audio amp study

[Herschel]

what should i adjust first, bias or 200k for center point?

[boB]

Not sure about the 200K but if you are looking to see more about the output  bias for crossover provided by those 2 diodes in series, look at the DC voltage across the 0.5 Ohm output transistor emitter resistor(s) which will show the quiescent current.   Too much bias and it will get too hot.  Too little and you get more crossover distortion of course.

Fun stuff !

 

[MrAl]

Center voltage is the primary effect and bias is secondary.

For reasonable operation the output has to be near 15V (on a 30V supply). This means 3mA through R7+R8, and you could tweak it from 2mA to 4mA by shifting the output between 10V and 20V, but why? If you want to play with bias, change the value of R13 or R8.

This output stage is not particularly great anyway. Common wisdom is, if you really must use this quasi-complementary configuration, at least add a Baxandall diode and make the emitter resistors more symmetric.

Hello there,

Interesting reply.  Can you prove that the center DC output bias point is more important to be able to adjust than the crossover distortion?

I ask that because I do not recall ever seeing one post before this where the designer/builder was concerned with an adjustment for the DC output DC bias point, while I can recall several posts that talked about getting the crossover distortion minimized while keeping the quiescent output transistor currents low, and that's what most of these circuits have or they have no adjustments at all.  They almost always have an adjustment for the crossover distortion.  I agree though that they are usually better than this circuit (ha ha).

I do recall one instance where the student had to design the output to have maximum output swing, but that was to be by solid design not with an adjustment.  Usually, the design is made to get approximately to the output center point while the crossover distortion requires a finer adjustment.

I don't want to argue these points anymore, but if you could PROVE that the output DC bias point adjustment is more important than the crossover distortion adjustment, especially when there is only ONE adjustment in the circuit, then I would HAVE to agree fully with your opinion.

BTW, in this circuit the output 0.5 Ohm resistors are the same value so not sure why you would object to those values.  If they are both 0.5 Ohms, how could they be more 'symmetrical'.  Just a quick question for you because I was wondering what you meant there.

[MrAl]

what should i adjust first, bias or 200k for center point?

Hi,

Because we have no definite analysis yet, you can measure the output voltage and see if it is off center by a very large amount.  If it is, you would want to keep an eye on that as you adjust, but also look at the quiescent transistor currents to see that they are not too high.  You can then decide if you want better center adjustment or lower power dissipation in the output transistors.  There is a chance they will both come in good enough.

If you build more than one circuit you could see variations on this too as the output transistors and other transistors will not be exactly the same as the very first set.

[Herschel]

ok guys i tried to adjust the center voltage 200k and seems to be, the output power transister are over heating and drawing a nice amount of current. it only happens when i try to adjust the voltage to center i.e to half the supply voltage. else it dont over heat..
is it because of i'm using some random trasistors with high hfe values like 150+ ?
the transistors they are using only have 70 to 80 hfe and they are too ancient they are not available in the local markets
are there any new substitutes for these?

[magic]

β of the transistors shouldn't matter, but maybe the output stage could oscillate if you replace 2N3055 with something faster. The small transistors don't seem critical, the original parts look like any other small transistor and could likely be replaced with any TO92 jellybean as you appear to have done.

actually i have replaced the two 1n4148s and 22R resistor with a single 1n4007 and a 2.2k pot " The two diodes don't generate the necessary 2 V but about 1.4 V" yeah, i had already noticed that so i did that modification. i had to wait for some time to get the amplifier a bit warm to do that

And how did you set this 2k2 pot?
It needs to be re-tuned every time you change VR1, because tweaking VR1 changes R7 and R8 current and hence it changes voltage across your 2k2.

This mod may not be reliable in practice. It may be necessary to replace it with a proper Vbe multiplier mounted on the heatsink (BD139 is often used for this purpose).
edit
But then Q3 and Q4 would also need to be on the heatsink. Or make it two diodes on the PCB and one diode on the heatsink and hope that it will track well enough. I think there is reason why this is an "easy" circuit, and the reason is because the output stage was meant to run underbiased (only two diodes), eliminating the problem of bias adjustment completely...

