Author Topic: TL071 distortion - bad amplifier design?  (Read 4129 times)

0 Members and 1 Guest are viewing this topic.

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 11731
  • Country: gb
  • Hero999
Re: TL071 distortion - bad amplifier design?
« Reply #25 on: December 10, 2018, 11:21:25 pm »
And for god's sake put limiting resistors on base on final resistors, once they saturate you're in trouble.
On input I will use in your place bigger capacitors like 10 uF


Envoyé de mon iPad en utilisant Tapatalk
You mean the final transistors, not resistors, i.e. the 2SC52EE and 2SA1943?

If so, how can they saturate? They're emitter followers.
They can.for example, from thermal considerations or if driver goes short.that's why between emiter driver and final transistor base it's good idea to put resistor.see the second schematic that does have.
Pierre


Envoyé de mon iPad en utilisant Tapatalk
If the driver goes short circuit, the whole thing is toast anyway, whether there's a small base resistor before the output stage or not. Suppose the TIP41 goes short. How will a 2R2 resistor in series with the 2SC52EE's base possibly offer any protection? A huge current will flow through the 2SC52EE. Hopefully the amplifier will blow up before the speaker is destroyed.

The output stage is an emitter follower so can't go into saturation. Even if the base of one of the output transistors is connected directly to +V, it will still not saturate.  The emitter voltage will rise, thus reducing the base-emitter voltage, until the emitter will sit a couple of volts below the collector.
https://en.wikipedia.org/wiki/Common_collector#Basic_circuit

There may be other reasons for including a base resistor, such as reducing the gain and making it less susceptible to oscillation, but preventing saturation is not one of them.
 

Offline perieanuo

  • Frequent Contributor
  • **
  • Posts: 271
  • Country: fr
Re: TL071 distortion - bad amplifier design?
« Reply #26 on: December 10, 2018, 11:36:43 pm »
of course it can saturate
https://www.allaboutcircuits.com/textbook/semiconductors/chpt-4/common-collector-amplifier/
https://www.electronics-tutorials.ws/amplifier/amp_2.html
get yourself a real amplifier, put it on your basement and put it to over 80% for a while with nominal load (cut the cooler fan to do it quickly), buy a multimeter or scope and tell us the result
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 11731
  • Country: gb
  • Hero999
Re: TL071 distortion - bad amplifier design?
« Reply #27 on: December 11, 2018, 12:50:58 am »
of course it can saturate
https://www.allaboutcircuits.com/textbook/semiconductors/chpt-4/common-collector-amplifier/
https://www.electronics-tutorials.ws/amplifier/amp_2.html
A common emitter amplifier yes, it can saturate, but a common collector amplifier can't saturate, unless the input voltage is taken beyond the supply voltage. The tutorial you've linked to actually explains this.

Quote
Given the voltage polarities across the base-emitter PN junction and the load resistor, we see that these must add together to equal the input voltage, in accordance with Kirchhoff’s Voltage Law. In other words, the load voltage will always be about 0.7 volts less than the input voltage for all conditions where the transistor is conducting. Cutoff occurs at input voltages below 0.7 volts, and saturation at input voltages in excess of battery (supply) voltage plus 0.7 volts.

In the circuit posted at the start of the thread, the input voltage can't exceed the supply voltage, so the output state can't saturate.

Quote
get yourself a real amplifier, put it on your basement and put it to over 80% for a while with nominal load (cut the cooler fan to do it quickly), buy a multimeter or scope and tell us the result
I seems like you're confusing clipping distortion and overheating with saturation of a BJT. If an emitter follower output stage is over-driven, it will clip, distort and possibly overheat, if the heat-sinking and power rating of the transistors is insufficient, but neither of the transistors will ever enter the saturation region. They will both be operating in the active region the entire time.

In any case, adding base resistors to  2SC52EE and 2SA1943 will do little to protect them against overheating. It may improve the stability, at the expense of a reduced voltage swing.


EDIT:
Look here and try to understand in practice it can arrive this situation"The bias resistor must connect to a voltage that is at least 0.7V higher than the supply voltage for the upper emitter-follower transistor to saturate." (eg if you have voltage loss at high power on your wires inside the amplificator, yes, the wires between the final transistor and the +Vdd, I seen this in a clone philips schematic, in practice) :
Sorry, I missed this.

The wires between the final transistor and the supply voltage would have to be a very high resistance, or carrying a large current to drop nearly 1V (more more than 0.7V is required to saturate a transistor). An additional circuit would be required to protect against the high current scenario, as without it the amplifier would be toast anyway.

