good evening DIYers,
i just made an 2 channel amplifier with complementary bjt output. i just have one problem. channel 2 is giving 30mv and the channel 1 is giving 238mv at the speaker terminals. after i counter checked the connections and components of channel 1 with channel 2, and all was well. i tested for the second time and still got the same 238mv on channel 1. now i read an article that was produced by Rod Elliot. He said that any output voltage below 100mv is fine and any above that is exessive even though some amplifiers are seen to work well at output voltages exceeding 100mv. i tried changing the bias voltage buh still there was no change. what could be the problem?
i just made an 2 channel amplifier with complementary bjt output. i just have one problem. channel 2 is giving 30mv and the channel 1 is giving 238mv at the speaker terminals. after i counter checked the connections and components of channel 1 with channel 2, and all was well. i tested for the second time and still got the same 238mv on channel 1. now i read an article that was produced by Rod Elliot. He said that any output voltage below 100mv is fine and any above that is exessive even though some amplifiers are seen to work well at output voltages exceeding 100mv. i tried changing the bias voltage buh still there was no change. what could be the problem?
Last edited:
That's a hard question to answer without knowing what amplifier it is. Can you post a schematic? Bias has nothing to do with DC offset. Does the front end have a differential transistor pair?
Craig
Craig
I would imagine that's low enough anyway, but as already suggested post the schematic - it's normally pretty easy to add an adjustment (or you might just need to swap a couple of transistors round).
Hi Shelah,
This sounds like a classic case of input transistors that were not matched. If your amplifier uses a differential pair of transistors for the input stage, it is very important that they are matched in Beta. This stage compares the input signal with the output signal and performs subtraction to generate a correction to reduce the distortion. DC offset is just one form of correction the input stage performs.
One thing you can read into what was just said is that your distortion performance is much improved with a matched pair of transistors. This is a true statement.
One other thing can contribute to higher DC offsets. Low beta transistors in that differential pair will have higher base current (if they are BJTs). This current drops voltage across the resistor between the base and signal ground, subtracting from the inverting side. The voltage between the inverting transistors base and speaker output (Assuming there is a DC blocking capacitor between that base plus resistor and signal ground). Mismatched resistor values will exacerbate differences in base currents.
It can be difficult to match transistor beta because transistors are very sensitive to temperature. In order to test the beta, the different transistors will need to be at the same temperature. All you can do is test and leave the parts in there long enough to stabilize their temperature. Once you have a matched pair, they need to be held in contact with each other to maintain the same temperature across both of them.
Sorry for the long post, but it's important to know the why and how as well as the reason for your DC offsets. The numbers you posted are quite common for mismatched transistors.
-Chris
This sounds like a classic case of input transistors that were not matched. If your amplifier uses a differential pair of transistors for the input stage, it is very important that they are matched in Beta. This stage compares the input signal with the output signal and performs subtraction to generate a correction to reduce the distortion. DC offset is just one form of correction the input stage performs.
One thing you can read into what was just said is that your distortion performance is much improved with a matched pair of transistors. This is a true statement.
One other thing can contribute to higher DC offsets. Low beta transistors in that differential pair will have higher base current (if they are BJTs). This current drops voltage across the resistor between the base and signal ground, subtracting from the inverting side. The voltage between the inverting transistors base and speaker output (Assuming there is a DC blocking capacitor between that base plus resistor and signal ground). Mismatched resistor values will exacerbate differences in base currents.
It can be difficult to match transistor beta because transistors are very sensitive to temperature. In order to test the beta, the different transistors will need to be at the same temperature. All you can do is test and leave the parts in there long enough to stabilize their temperature. Once you have a matched pair, they need to be held in contact with each other to maintain the same temperature across both of them.
Sorry for the long post, but it's important to know the why and how as well as the reason for your DC offsets. The numbers you posted are quite common for mismatched transistors.
-Chris
A schematic can be very helpful.
