:
None of the designs in his books or articles use matched devices and the results are pretty good. When he doubles up or expands the output stage with more pairs, then he adds emitter resistors for current sharing. The fact is that for BJTs, the turn-on characteristic is very uniform from one sample to another, and even from one device to another.
When you order BJTs from a supplier, you will get ones from the same rail and thus the same batch, and this is usually matched as close as you need for 99% of circuits.
Small-signal transistors should be matched for lowest distortion, particularly in symmetric circuits. If they are not matched, then hopefully there is emitter degeneration in use to reduce the sensitivity to device parameter spreads. Leach says he never matched any devices in his circuits and had pretty good performance.
Note that for output BJTs, higher Re will enforce better current sharing but will increase output impedance. Having 14 pairs versus 2 pairs allows the multi-pairs to smooth out variations so the push-pull sides are likely to be more balanced.
With proper cooling and protection circuit design, nearly the full rating of the output device can be utilised on a peak basis.
Have fun
Kevin O'Connor
Many of the questions you ask here have been answered so many times that people have written books and webpages about them. What's more, the answers likely are better thought out than just good guesses or suggestions guys try to offer when they are uncertain what you are really after.
Why not Search or Google your questions first before starting dozens of threads with the same questions others ask here regularly? The right answers and discussions are just seconds away. If you still need more information, then post in an old thread that is close to what you want so the info. remains together for others to search conveniently. Here's most of what you need to know about matching; Matching Power and Driver Transistors
In fact, if you read that site systematically, you'd probably learn enough to tackle anything you might want to DIY.
Why not Search or Google your questions first before starting dozens of threads with the same questions others ask here regularly? The right answers and discussions are just seconds away. If you still need more information, then post in an old thread that is close to what you want so the info. remains together for others to search conveniently. Here's most of what you need to know about matching; Matching Power and Driver Transistors
In fact, if you read that site systematically, you'd probably learn enough to tackle anything you might want to DIY.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.