Tone controls - what's going on?

Ooh, well the thought of an amplifier to add onto a PC sounds like a nice little project - see, I'd rebuild it and add tone controls, now that I have learned how etc - run the whole lot through two (for stereo) of those TDA200x amplifier chips and a little unity gain op amp as a pre amp - lovely stuff. Never used a digital pot, I have no idea how they function lol!

Motorboating is a pretty standard problem with audio amplifiers I think, the US based messageboard I frequent quite often is full of guys saying their old tube radio is motorboating - the usual cure is as you say, replace an out of spec resistor or a leaky capacitor.

I've tried to build quite a few little amps for solid state radio projects based on the LM386, but alas they just seem to be so unstable. The TBA820M is a nicer chip to work with.

Harry's little half Watt audio amp is well behaved however, I've never had a problem with that circuit, it doesn't seem to matter if I build it on protoboard or bother to etch a board - works first time - more than cam be said of the '386 chip.

Glad you sorted out that organ, I've seen the wiring inside one of those things once - made spaghetti look neat!

I have an old plain matrix board (perf. board) 100 watt quad board in my computer desk that I build 10 years ago and it's still running. it has no tone controls and it used to have volume pots but they got kicked so much they went intermittent and they are now bypassed. I have thought about adding some digital pots but haven't gotten around to it yet.
I recently ran into a strange problem. It was an organ with 5 60w amps. Each one had a filter but the bass amp had a low pass filter in front of it. One amp had an open bios resistor but it was not motor-boating and only the bass amp was. I finally saw it on the power supply and then on the input of the amp I thought was good and had been causing the whole problem. The LP filter was evidentially passing it through but the others must have blocked it. Of course there was a mixer in front of all these amps making the problem much harder to find.

That small board is made on strip board, it has copper strips on the back, running the length of the board. I too reuse the stuff for prototyping, but after a while it gets full of track breaks and becomes unusable - so I like to build on plain matrix board, move to stripboard, and then make a PCB when I'm happy with it.

I don't mind doing the dead bug thing, but I like "neat". Dead bug is ok to get some feedback on an idea or whatever, but if I'm going to stuff it inside a case and use it, I want wires neat and tidy, gold bands on resistors all facing toward ground, all markings on caps etc to face either to the right, or to ground - it's all part of my OCD

The Baxandall unit you see in those pictures have off-board pots, and that means wires - 6 to each pot! Plus the phono jacks and power requirements means lots of wiring - my pet hate! So in the new design (uses op-amps instead of transistors) I will try to route the tracks so that the control pots are board mounted. The power supply will also be board mounted as I plan to use a split rail supply. So far everything has used a single rail supply. There will be wires, lots of them - there will be at least 4 inputs (CD, Tape, Tuner and AUX), and I may even build an RIAA phono pre-amplifier so I can put my turntable into it too. Then there's the outputs - one for the power amp AUX input, one for the "through" (which will also double as the power switch for the unit). There's a long way to go as yet, but at least the actual Baxandall circuit works - the rest is just switching inputs and outputs. Oh, and then there's the case to consider I could go on, and maybe I will - once I'm actually closer to getting things finalised in my mind.

I don't really listen to vinyl anymore so the RIAA preamp is not really a prority - but just in case.....

So this all leads on to a few questions really - maybe best start a new thread on op-amps.

Hi Dragonforce
I see you have one board there using perf. board. Is that one with foil on the back or just a bare board. I really like building with that board because I can easily remove parts and and run my buses where ever I need and make my grounds heavier where I need them and so-forth. It's not good for RF but works just fine and neat for everything else. I have boards that have been re-used many times and still work for new circuits. Even the V5 works very well on them but that's pretty much pushing the limit. Anything that has frequencies that can cross couple will not work on it even when you try to expand the ground plain.
I am mostly an experimenter and I like to make too many changes after the board is etched for PCB's to work well for me. I also find dead-bugging too messy but do it sometimes.

These are the small single stage modules I was talking about. All input and output signals, ground and +ve are all routed to the pin connector on the left of the board. Not the Baxandall board has no output cap since there is one on the LINE amp board. A zero ohm resistor is used to switch the ground pin between -VE and the virtual ground required for the op-amp based LINE amp.