(... which also demonstrates that VR1 was solely for setting the half-rail voltage and nothing else )

[MathWizard]

I'm amazed at seeing through hole parts put through copper clad boards. It never occurred to me, since I don't have etching materials/tools, or good tools to cut islands on these boards.

I've done some deadbug style stuff with them, but yeah. 

[cbutlera]

Another method to find a complete schematic is looking into the service manuals of vintage audio equipment, e.g. Denon, Sony and Technics.
...

The schematic looks similar to the one below, which is the power amplifier section of my ancient Sony TA-1120A (purchased by my late father in 1972 and still in every day use by me, albeit with a few repairs from time to time).  There seem to be some critical parts missing though, such as the primary AC feedback path, which in the TA-1120A sets the gain to about 20.  Note that the SV-06 components are not normal diodes, but silicon varistors.

The full service manual for the TA-1120A is available on Elektrotanya https://elektrotanya.com/sony_ta-1120a_amplifier.pdf/download.html

Edit: I missed that Q4 in the OPs schematic was a PNP, so I think the design has a reasonable gain and my comment was wrong.

[dietert1]

Once the bias current is adjusted to 40 or 50 mA, the output transistors get about 30 V * 45 mA = 1,35 W. So that heatsink will get warm but not really hot.
Yes, an Ube multiplier like shown in the ESP amplifier schematic linked above works much better. I think the transistor doesn't need to be on the heatsink. The heat generated by the nearby driver transistors should be enough. By the way that Ube multiplier can be made with fixed resistors. E.g. resistors of 5K6 and 12K will give you the required multiplication by three, about 2 V. I remember that this works pretty well when using 1 % metal film resistors, as Ube is a physical constant. No real need for adjustment.

Regards, Dieter

[MrAl]

β of the transistors shouldn't matter, but maybe the output stage could oscillate if you replace 2N3055 with something faster. The small transistors don't seem critical, the original parts look like any other small transistor and could likely be replaced with any TO92 jellybean as you appear to have done.

actually i have replaced the two 1n4148s and 22R resistor with a single 1n4007 and a 2.2k pot " The two diodes don't generate the necessary 2 V but about 1.4 V" yeah, i had already noticed that so i did that modification. i had to wait for some time to get the amplifier a bit warm to do that

And how did you set this 2k2 pot?
It needs to be re-tuned every time you change VR1, because tweaking VR1 changes R7 and R8 current and hence it changes voltage across your 2k2.

This mod may not be reliable in practice. It may be necessary to replace it with a proper Vbe multiplier mounted on the heatsink (BD139 is often used for this purpose).
edit
But then Q3 and Q4 would also need to be on the heatsink. Or make it two diodes on the PCB and one diode on the heatsink and hope that it will track well enough. I think there is reason why this is an "easy" circuit, and the reason is because the output stage was meant to run underbiased (only two diodes), eliminating the problem of bias adjustment completely...

(... which also demonstrates that VR1 was solely for setting the half-rail voltage and nothing else )

Hi,

Although we can't still be sure about that last statement, I do not see how you came to that conclusion through a test of a circuit that is not the right circuit now.  We (you and I both) could change the design such that the pot changes the output zero DC center, but then that would be a different circuit (or even the same circuit).
As was pointed out before, when we adjust the output DC center point, the quiescent current though both of the output transistors changes also.  This makes it hard to nail down the use of the 200k pot theoretically, or at least simply.  That makes it necessary to adjust while keeping an eye on both of these things.
The bottom line is, if we adjust the output DC voltage to one-half of Vcc, then if the quiescent current is too high we would have to back off with that and settle for a less-than-perfect output DC voltage center point.

I don't know if you do any bench testing and have the parts to do this, but if you did you could try it and see what happens

[Xena E]

I'm not going to comment on the crap design as a whole, though I will compliment the OP on their neat assembly work.