In your negative feedback, do you use AC grounding or you do DC negative feedback. If you use DC negative feedback, you have a lot of DC gain, any mismatch at the input LTP will be amplified.
If you use complementary IPS, it is possible to have that kind of offset even if it is AC negative feedback. In fact I was just mentioned to Mr. Bob Cordell in his thread two weeks ago and he acknowledge it can happen. The problem is not even mismatch of the LTP, it's the mismatch of the tail current of the two complementary LTPs. Go to Mr. Cordell's thread, back like 6 pages and you'll see the discussion.
In your negative feedback, do you use AC grounding or you do DC negative feedback. If you use DC negative feedback, you have a lot of DC gain, any mismatch at the input LTP will be amplified.
If you use complementary IPS, it is possible to have that kind of offset even if it is AC negative feedback. In fact I was just mentioned to Mr. Bob Cordell in his thread two weeks ago and he acknowledge it can happen. The problem is not even mismatch of the LTP, it's the mismatch of the tail current of the two complementary LTPs. Go to Mr. Cordell's thread, back like 6 pages and you'll see the discussion.
Hi Shelah,
In this circuit where you have complimentary diff pairs, matching all four (very time consuming, need lots of transistors to pick from) makes a big difference in performance. Any imbalance of Beta will cause DC offset. So, get all four as close as you can.
This general form of input was popular with SAE and Phase Linear just to name a few. Many years of experience with these has taught me that there is no other way to make these sound great other than doing it the right way. I must have bought thousands of those transistors (2N5551 and 2N5401). I still have over 100 of each in order to find matched quads. Get the like transistor types as close as you can, then find the closest matched other pair. Expect to achieve about 5 mV DC offset, <10 mV really if the diff pairs are properly matched. This also improves the sound quality of the circuit.
Hi Alan0354,
Unequal base current is another way of saying non-matched Beta in these circuits. Attempting to unbalance a pair might help with DC offset, but will hurt in the distortion department.
-Chris
In this circuit where you have complimentary diff pairs, matching all four (very time consuming, need lots of transistors to pick from) makes a big difference in performance. Any imbalance of Beta will cause DC offset. So, get all four as close as you can.
This general form of input was popular with SAE and Phase Linear just to name a few. Many years of experience with these has taught me that there is no other way to make these sound great other than doing it the right way. I must have bought thousands of those transistors (2N5551 and 2N5401). I still have over 100 of each in order to find matched quads. Get the like transistor types as close as you can, then find the closest matched other pair. Expect to achieve about 5 mV DC offset, <10 mV really if the diff pairs are properly matched. This also improves the sound quality of the circuit.
Hi Alan0354,
Unequal base current is another way of saying non-matched Beta in these circuits. Attempting to unbalance a pair might help with DC offset, but will hurt in the distortion department.
-Chris
You need to put in the reference number of the components.
This is exactly my finding about the offset problem and Mr. Bob Cordell agree with it. read this post http://www.diyaudio.com/forums/solid-state/171159-bob-cordells-power-amplifier-book-520.html
Start at post # 5192. That's likely your problem. It's the two tail current source. You use the red LED that has less than 2V. Any mismatch of the two LED, the two CCS transistor can create big difference the current. I change my design using 10V zener diodes instead of LED. I even hand match the zeners. Your offset adjustment pot do nothing to fix this.
Last edited:
Alan,
The diff pairs must be matched. There is more latitude in tail currents than there is in beta matching for the diff pairs.
I have and have read Bob's excellent book, as well as Doug's books. Actual experience over 35 years with audio amplifiers confirms what the issues most likely are. BTW, there is zero wrong with using red LEDs for a reference voltage. They are stable, low in noise (less than your zeners) and partially correct for tempco depending on current densities. What the higher voltages using zener diodes will buy you is higher impedance for your current sources. This really isn't an issue here. The different Vbe for the two current source transistors is more of an issue that will require the emitter resistor value to change between + and - sources. A current mirror might be a better solution for one of the current sources that uses the other as a reference.