The prototype stereo active Baxandall unit - notice the use of plug sockets. The main board is constructed on a PCB identical to the second one you can see in the foreground - it's a spare



A close up of the board - included because I think it looks pretty lol

This, is the progress so far. Other boards have been constructed but robbed of bits etc, so were not included in the "gallery"

At the moment, I'm planning a discreet transistor pre amp and LINE amp based on a pair of BC109C's.

First I built the preamp on stripboard, and laid it out so that it was a module that could be plugged into a test board.

Then I built the Bandaxall control in the same way, followed by the main LINE amp.

On the test board, they're arranged like a letter "E", maybe I'll get some photos.

Well this worked just fine, so I went ahead and designed a PCB layout for the whole thing.

The point of doing the little boards as individual modules was so that I could test various op-amp LINE amps without wasting too many components - it's okay to use a breadboard but I like things to be a little more permanent, I hate messy wiring (an occupational hazard with breadboard layouts) plus I can use them in other projects. I'm still planning on a little headphone amp, two of the Bandaxall test boards can be used in that project for example.

Right, so as it stands right now, the plastic sheet has just finished printing off, now I'm going to take it to the utils room and give it a quick blast with my hot air stripper, and then I'm going to etch a board

If it works, and I'm happy with it (I was when it was module based), I shall share details here.

Thanks to all those who have followed me so far - it's appreciated.

Hi
I used the circuit posted here on my amp. I used the LM358 or LM 324. I run them at + - 16 that's pretty much max but I haven't had any go out. I used them for the drivers in the 80W amp too and it sounds great as long as the volume and bass is low. They were really not good for drivers as on high volume they will still clip. I like those chips and use them for lots of things. It can still go to where your ears begin to hurt before you hear any clipping. LOL

Final point, I used to work with early OpAmps. They were not designed for amplifying but for instrumentation in analogue computers. Early OpAmps (and some of the cheaper amps today) tend to suffer from crossover distortion. This becomes apparent when playing very low AF signals.

The solution is to change them from Class-AB push-pull to Class-A. Connect a 1K0 resistor from the +10V terminal to the output pin.

BR Harry

Thanks for that Harry

On the bright side, I'm learning a lot about op-amps!

The only things I have to test distortion with is my ears. The original circuit certainly produced a lot of bass in that 20% of arc, before finally getting heavily distorted and then seemingly, shutting down - audio output fell to zero. Tuning the bass pot was like fine tuning the comparator in my continuity tester - a fraction too far and it would avalanche into silence. I doubt it was over driven, but I suppose it's possible. I'm driving the latest incarnation from the same source, and I have to be honest, I'm happy with it. I shall try a BC109 gain stage as well as a TL072 and compare, but if I had to go with these little 741's then that wouldn't be too much of a heartbreak. I'm sure it will make a difference when I use the CD player rather than the laptop, although it's good, my laptop isn't in the same class as my Technics CD deck.

The unit doesn't distort at a LINE input level, however, if I up that by around 50%, then it gets noticeable to me, so yeah, keeping a tight control on the input level is a good idea. I am running the 741's at +12V though, so maybe I should have some headway?

With no input, the power amp volume at maximum, there's nothing - not even a hiss or a buzz. None that I can hear, anyway. I'm very pleased with it to be honest, I don't understand why Marantz never built this into the amplifier in the first place - my old Pioneer amp had bass and treble.....

The original plan was to build a pass-through active tone control box, that idea evolved into making a 1 Watt stereo headphone amplifier, doing away with the need to use the main power amp. We shall see It's a rats nest at the moment, makes spaghetti look neat lol - I shall design some custom PCBs to solve that problem

I have been watching this post (on and off) and I think it is time for me to chip in a couple of comments.

In the 80s I used to build stereo amplifiers and sell them. I used the BD139/140 output stage, CB108/109 throughout and also a Baxandall tone controll. I found that the tone control tended to distort so I changed to a 100% passive design. I found that when the tone controls were taken to one extreme there was a tendency to distort. I reasoned that due to the boost there was exessive gain.

I used 100K pots:
22k + pot + 2k2 for the bass circuit (to ground) with a single 47nf across the pot.
2n2 + pot 0 22n for the trebble circuit (to ground).

Output mix was from the trebble pot via 2K2 and the base pot via 6k8.

The last two were selected to give a flat response.