However:

Unfortunately if there is no means of fixing the quiescent bias votage, then trying to alter the mid point will change that also. In fact, with just a simple resistive setting of the output stage current, It will also change asymmetrically in sympathy with the signal, causing amplitude nonlinearity.

The op needs to replace the simple pot/diodes whatever, with a transistor Vbe multiplier circuit, as previously suggested, to minimise the change of bias voltage, (output quiescent current bias), as happens when the current in the driver stage changes.

Without either underrunning the bias or having a correctly designed circuit that has thermal feedback this amplifier could eventually die from thermal runaway.

To answer the OP question regarding which is the more important, then that would be the quiescent current bias, the centerpoint voltage being incorrect will only limit the output swing in one or other direction from the quiescent output voltage point.

Just search 'amplified diode circuit', implement it correctly, and life will become much simpler.

Regards,
Xena

[Herschel]

β of the transistors shouldn't matter, but maybe the output stage could oscillate if you replace 2N3055 with something faster. The small transistors don't seem critical, the original parts look like any other small transistor and could likely be replaced with any TO92 jellybean as you appear to have done.

actually i have replaced the two 1n4148s and 22R resistor with a single 1n4007 and a 2.2k pot " The two diodes don't generate the necessary 2 V but about 1.4 V" yeah, i had already noticed that so i did that modification. i had to wait for some time to get the amplifier a bit warm to do that

And how did you set this 2k2 pot?
It needs to be re-tuned every time you change VR1, because tweaking VR1 changes R7 and R8 current and hence it changes voltage across your 2k2.

This mod may not be reliable in practice. It may be necessary to replace it with a proper Vbe multiplier mounted on the heatsink (BD139 is often used for this purpose).
edit
But then Q3 and Q4 would also need to be on the heatsink. Or make it two diodes on the PCB and one diode on the heatsink and hope that it will track well enough. I think there is reason why this is an "easy" circuit, and the reason is because the output stage was meant to run underbiased (only two diodes), eliminating the problem of bias adjustment completely...

(... which also demonstrates that VR1 was solely for setting the half-rail voltage and nothing else )

yep, that overheating issue makes sense.. because i was using a ksh13009h transistor i found inside a powersupply instead of 2n3055 so i could test the circuit...

alse here is how i adjusted the bias:
first before turning the amp on, both the presets i.e 2.2k and 200k are adjusted to its least value
then i adjusted the 200k pot to the center voltage
after that, i adjusted the 2.2k pot to get 2volts across the 1n4007 diode and 2.2k pot.
  the 200k votage dont vary every time when i adjust the bias (2.2k)

[Herschel]

I'm amazed at seeing through hole parts put through copper clad boards. It never occurred to me, since I don't have etching materials/tools, or good tools to cut islands on these boards.

I've done some deadbug style stuff with them, but yeah. 

i dont have any fancy tools, the only thing i use is a small plastic box and an over concentrated fecl3 sol. which i made by mixing too much fecl3 powder with a little amount of water. so i could get the job done within a minute. Even the first try was pretty easy for me...
for me, zero pcbs are a time killer, as I'm too much busy going to school just to learn s###
so i have to finish every thing within an hour a week...
in india, most schools are against students creativity

[Herschel]

Another method to find a complete schematic is looking into the service manuals of vintage audio equipment, e.g. Denon, Sony and Technics.
...

The schematic looks similar to the one below, which is the power amplifier section of my ancient Sony TA-1120A (purchased by my late father in 1972 and still in every day use by me, albeit with a few repairs from time to time).  There seem to be some critical parts missing though, such as the primary AC feedback path, which in the TA-1120A sets the gain to about 20.  Note that the SV-06 components are not normal diodes, but silicon varistors.

The full service manual for the TA-1120A is available on Elektrotanya https://elektrotanya.com/sony_ta-1120a_amplifier.pdf/download.html

Edit: I missed that Q4 in the OPs schematic was a PNP, so I think the design has a reasonable gain and my comment was wrong.

i have seen this circuit before, actually this circuit led me to make the circuit that now i have made so why i havent used this circuit? instead of stepping up voltage this high,94 volts, its better to use a split supply one which i already have coppied from another sony poweramp because it just sounds good for some reason...