-Chris
The diff pairs must be matched. There is more latitude in tail currents than there is in beta matching for the diff pairs.
I have and have read Bob's excellent book, as well as Doug's books. Actual experience over 35 years with audio amplifiers confirms what the issues most likely are. BTW, there is zero wrong with using red LEDs for a reference voltage. They are stable, low in noise (less than your zeners) and partially correct for tempco depending on current densities. What the higher voltages using zener diodes will buy you is higher impedance for your current sources. This really isn't an issue here. The different Vbe for the two current source transistors is more of an issue that will require the emitter resistor value to change between + and - sources. A current mirror might be a better solution for one of the current sources that uses the other as a reference.
You are being silly.
-Chris
Of cause it is good to match, but it's not the issue. Did you read the thread? this is not the issue of match pairs. You try my simulation? Changing the current is the major contribution of the offset. Even if the transistor pair is mismatched by 20mV, the offset is going to be only about 20mV. This is 200mV.
Last edited:
good evening,
i did a comparison of ltp transistors i used in both channels and i discovered that i used two different brands despite having the same numbers. so i changed the brand to those on the low dc offset amp and when i tested it at +/- 47vdc and i got -58.7mv. i have not tested the amp at the max voltage coz i want to find out if -58.7mv is okay.
i did a comparison of ltp transistors i used in both channels and i discovered that i used two different brands despite having the same numbers. so i changed the brand to those on the low dc offset amp and when i tested it at +/- 47vdc and i got -58.7mv. i have not tested the amp at the max voltage coz i want to find out if -58.7mv is okay.
i changed the all tansistors on the one channel with high dc offset and when i tested it at +/-47dc, i got -57.5mv. is this okay?
Alan, Shelah has proved my point I think. Beta match is critical. Tail current match is less of a problem.
Hi Shelah,
Good first steps. Yes, those offset numbers are fine to use your amplifier.
Look into ways of matching those transistors more closely. This will improve the sound quality and drop your DC offset numbers. I'm generally not happy until DC Offset is less than 20 mV. By close matching you can attain 5 ~ 10 mV DC offset and much improved sound quality. This last bit is optional for you, but now you can see how important matching the input diff pairs is. Remember, you now have similar transistors installed. They are not matched, so imagine what matching can do.
Thank you for trying my suggestion.
-Chris
Hi Shelah,
Good first steps. Yes, those offset numbers are fine to use your amplifier.
Look into ways of matching those transistors more closely. This will improve the sound quality and drop your DC offset numbers. I'm generally not happy until DC Offset is less than 20 mV. By close matching you can attain 5 ~ 10 mV DC offset and much improved sound quality. This last bit is optional for you, but now you can see how important matching the input diff pairs is. Remember, you now have similar transistors installed. They are not matched, so imagine what matching can do.
Thank you for trying my suggestion.
-Chris
good morning, i tested the amplifier at the max voltage i am using and i still got 200+ mv. what could be the problem this time around?
Good evening, i managed to sort out the high dc offset. the two 470p capacitors in the input section were not okay so i changed and put another brand of ceramic and i got the result i wanted. thanks for the contributions, i have learned this well.
Hi Shelah,
Match those transistors! 🙂
Aaah, you had some oscillation then. Shorted or leaky caps would have reduced the DC Offset. I typically use an oscilloscope for rough readings. This would have revealed your problem. Many meters are average (but peak detecting types) so they can rectify any HF AC. You end up with fictitious readings.
So, what is your DC offset now?
-Chris
Match those transistors! 🙂
Aaah, you had some oscillation then. Shorted or leaky caps would have reduced the DC Offset. I typically use an oscilloscope for rough readings. This would have revealed your problem. Many meters are average (but peak detecting types) so they can rectify any HF AC. You end up with fictitious readings.
So, what is your DC offset now?
-Chris
- Status
- Not open for further replies.