I built about 30 or 40 of these over the years I worked at PYE telecom in Cambridge. I did try the active version a couple of times when I began to experiment with OpAmps but again gave up in disgust. The tone control gives a 20dB trebble boost/cut and 25dB bass boost/cut.

There was a lot of disco beat around in the 80s and people wanted excessive bass boost without distortion. So I chose a stage with higher "flat" loss.

Now back to your circuit - if the original circuit was correctly wired then you may find that reducing the input level would reduce the distortion. The "flat" gain of the OpAmp is controlled by the feedback through the tone circuit and this can be quite heavy, causing distortion at frequency boost extremes. Attenuating the input cures it but also causes more noise due to the required gain of the following amplifier.

Just a few observations from my own past :-)



BR Harry

Right, I have taken the modular approach.

First up was a mono Baxandall circuit, input (via a 10uF cap) from my laptop, then output direct to the power amp. This circuit worked, but the volume was down - as to be expected.

A stereo version was then built - and tested, working properly. This is now on a single bit of strip board.

Then, a small amp was built, based around a 741 op-amp - this boosted the output signal very nicely, no cutting out as the bass was advanced as per the previous circuit.

A second amp was built on a separate bit of strip board and the whole thing connected up - working very nicely, but jeez, what a birds nest of wires!

So this completes prototype 1 - I shall build another feedback amplifier based on a pair of TL072's and see how much better the quality is. 741's were pretty impressive though, I have to say.

If the TL072 design works, I shall build a final test amp using a pair of BC109C's before making a final decision on the final unit.

Quite why the first attempt failed so miserably is beyond me, although I suspect Ivan had the answer to that. Maybe a wiring error, who knows - this 741 circuit is pretty much identical except for the cap goes directly to pin 3, and the bias is set to half +ve via a pair of 100K resistors rather than the 68K used in the original.

Just a question that remains - the TL072 is a dual op-amp and I intend to only use one of the pair. Do I simply ignore the unused section or tie both inputs to ground? Or indeed, neither lol!

73's, from one happy DragonForce

Hi DragonForce
So far when I have had one of these that didn't work right I have always been able to find the problem by adjusting the suspected parts. Sort of trial and error, My favorite method. LOL
I have also build a 10 channel equalizer that I though sounded read good. The biggest problem I had with that one was to find the right slide pots and capacitor values.

Hi Dragon Force,
it is very strange that the capacitor on the input connector did not help.
In any case, attach a DC voltmeter between pin 6 of the opamp and GND. The reading must be Vcc/2 in any position of the bass potentiometer, with the signal source attached. Maximum deviation is in the millivolts region. If this is not the case, you have wrong biasing and you must stop DC leaking. Keep in mind, that the difference of DC voltages on pins 2 and 3 is amplified and added to the signal on pin 6.

BR from Ivan

Hi Ruud

The op amp is a 741, just for testing purposes, but the original circuit diagram was taken from a book called 52 Projects Using IC 741. I mention this since you mention FET op-amps. I have some TL072's on order.

I did try using an input capacitor, 10uF - but the circuit still died at around 20% swing on the bass control pot.

Output was taken from the headphone output on my laptop and driven at various levels. The signal quality at the output of the 741 was OK, so long as I didn't turn up the bass.

Ivan mentioned the output capabilities of the 741 isn't so good, but my phones are 32R and the output was pretty good.

The signal source of the finished project will be LINE. If necessary I can always build a small buffer amplifier.

Many thanks to all those people who have given their time to help me



I plan to rebuild the circuit again using a BC109 amplifier, and also the TL072 to see what I can make happen.

In principle this circuit should work good as it is.
The OpAmp is biased at about 1/2 Vb, so no problem there.
It should also be no problem to vary the resistance between the output (pin 6) and the inverting input (pin 2). No need for a capacitor in the feedback loop. The input however also has the 1/2 bias, so an input capacitor will be necessary.
With a FET OpAmp, the feedback resistance can be anything between zero ohms and a couple of mega ohms.
The only thing to take care of is that there is no DC from a previous stage. You could easily measure this at the output; simply connect a voltmeter and vary the settings of the tone controls.
Also the circuit should be driven from a reasonable low impedance.
In some cases circuits like this will oscillate with a capacitive load. A 100 ohms resistor in series with the output usually cures this.