[magic]

alse here is how i adjusted the bias:
first before turning the amp on, both the presets i.e 2.2k and 200k are adjusted to its least value
then i adjusted the 200k pot to the center voltage
after that, i adjusted the 2.2k pot to get 2volts across the 1n4007 diode and 2.2k pot.

You can measure actual output stage bias by measuring voltage drop across R19 (or R17 - they should be almost equal).
Try to adjust 2k2 again, watching R19 voltage on a DMM. About 50mA or even 100mA bias (25mV to 50mV at R19) is only a few watts of power at this supply voltage and should be good enough. Beware that bias may drift as the amplifier warms up.

As was pointed out before, when we adjust the output DC center point, the quiescent current though both of the output transistors changes also.  This makes it hard to nail down the use of the 200k pot theoretically, or at least simply.  That makes it necessary to adjust while keeping an eye on both of these things.

Here's my point: at least in the original design shown in the first post, the output stage is inevitably underbiased because there are only two diodes and 22Ω between the bases of Q3 and Q4, but three PN junctions in series in transistors Q3, Q5, Q4. As a result, only the drivers Q3 and Q4 are active with maybe a few mA through them. It makes little difference whether Q2 current is 2mA or 4mA, there is no way to achieve proper bias in Q5 and Q6, no critical adjustment to make for minimum distortion and surely no risk of overbiasing and blowing things up.

Note that the OP modified the circuit, replacing one of the diodes with a 2k2 pot.

[Herschel]

i checked the voltage across R17 and R19, and there was 250mV going through each of them. is it supposed to 14.3mV going through each of them?. this all happens when i try to adjust 200k pot to the center.
now its adjusted to 9v insted of 15v... still the bias adjust dont drift when i adjust 200k pot

[MrAl]

alse here is how i adjusted the bias:
first before turning the amp on, both the presets i.e 2.2k and 200k are adjusted to its least value
then i adjusted the 200k pot to the center voltage
after that, i adjusted the 2.2k pot to get 2volts across the 1n4007 diode and 2.2k pot.

You can measure actual output stage bias by measuring voltage drop across R19 (or R17 - they should be almost equal).
Try to adjust 2k2 again, watching R19 voltage on a DMM. About 50mA or even 100mA bias (25mV to 50mV at R19) is only a few watts of power at this supply voltage and should be good enough. Beware that bias may drift as the amplifier warms up.

As was pointed out before, when we adjust the output DC center point, the quiescent current though both of the output transistors changes also.  This makes it hard to nail down the use of the 200k pot theoretically, or at least simply.  That makes it necessary to adjust while keeping an eye on both of these things.

Here's my point: at least in the original design shown in the first post, the output stage is inevitably underbiased because there are only two diodes and 22Ω between the bases of Q3 and Q4, but three PN junctions in series in transistors Q3, Q5, Q4. As a result, only the drivers Q3 and Q4 are active with maybe a few mA through them. It makes little difference whether Q2 current is 2mA or 4mA, there is no way to achieve proper bias in Q5 and Q6, no critical adjustment to make for minimum distortion and surely no risk of overbiasing and blowing things up.

Note that the OP modified the circuit, replacing one of the diodes with a 2k2 pot.

Hello again,

That sounds reasonable, and I could check that idea (about the bias not changing) but not sure if I can get to it today.  I can see your point but you did not really mention WHY you thought that.  If we just say, "It does this..." that sounds like JUST an opinion, and that is fine I guess, but better is to say WHY we say something, like "Well the current increases in A and therefore the current increases in B and then finally in C the voltage goes down."  That would explain it a little better and then we would all know the basis of that opinion.  If we just say, "The voltage in C goes down", that would not be as good as mentioning the other two points with it.
A leads to B leads to C, and that tells us the whole story, hopefully

I'm not sure how interested I am in discussing this circuit anymore though because the circuit changes too much and replies do not always reflect which version is being discussed.  If I do a more in-depth study I'll post the actual circuit I use with it.

