Amplifier Distortion, DC-Offset, and You!
- Thread starter EchoWars
- Start date
Donald&Lydia
Member
Pardon my spelling , Combination of arcing and aching i guess , LOL. Off the topic of DC offset , the Sansui G-9000 switch is an easy fix.
Ylli
Super Member
No inputs, volume set at minimum.
70's Rockfan
Rough N Ready
Thanks EchoWars for posting this thread.
I am new to electronic testing but would like to learn
more about testing and even minor adjusting or repair.
I checked these 3 today following your instructions in post 1
all 3 are completely stock. I only have $30 in the 580 so
that will be a good place for me to start learning on.
Technics SA-5460 5.8mv R and 2.8mv L
Pioneer SX-750 31.5mv R and 32.1mv L
Pioneer SX-580 86.9mv R and 106.4mv L
I am new to electronic testing but would like to learn
more about testing and even minor adjusting or repair.
I checked these 3 today following your instructions in post 1
all 3 are completely stock. I only have $30 in the 580 so
that will be a good place for me to start learning on.
Technics SA-5460 5.8mv R and 2.8mv L
Pioneer SX-750 31.5mv R and 32.1mv L
Pioneer SX-580 86.9mv R and 106.4mv L
frunobulax30
New Member
Kenwood basic M2A, completely original 3.2mv, 6.3mv
Yes
Im an amateur, but both are important, but seems to be more focus on VBE
Found this thread
https://audiokarma.org/forums/index...nsistor-subs-important-for-comp-pairs.757654/
Modified MCS 3248 with KSA992s replacing the SA768 diff pairs. Matched the new transistors using a Peak Atlas DCA75 and a simple Vbe matching circuit I found online (details here: https://audiokarma.org/forums/index.php?threads/diy-transistor-matching-circuit.844771/)
Before mods: 34mV/44mV
After mods: 2mV/4mV
Mods were done to balance the voltages at the bases of the pair and to improve the current matching through each side of the pair.
Before mods: 34mV/44mV
After mods: 2mV/4mV
Mods were done to balance the voltages at the bases of the pair and to improve the current matching through each side of the pair.
I'm an IC designer- matching is what we do. The answer is- it depends. Generally, VBE at a given IC is what you care about, but there are times when Hfe is also what you care about- for example if the source impedance is high. An example of this complexity is in the differential input stage of a bipolar amp where, for example, a pair of "matched" devices are used. Offset will, in large part, consist of the difference in VBEs of the two devices at the same operating collector current, together with a component which is the difference in the external voltage drop due to the base current and the resistance "seen" at each of the input terminals.
If you go to the Leach preamp thread;
https://www.audiokarma.org/forums/index.php?threads/the-leach-moving-coil-pre-preamp.838004/
there's some discussion of these concepts.
In IC design we generally care about VBE (or VGS for that matter) matching at a given output current (collector, drain) and the current gain is whatever it is.
This is most often the case in Audio design.
To put this in perspective. Bipolar devices are NOT (I repeat NOT) best viewed as current in current out, but as voltage in current out with the base current just a consequence of the processes that make that happen. Yes, I know that some people claim otherwise, but they know not of what they speak.
Wyn, if a given pair of transistors matches Vbe at a given input voltage/current, what is the likelihood that they'll match at a different operating point? Do they tend to behave similarly across the board if they match at a given point or is is critical to match them at the operating point they'll be running at in-circuit?
I'm thinking of this from the standpoint of using the matching circuits I linked to a couple of posts up:
https://audiokarma.org/forums/index.php?threads/diy-transistor-matching-circuit.844771/
You could tailor the diode drop (multiple diodes, etc) or the resistor values used to get closer to the operating point in circuit, but I'm not sure how necessary that is to do the job. Obviously in the case of the amp you helped me with, things worked out pretty using it as is.
On another note. I do have a curve tracer in the form of the Peak Atlas DCA-75. For audio applications and specifically for diff pair matching, is there a specific set of curves I should be paying attention to or if a pair of devices matches on one curve is it likely they'll match for the others?
Cheers,
Nathan
I'm thinking of this from the standpoint of using the matching circuits I linked to a couple of posts up:
https://audiokarma.org/forums/index.php?threads/diy-transistor-matching-circuit.844771/
You could tailor the diode drop (multiple diodes, etc) or the resistor values used to get closer to the operating point in circuit, but I'm not sure how necessary that is to do the job. Obviously in the case of the amp you helped me with, things worked out pretty using it as is.
On another note. I do have a curve tracer in the form of the Peak Atlas DCA-75. For audio applications and specifically for diff pair matching, is there a specific set of curves I should be paying attention to or if a pair of devices matches on one curve is it likely they'll match for the others?
Cheers,
Nathan
The basic equation that relates Ic to VBE is an exponential with the same scaled dependencies, so if VBE matches at a given collector current it will match (to a first order) at all other collector currents. There are deviations from this under certain circumstances- there are several higher order effects that kick in, but I think that for the same kind of device it will be good enough.Wyn, if a given pair of transistors matches Vbe at a given input voltage/current, what is the likelihood that they'll match at a different operating point? Do they tend to behave similarly across the board if they match at a given point or is is critical to match them at the operating point they'll be running at in-circuit?
I'm thinking of this from the standpoint of using the matching circuits I linked to a couple of posts up:
https://audiokarma.org/forums/index.php?threads/diy-transistor-matching-circuit.844771/
You could tailor the diode drop (multiple diodes, etc) or the resistor values used to get closer to the operating point in circuit, but I'm not sure how necessary that is to do the job. Obviously in the case of the amp you helped me with, things worked out pretty using it as is.
On another note. I do have a curve tracer in the form of the Peak Atlas DCA-75. For audio applications and specifically for diff pair matching, is there a specific set of curves I should be paying attention to or if a pair of devices matches on one curve is it likely they'll match for the others?
Cheers,
Nathan
You could modify the test circuit to add a common mode resistor to ground at the junction of the two emitter currents. A simple calculation/basic NPN simulation would allow you to determine the resistors needed to decently approximate the desired operating point. The delta V between the emitter voltages would still give you the desired matching result.
Ylli
Super Member
This is a capacitively coupled output. The output caps will block any DC, and measuring the voltage at the outputs without a speaker (or other load) connected is meaningless. Go ahead and hook up those speakers then measure the output offsets.