When I was young, I made myself a 10-channel stereo audio mixer, with tone controls of this type in every channel. At the time the price of those mixers was very high, if you could even buy them. The quality was so good, that I even recorded material with it that was later published on records!

Jalex2 wrote:Hi DragonForce
I built a few circuits like these and had great results on all of them. I had a computer that had a terrible on board sound system and I built one of these and it sounded as good as any expensive system I have heard from then on. I just put it on the line between the computer and the Amp and it worked just fine there. I was never sure but that problem could have been because of a poor match between the computer and amp and the circuit took care of that.

Hi Jalex, hope you're well

The reason I want to build this is because my (new to me) hifi amplifier does not have a bass or treble control. Quite why this is defies logic - to me at least. In order to get the audio sounding the way I like it (boosted around 6 KHz due to a hearing deficiency on my part), I'm going to need to put a graphic equaliser into the system. That's "simple" enough, but will require me to use the tape monitor function of the amplifier in order to use it. All Marantz had to do was bung in a few pots and a few resistors and caps, and it wouldn't be PITA it's being!

What I propose to do is build an amplifier that allows me to switch the outputs of various hifi separates (Tuner, CD, Tape, Phono and Aux), allow me to set the balance, bass and treble to my liking, and feed the boosted output to the aux input of the Marantz amplifier. Now it will need to be boosted a little to get the signal back up to the level it should be - the Baxandall filters being a little attenuative (is that even a word?). I won't need to use the controls on the amplifier then, just set it to AUX, set the volume and it's good to go.

The problem I have found with the experimental (mono) version is as described in my first post. Ivan did his best to explain it, I've read Harry's Op-amp tutorial, and still I don't really understand what Ivan is trying to tell me. I've added a link to an alternate active Baxandall version, which Ivan likes, but apart from an input capacitor, I can see no real difference in the two circuits. Both have the output from pin 6 being fed back to the Baxandall network without a DC iso cap, and I did try adding one at the input to the version I built, with no change in the result - mine cuts off the output when the bass control gets to maybe 20% of its' travel.

Now I've found several versions that use a single NPN transistor in place of an op-amp. It's funny, because in the original paper that Baxandall wrote back in the 50's, his original design used valves/tubes. Both Baxandall's original and the transistor versions I've seen use a DC blocking cap in the feedback loop - just as Ivan suggested. The op amp designs do not seem to have this cap. Now, Ivan said about using one of these, and so I'll have to get out my breadboard and carry on playing with this. It's not going to beat me LOL!

I have some new caps, several BC109C low noise transistors and a few TL072 op-amps on the way, along with sone dual gang pots. This week is going to be fun!

Hi DragonForce
I built a few circuits like these and had great results on all of them. I had a computer that had a terrible on board sound system and I built one of these and it sounded as good as any expensive system I have heard from then on. I just put it on the line between the computer and the Amp and it worked just fine there. I was never sure but that problem could have been because of a poor match between the computer and amp and the circuit took care of that.

Hi DragonForce,
"I've looked at
several circuits of similar design, and they all follow pretty much the
same basic layout, with the output of the op-amp being fed back to the
Baxandall network directly as I have done."
The opamp circuits are often drawn as with symmetrical PSU, so the inputs and outputs are settled at GND (maybe with some offset). Using single rail PSU, you must "lift" all the circuitry connected to Vcc/2.
"It's my understanding that
the op-amp isn't there to simply boost the output from the entire
network, it's supposed to use negative feedback"
Yes.
"to alter the gain of the
amplifier"
No. It changes the frequency characteristics of the amplifier.
"I assume the bias on the non-inverting input has to vary
with the potentiometer settings."
NOOOO !!!
"Ivan, thanks for your time "
Not at all. I am glad if I could help you.

http://www.learnabout-electronics.org/Amplifiers/images/baxandale+amp.gif
is one such example - the basic layout doesn't differ too much from
whet I have drawn below - of course, I am assuming it works...... "
Yes, it seems it should work. Mention that GND is connected to Vee, the input and output connector. Nowhere else. All the circuitry "floats" at Vcc/2. The input and output are separated by capacitors.