[MrAl]

I'm not going to comment on the crap design as a whole, though I will compliment the OP on their neat assembly work.

However:

Unfortunately if there is no means of fixing the quiescent bias votage, then trying to alter the mid point will change that also. In fact, with just a simple resistive setting of the output stage current, It will also change asymmetrically in sympathy with the signal, causing amplitude nonlinearity.

The op needs to replace the simple pot/diodes whatever, with a transistor Vbe multiplier circuit, as previously suggested, to minimise the change of bias voltage, (output quiescent current bias), as happens when the current in the driver stage changes.

Without either underrunning the bias or having a correctly designed circuit that has thermal feedback this amplifier could eventually die from thermal runaway.

To answer the OP question regarding which is the more important, then that would be the quiescent current bias, the centerpoint voltage being incorrect will only limit the output swing in one or other direction from the quiescent output voltage point.

Just search 'amplified diode circuit', implement it correctly, and life will become much simpler.

Regards,
Xena

Yes apparently there is a little bit of a communication problem.  The better bias circuit would of course be, better, but apparently that didn't get through yet.

[dietert1]

The difficulty to adjust bias probably indicates oscillation. The schematics presented above are missing a local buffer cap for the supply. When using test leads for the supply, this missing cap can already cause instability and oscillation.

[Herschel]

i have already adjusted the bias... for me, the problem was adjusting the 200k to center the voltage.

later the reason was found to be the stepup boost converter causing the problem. when i tweak the 200k pot to adjust the center voltage, for some reason, the voltage increases in the stepup converter but why?, anyway i replaced the boost converter, and it was working just fine

[Herschel]

β of the transistors shouldn't matter, but maybe the output stage could oscillate if you replace 2N3055 with something faster. The small transistors don't seem critical, the original parts look like any other small transistor and could likely be replaced with any TO92 jellybean as you appear to have done.

actually i have replaced the two 1n4148s and 22R resistor with a single 1n4007 and a 2.2k pot " The two diodes don't generate the necessary 2 V but about 1.4 V" yeah, i had already noticed that so i did that modification. i had to wait for some time to get the amplifier a bit warm to do that

And how did you set this 2k2 pot?
It needs to be re-tuned every time you change VR1, because tweaking VR1 changes R7 and R8 current and hence it changes voltage across your 2k2.


This mod may not be reliable in practice. It may be necessary to replace it with a proper Vbe multiplier mounted on the heatsink (BD139 is often used for this purpose).
edit
But then Q3 and Q4 would also need to be on the heatsink. Or make it two diodes on the PCB and one diode on the heatsink and hope that it will track well enough. I think there is reason why this is an "easy" circuit, and the reason is because the output stage was meant to run underbiased (only two diodes), eliminating the problem of bias adjustment completely...

(... which also demonstrates that VR1 was solely for setting the half-rail voltage and nothing else )

that transistors did actually contributed the overheating issue. I changed the transistors from 2n5551 to 2n2222 which had a really low hfe compared to 2n5551 which had 150 hfe min. which actually reduced the heating issue for some reason

the overheating was actually caused by the failing boost stepup converter which i used. later changed to one which can handle 6 amps.

[Herschel]

also, now it plays music really well!

[kevin.gibbs]

With a 200kOhm resistor, you will set half of the power supply at the amplifier output. Selecting a 20-ohm resistor (in series with the diodes) will set the amplifier's initial current to about 50mA (when there is no signal)

[Herschel]

With a 200kOhm resistor, you will set half of the power supply at the amplifier output. Selecting a 20-ohm resistor (in series with the diodes) will set the amplifier's initial current to about 50mA (when there is no signal)

that 20 resistor + 2* 1n4148 zeners are already replaced with a in4007 diode and a 2.2k pot, as there was only 1.3volt for biasing but i need atleast 1.8 volts. now everything works pretty well...