So, you just used the 10 ohm base stop resistor's in there
because it was recommended?
or
You didn't use 10 ohm resistors and were fine, then 8 generations later, you started using them.
or
You used the 10 ohm base stoppers and didn't think they did much.
Then you omitted them or started using smaller values until things
started blowing up...leading to your use of them again, albeit at
a smaller value, 5 ohms perhaps?
Question
The base stopper helps prevent thermal runaway in amps?
It is appear these also help control output as well. We are talking
We have to be talking about some hi wattage resistors here aren't
we? I assume you want these heat synched away from the transistors.
In part, this also helps me understand why it is important to Strap/ thermally balance the output transistors.
Question
Does it matter much if rectifying devises are thermally balanced? For example, hexfreds in a bridge rectifier?
Thanks,
Sync
because it was recommended?
or
You didn't use 10 ohm resistors and were fine, then 8 generations later, you started using them.
or
You used the 10 ohm base stoppers and didn't think they did much.
Then you omitted them or started using smaller values until things
started blowing up...leading to your use of them again, albeit at
a smaller value, 5 ohms perhaps?
Question
The base stopper helps prevent thermal runaway in amps?
It is appear these also help control output as well. We are talking
We have to be talking about some hi wattage resistors here aren't
we? I assume you want these heat synched away from the transistors.
In part, this also helps me understand why it is important to Strap/ thermally balance the output transistors.
Question
Does it matter much if rectifying devises are thermally balanced? For example, hexfreds in a bridge rectifier?
Thanks,
Sync
The 10 ohm resistors did NOT cause any obvious problem for about 20 years of implementation, or 8 generations of power amps. However, the potential was there, with poor beta mismatching or poor thermal management. The early JC-1 had both. This was fixed, and the 10 ohm resistors remain. SOME resistance or its equivalent must be in the base to avoid high frequency oscillation. How much, is the question.
jam said:
Hi Jam.
In general terms, a smaller value of emitter resistor is worse for matching issues, both Vbe matching and beta matching. In general, smaller emitter resistors tend to reduce both global thermal stability (i.e., like heat sink temperature and feedback to the Vbe multiplier) and local thermal stability (like current hogging among paralleled transistors and thermal runaway that can happen faster than the heat sink can react to it).
However, the use of more agressive values of emitter resistors (i.e., smaller values) often reduces crossover distortion.
BTW, the optimum class AB biasing approach that Barney Oliver came up with and that Doug Self discussed, is largely based on an ideal transistor. Unfortunately, the pesky base spreading resistance and any base stopper resistance is transformed by transistor beta into an equivalent ohmic resistance in series with the emitter. This tends to mean that the optimum voltage drop across the real RE will be less than the 26 mV number that is often quoted.
For better thermal stability, the resistance seen looking into the transistor would like to be significantly larger than the sum of base spreading resistance and base stopper resistance. For example, if you have transistor beta of 50, bias current of 125 mA and RE of 0.22 ohms, the resistance seen looking into the transistors intrinsic base (i.e., ignoring base spreading resistance) is about 25 ohms, if I've done my math right. In that case, if base spread resistance is 5 ohms and base stopper is 5 ohms, for a total of 10 ohms, things are not too bad. The transistor feels like it is more voltage driven (bias-wise) than current driven.
This simplified explanation assumes that the dynamic resistance of the driver EF is low in regard to DC biasing issues.
When a transistor is more voltage driven, beta and beta matching will have less effect.
Larger base stopper resistance makes achieving thermal stability more difficult because it makes the transistor somwhat more in the direction of being current driven, increasing beta-dependent effects.
I hope this helps, but I know the explanation is a bit over simplified.
BTW, base spreading resistance is often poorly modeled for power transistors because it is easy to confuse with a non-deal slope (NE=1.0) in the collector current Gummel plot. Accurate Ic vs. Vbe measurements usually need to be done down to quite low currents (for a power transistor) in order to separate out the two effects.
I have often quoted base spreading resistance as being on the order of 5 ohms at collector currents representative of quiescent bias, but I could be wrong. BTW, base spreading resistance decreases at higher collector current due to emitter crowding.
Cheers,
Bob
Hi Bob,
From personal experience, I can say that matching the output transistors does bring about a reduction in THD. It had been suggested before that allowing a mismatch here would allow for lower THD. On-Semi has made a point of this fact in their literature here and there.
The value of emitter resistance would seem to have a greater influence on sharing the load at higher currents. When you are passing lower currents, the beta match between output transistors has a greater effect on current sharing, right in the crossover region. I wonder how many designs suffer from mild current shoot through due to mismatched output transistors? Anyway, this seems to support what you are saying.
Regarding the LT-1166, I did service an expensive Italian integrated amplifier once that used these. They had removed the lettering on their "secret" chip, but it's function was pretty clear. After this introduction to the LT-1166, I determined that I'd never use one. They didn't optimize the application for lower THD and the data sheet didn't mention this, so no matter how attractive the idea was, it didn't perform well.
Another pint about this amplifier design. It seems that these ICs were the reason why the amplifier failed. The LT-1166 in the other channel had also been replaced earlier in time.
Just goes to show you how important application information is when you are trying to sell parts. Not everyone has the time to figure out the unknown features if the part does not perform well in it's basic configuration.