I see what you mean now and agree: a base resistor would protect against a high resistance or open circuit collector connection, because the maximum base current rating would be exceeded. Are you sure the extra resistor in the Philips amplifier wasn't to improve the stability?

There will be plenty of other points in the amplifier which can be destroyed by a bad connection. For example, what will happen if the connection to one of the zener diodes is bad? The TL071 will get 40V on one of its supply rails.

It's not feasible to protect against everything that can possibly go wrong!
« Last Edit: December 11, 2018, 07:15:05 am by Hero999 »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 5099
  • Country: de
Re: TL071 distortion - bad amplifier design?
« Reply #28 on: December 11, 2018, 03:19:24 am »
The circuit is problematic: There is quite some gain inside the loop of the TL071. This requires quite some extra capacitors for the compensation to slow it down.  With those rather large capacitors the TL071 sees quite some capacitive load and it might thus still oscillate at a high frequency.

The output stage is class B and thus would need a rather fast feedback to keep cross over distortion low. However as shown the first transistor stage is rather slow.

So the circuit is more like a poor example and can not expected to give low distortion output.

Another point is that the biasing current for the C1815 and A1015 depends on the supply voltage. So voltage drops (e.g. due to small filter caps) could make it leave the reasonable working range. Too high a voltage could cause excessive current and blow the small transistors.
For understanding the circuit it might help to simulate it - however there is likely little success in improving this type of circuit.
 

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #29 on: December 11, 2018, 03:23:17 am »
Could you please explain why the first transistor stage is slow?
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 11731
  • Country: gb
  • Hero999
Re: TL071 distortion - bad amplifier design?
« Reply #30 on: December 11, 2018, 03:33:55 am »
The circuit is problematic: There is quite some gain inside the loop of the TL071. This requires quite some extra capacitors for the compensation to slow it down.  With those rather large capacitors the TL071 sees quite some capacitive load and it might thus still oscillate at a high frequency.

The output stage is class B and thus would need a rather fast feedback to keep cross over distortion low. However as shown the first transistor stage is rather slow.

So the circuit is more like a poor example and can not expected to give low distortion output.

Another point is that the biasing current for the C1815 and A1015 depends on the supply voltage. So voltage drops (e.g. due to small filter caps) could make it leave the reasonable working range. Too high a voltage could cause excessive current and blow the small transistors.
For understanding the circuit it might help to simulate it - however there is likely little success in improving this type of circuit.
Are you sure the final stage is class B? Have you done any calculations?

Whether the final stage is class B or AB, will depend on how much current is flowing through the 100R resistors in the TIP41 and TIP42 stage. I suspect if it is class B, it can be tweaked into class AB fairly easily.

EDIT: I see how it's class B now.
« Last Edit: December 11, 2018, 04:03:57 am by Hero999 »
 

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #31 on: December 11, 2018, 03:39:06 am »
Well, on the scope, the crossover distortion was clear. There were bumps on the side of the waves near the 0V line. I guess lowering the 220R resistor would help bring it to the "AB" side.
 

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #32 on: December 11, 2018, 03:57:09 am »
Are you sure the final stage is class B? Have you done any calculations?

Whether the final stage is class B or AB, will depend on how much current is flowing through the 100R resistors in the TIP41 and TIP42 stage. I suspect if it is class B, it can be tweaked into class AB fairly easily.
[/quote]


Arent the 220R resistor also responsible for the bias of the TIPs? Or they are so low that the 1015 and 1815 have plenty of base current and what is limiting their output is just the 100R on their emitters?
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 11731
  • Country: gb
  • Hero999
Re: TL071 distortion - bad amplifier design?
« Reply #33 on: December 11, 2018, 04:09:13 am »
Looking at it again, Kleinstein was right, it's definitely class B. The diodes on the bases of the TIP4x transistors hold their bases at 1.4V apart, so hardly any current flows through the 100R resistors. Adding a couple of extra diodes should increase the current flow and bias it into class AB operation.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 5099
  • Country: de
Re: TL071 distortion - bad amplifier design?
« Reply #34 on: December 11, 2018, 04:49:13 am »
The TIP41/42 run kind of class AB, but the SD52xx / SA19.. are running class B. One could at least add a 3 rd diode. To make is class AB it would need something like an VBE multiplier instead of the diode, set to a little more than 4 diode drops to adust the standing current. It won't be very stable however, as the VAS current is not stable.