BR from Ivan

Ivan wrote:Hi DragonForce,
the bias of the noninverting input (pin 3) is OK, although I do not see the reason for dividing the upper resistor into two parts. The electrolytic cap can be attached directly to the pin 3.
The bias of the inverting input (pin 2) is taken from the output (pin 6) and goes via a 10K resistor, the upper poti and another 10K resistor. Therefore it depends on the poti setting! Moreover, there is no DC separation on the corrector input, so the bias depends on the device attached ! The dependance of the DC bias at pin 2 on potentiometer setting corresponds with the symptoms: at one end the circuit works nearly OK. When the poti moves, strong distortion appears as the bias approaches one rail. At the another end, the bias gets totally insufficient and the AC transfer disappears at all.
My advice is:
1/ separate the input by a capacitor;
2/ change the circuit so that the DC bias on pin 2 is always the same as on pin 6 and AC only is passed via the corrector circuitry.

BR from Ivan

I tried a capacitor on the input, made no difference.

I never really have understood the use of op-amps for this kind of use, it's not something I've ever needed to understand. Now as a comparator, yes - I certainly understand that!

I've looked at several circuits of similar design, and they all follow pretty much the same basic layout, with the output of the op-amp being fed back to the Baxandall network directly as I have done. It's my understanding that the op-amp isn't there to simply boost the output from the entire network, it's supposed to use negative feedback to alter the gain of the amplifier, so I assume the bias on the non-inverting input has to vary with the potentiometer settings. I find it all very confusing.

I think I'll have a few more attempts using various circuits I find online and then share results, this isn't going to beat me.

Ivan, thanks for your time

http://www.learnabout-electronics.org/Amplifiers/images/baxandale+amp.gif is one such example - the basic layout doesn't differ too much from whet I have drawn below - of course, I am assuming it works...... I suppose it might not lol

And 3/ low impedance headphones present too big load to common opamps. You need a small PA stage or at least an opamp with current buffer (complementary pair of bipolar transistors on the output, inside the feedback loop) to drive modern headphones.

BR from Ivan

Hi DragonForce,
the bias of the noninverting input (pin 3) is OK, although I do not see the reason for dividing the upper resistor into two parts. The electrolytic cap can be attached directly to the pin 3.
The bias of the inverting input (pin 2) is taken from the output (pin 6) and goes via a 10K resistor, the upper poti and another 10K resistor. Therefore it depends on the poti setting! Moreover, there is no DC separation on the corrector input, so the bias depends on the device attached ! The dependance of the DC bias at pin 2 on potentiometer setting corresponds with the symptoms: at one end the circuit works nearly OK. When the poti moves, strong distortion appears as the bias approaches one rail. At the another end, the bias gets totally insufficient and the AC transfer disappears at all.
My advice is:
1/ separate the input by a capacitor;
2/ change the circuit so that the DC bias on pin 2 is always the same as on pin 6 and AC only is passed via the corrector circuitry.

BR from Ivan

Hi Ivan

I've updated my previous post with a circuit diagram. Output from pin 6 taken to headphones via a 100uF cap, not shown on the diagram.

Hi DragonForce,
how is the biasing of the inputs to the half of supply voltage done ? The schematics would be fine.

BR from Ivan

Here's one for the audio experts

I've built an active tone control circuit. It seems to work in that the treble control does indeed control treble, and the bass control does indeed control lower frequencies. But only to a point - when the bass control is turned up around a third of the way, the sound heavily distorts and eventually dies. The Treble control is fine over its' entire swing, and the setting of this control does not have any effect on where in its swing the bass control causes the circuit to die.

Now I've robbed the circuit from one of those little Babani books, 52 Projects Using IC 741, but Rudi & Uwe Redmer. I have modified some of the component values simply because I do not have 3.3 and 33nF available at the moment, so I've used 4n7's and 47n's. Both variable resistors are 100K LIN law devices.

Now I'm assuming that my component modifications will only have an effect on where in the frequency spectrum the filtering takes place, and not cause the circuit to die. Is it worth me giving the circuit another try with the specified values, or is something else going on here. I suspect something is happening to the 741 side of things.

The circuit can be run on anything from 9 - 30 Volts - I'm using a regulated 12V supply, using 9V seemed to make no difference. It's a single rail supply.

The point of this circuit is to build myself a decent headphone amp to use with a CD player - of course the final project won't use a 741 What does the experts think is a decent quality audio op amp?

If you guys would like a circuit diagram, I'll make a scan of the book page, although it's a standard Baxandall design.