I'd like to thank you for explaining the finer points on the base resistors compared to the emitter resistor value. It seems that Yamaha went through a period where they would lose amplifiers often, they were using 0.22R non-inductive (plate) emitter resistors and larger base resistors. I think they finally settled on 4.7 R resistors in the end. So this makes sense now.
-Chris
From personal experience, I can say that matching the output transistors does bring about a reduction in THD. It had been suggested before that allowing a mismatch here would allow for lower THD. On-Semi has made a point of this fact in their literature here and there.
The value of emitter resistance would seem to have a greater influence on sharing the load at higher currents. When you are passing lower currents, the beta match between output transistors has a greater effect on current sharing, right in the crossover region. I wonder how many designs suffer from mild current shoot through due to mismatched output transistors? Anyway, this seems to support what you are saying.
Regarding the LT-1166, I did service an expensive Italian integrated amplifier once that used these. They had removed the lettering on their "secret" chip, but it's function was pretty clear. After this introduction to the LT-1166, I determined that I'd never use one. They didn't optimize the application for lower THD and the data sheet didn't mention this, so no matter how attractive the idea was, it didn't perform well.
Another pint about this amplifier design. It seems that these ICs were the reason why the amplifier failed. The LT-1166 in the other channel had also been replaced earlier in time.
Just goes to show you how important application information is when you are trying to sell parts. Not everyone has the time to figure out the unknown features if the part does not perform well in it's basic configuration.
I'd like to thank you for explaining the finer points on the base resistors compared to the emitter resistor value. It seems that Yamaha went through a period where they would lose amplifiers often, they were using 0.22R non-inductive (plate) emitter resistors and larger base resistors. I think they finally settled on 4.7 R resistors in the end. So this makes sense now.
-Chris
Hi Anatoliy,
No, but voltage drops at high, peak currents may matter. You can only accept so much loss in the output circuit.
I do know that what we are talking about is less than 1 dB, but I wonder how the circuits react to a wildly fluctuating voltage between emitter and output buss. That depends greatly on the application.
-Chris
No, but voltage drops at high, peak currents may matter. You can only accept so much loss in the output circuit.
I do know that what we are talking about is less than 1 dB, but I wonder how the circuits react to a wildly fluctuating voltage between emitter and output buss. That depends greatly on the application.
-Chris
john curl said:
john curl said:
It's funny that both quotes are from the same author.
Read chapter3 of "Wideband Amplifiers" by Peter Staric and Eric Margan,
<http://www.amazon.de/Wideband-Ampli...ie=UTF8&s=books-intl-de&qid=1249767784&sr=1-1>
They used to be with Tektronics, (scope Y-amplifiers and such).
Sorry, this is RF.
regards, Gerhard
Hi Gerhard,
Sorry, but it's way over my pay grade for one.
$234.50 CDN, might get one for $176.85 CDN. Used, lowest is $158.34 CDN -USED!
It's a book, just a book.
-Chris
Edit: You know, if it weren't for the price, the contents suggest that it is a very useful book. Still, it is just a book.
Sorry, but it's way over my pay grade for one.
$234.50 CDN, might get one for $176.85 CDN. Used, lowest is $158.34 CDN -USED!
It's a book, just a book.
-Chris
Edit: You know, if it weren't for the price, the contents suggest that it is a very useful book. Still, it is just a book.
Hi Anatoliy,
I see this as overhead. Wasted as heat.
Having said that, I'm all for keeping bias currents under control. I just don't want to waste more than really required to keep a lid on an amp.
Thermal feedback is being .... well, kind. Something a "spin doctor" might say. Just humor, no insult intended here.
-Chris
I see this as overhead. Wasted as heat.
Having said that, I'm all for keeping bias currents under control. I just don't want to waste more than really required to keep a lid on an amp.
Thermal feedback is being .... well, kind. Something a "spin doctor" might say. Just humor, no insult intended here.
-Chris
Hi Gerhard,
I still have problems when the price of a technical book is very high. It removes the knowledge from the grasp of those who are still learning and young. It's like a class barrier (I did a report on that very thing here in Canada). Remembering back to when I was going to Ryerson in Toronto, the books and tuition were so expensive, I had to work every second year. Even then the books were a real stretch for me. I know others who couldn't swing it.
Later in life, I see things in a similar way. Now here is the question that begs an answer. How much do the authors get in royalties, and how much does the publisher rake in? It is what it is, but still effectively a class barrier.
I will see if my library can order it in. Now there is the question.
-Chris
You know, I do believe you on that. If I was able to work, then I might be able to swing it.
I still have problems when the price of a technical book is very high. It removes the knowledge from the grasp of those who are still learning and young. It's like a class barrier (I did a report on that very thing here in Canada). Remembering back to when I was going to Ryerson in Toronto, the books and tuition were so expensive, I had to work every second year. Even then the books were a real stretch for me. I know others who couldn't swing it.
Later in life, I see things in a similar way. Now here is the question that begs an answer. How much do the authors get in royalties, and how much does the publisher rake in? It is what it is, but still effectively a class barrier.
I will see if my library can order it in. Now there is the question.
-Chris
- Status
- Not open for further replies.