Changing the 220 Ohms resistors is tricky, as there correct value depends on the supply voltage - which is one of the weak points. Too much and the 1015/1815 overheat. To little and the VAS stage is also kind of working in an class B mode with more distortion.
It might get a little better with a diode in series with a smaller resistor - though this might load the OP too much. The rabbit hole just starts here.
 

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #35 on: December 11, 2018, 06:15:23 am »
And for god's sake put limiting resistors on base on final resistors, once they saturate you're in trouble.
On input I will use in your place bigger capacitors like 10 uF


Envoyé de mon iPad en utilisant Tapatalk
The output transistors are fine. But it is quite true, that in common with many audio amplifiers, that this amplifier is not short circuit protected
 
The following users thanked this post: AngraMelo

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #36 on: December 11, 2018, 06:21:34 am »
Well, on the scope, the crossover distortion was clear. There were bumps on the side of the waves near the 0V line. I guess lowering the 220R resistor would help bring it to the "AB" side.
The bumps 'on the sides' are crossover distortion. If you can see them on the scope, it is gross distortion.
The amplifier as it stands is essentially class B, not class AB.

On the schematic that you originally posted there is a missing dot connection between the two 0.22R resistors. Is this a typo and is that connection actually made in practice?

Did you try changing the capacitors as I suggested in reply #21?
« Last Edit: December 11, 2018, 01:36:00 pm by spec »
 
The following users thanked this post: AngraMelo

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #37 on: December 11, 2018, 06:54:03 am »
It is a typo. The resistors are connected to the output.
Im trying your suggestin right now. Ill be back in a moment
 

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #38 on: December 11, 2018, 07:43:42 am »
It is a typo. The resistors are connected to the output.
Im trying your suggestin right now. Ill be back in a moment
Thanks

Attached below is the mod for converting the amplifier from deep class B to class AB. Set Iq to 30mA to 40mA. Physically place the flat part of the BC337 case in contact with one of the driver transistor cases.
« Last Edit: December 11, 2018, 07:46:00 am by spec »
 
The following users thanked this post: AngraMelo

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #39 on: December 11, 2018, 07:47:29 am »
Thank you so much!!!! Im about to try it!!!

I just soldered the 2 47uf capacitors.
Should I keep them on your desing? Or I go back to the 4.7nf?
 

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #40 on: December 11, 2018, 08:10:59 am »
New 47uF capacitors are a success. Much better sound and no more noise at lower volumes.
Now to try the BC337.

Oh man, Im so freaking happy.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 8331
  • Country: us
  • DavidH
Re: TL071 distortion - bad amplifier design?
« Reply #41 on: December 11, 2018, 09:36:56 am »
The circuit is problematic: There is quite some gain inside the loop of the TL071. This requires quite some extra capacitors for the compensation to slow it down.

It is not quite dire.  The extra voltage gain inside the loop is attenuated by the feedback network before reaching the non-inverting input.  Indeed, if there was no extra voltage gain, then the operational amplifier would be overcompensated.

But I agree there are better configurations which use local feedback to control the gain added by the output stage.  Below is the example I was thinking off where local feedback is applied to the *output* of the operational amplifier which becomes an input.  How weird is that?
 

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #42 on: December 11, 2018, 12:06:53 pm »
New 47uF capacitors are a success. Much better sound and no more noise at lower volumes.
Now to try the BC337.

Oh man, Im so freaking happy.
:) Great news- it is nice when theory and practice match.
 
The following users thanked this post: AngraMelo

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #43 on: December 11, 2018, 12:08:12 pm »
I sent you a message inbox!
 

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #44 on: December 11, 2018, 12:27:32 pm »
CROSSED POSTS

UPDATE #1 2018_12_11 (two resistor values changed on amplifier schematic)
UPDATE #2 2018_12_11 (BC556 transistor added to better controlled constant current)

Attached below is the amplifier circuit fully worked over, but some detailed components have been omitted, like decoupling capacitors and padding resistors, for example. You will notice that there are no 47uF capacitors, but an extra transistor has been added. This transistor has no voltage or current gain and mainly functions as a level shifter. It has two other functions which I can describe if you want to know.

You will also notice that the supply rails have been reduced to 23V and -23V.
The reason for this is that, while looking at the transistor datasheets I noticed the following:

2SC5200: 230VCE
2SA1043: 230VCE
TIP41     : 60VCE
TIP42     : 60VCE
A1015    : 50VCE
C1815    : 50VCE

The two output transistors are Toshiba beauties and, obviously have no problem with +-40V supply lines, but I haven't checked the other parameters. I presume you are using 8 Ohm speakers.
The other transistors are only average in terms audio amplifiers and they limit the supply line voltage.

But, as you say that the amp is working OK with +-40V supply lines, that means that your particular transistors probably have a high enough VCE rating, so you can continue testing at risk with +40 and -40 supply lines.
« Last Edit: December 11, 2018, 07:26:32 pm by spec »
 
The following users thanked this post: AngraMelo

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #45 on: December 11, 2018, 12:38:00 pm »
That solid tantulan is 2.2uf or 220uf? Did you see my message?
 

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #46 on: December 11, 2018, 12:42:00 pm »
That solid tantulan is 2.2uf or 220uf? Did you see my message?
220uF if you want a really low solid base. 100uF would be very good and 22uF would knock off some of the very low bass. You can make the choice quite freely and you may find that 2u2F suits your room/speakers/sound source/personal preference

Of course, you can go really high-end (low distortion and frequency response down to DC ) and do away with the capacitor altogether: just replace it with a short circuit.The consequence would be a worst case DC current of 42mA through your speakers, but that would not worry your speakers in the slightest.

The overall frequency response of the amplifier would be a single unit slope defined only by the input polypropylene capacitor and the opamp input bias resistor. This would give a beautiful open, rolling base with the right speakers.

And if you changed the opamp for ...

Will look at your message now.
« Last Edit: December 11, 2018, 07:36:16 pm by spec »
 
The following users thanked this post: AngraMelo

Offline AngraMelo

  • Regular Contributor
  • *
  • Posts: 176
  • Country: br
Re: TL071 distortion - bad amplifier design?
« Reply #47 on: December 11, 2018, 04:16:27 pm »
awesome! I will try different capacitors and check out the sound!
What is the op amp you were saying??? you got me curious!
 

Offline spec

  • Frequent Contributor
  • **
  • Posts: 833
  • Country: england
  • MALE
Re: TL071 distortion - bad amplifier design?
« Reply #48 on: December 11, 2018, 07:21:32 pm »
awesome! I will try different capacitors and check out the sound!
What is the op amp you were saying??? you got me curious!
Will post in about 10 hours.

Note the revised schematic in reply #44

Once this amplifier is working OK what do you want to do with it?

[1] Keep as is which will be a good quality audio amplifier with a sweet smooth sound

[2] More definition, lower distortion, more articulation, more impact (opamp change mainly)

[3] Maximum fidelity for this architecture (opamp change, component changes, and physical layout changes)

By the way, I meant to say before, all resistors, except perhaps the two 0R22 resistors, must be metal film types. A future modification will possibly include changing the 0R22 resistors for 0R1 low inductance resistors.

« Last Edit: December 11, 2018, 07:43:06 pm by spec »
 
The following users thanked this post: AngraMelo

Offline Mark Hennessy

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: gb
    • My electronics and audio website
Re: TL071 distortion - bad amplifier design?
« Reply #49 on: December 12, 2018, 12:07:26 am »
Just as a side-note, it's worth saying that there is a lot of confusion about amplifier class. Putting aside class A and just considering B and AB, Douglas Self shows that actually, class B is the ideal mode of operation.

Class B means that each device handles exactly half (180 degrees) of the signal. When there is crossover distortion, then the devices are carrying less than 180 degrees; hence the amplifier is in class C mode, not class B.

The other case - when the bias is increased and the devices overlap to handle more of the signal - is class AB. In an otherwise blameless amplifier, a class AB amplifier will have slightly more distortion than a perfectly biased class B amplifier because of gm-doubling.

All explained in the book: https://www.amazon.co.uk/Audio-Power-Amplifier-Design-Douglas/dp/0240526139 - PDF versions of older editions are easily found on-line, but I'd always recommend buying a copy of the latest. The books are based on his series of articles published in Wireless World back in 1993.

The difficulty in a practical amplifier is holding that exact bias point for class B operation under all circumstances, which is why most manufacturers play it safe by increasing the bias to put the amp into class AB, which is better than risk leaving it on the edge and risk having it fall into class C under certain circumstances. The book talks a lot about schemes to improve the thermal tracking so that this ideal point can be better maintained - it's really quite fascinating. Obviously, all this is well below what is audible, but it's intellectually satisfying to chase and minimise real distortion mechanism where easy and cheap to do so.

 
The following users thanked this post: Zero999, oPossum, AngraMelo


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf