Re: Re: No gain, No Pain...
sangram,
If I understand Kuei's impliment correctly, 18k resistor comes from the following reason: there are three resistors (two 220ks and one 22k) which are in parallel connection at amp's inverting input. So, the amp's inverting input "sees" the sum of these resistors which comes with 18.333k (18k in short). For impedance matching btwn inverting and non-inverting inputs, 18k resistor is placed btwn non-inverting input and ground.
If you want to use 56k resistors, replace 220k and 22k resistors with 56k and 5.6k, respectively. The resistor at non-inverting input will be 56k//56k//5.6k = 4.666k (4.7k in short).
Please correct my calculation, if there's something wrong. I'm still in a learning stage of diy audio. So, there could be something wrong.
sangram said:
sangram,
If I understand Kuei's impliment correctly, 18k resistor comes from the following reason: there are three resistors (two 220ks and one 22k) which are in parallel connection at amp's inverting input. So, the amp's inverting input "sees" the sum of these resistors which comes with 18.333k (18k in short). For impedance matching btwn inverting and non-inverting inputs, 18k resistor is placed btwn non-inverting input and ground.
If you want to use 56k resistors, replace 220k and 22k resistors with 56k and 5.6k, respectively. The resistor at non-inverting input will be 56k//56k//5.6k = 4.666k (4.7k in short).
Please correct my calculation, if there's something wrong. I'm still in a learning stage of diy audio. So, there could be something wrong.
Unity-gain inverting clone
Built the circuit as per Keui's schematic (using TDA 2030, not the LM series), with the resistor values at 5.6 (- to GND), 4.7 (+ to ground) and 56 K (feedback and input).
How does it sound? In a word, brilliant. Deep and rich bass and controlled and clean top end, the non-inverting clone sounds like cough syrup now. Sweet at first taste, but...
Now here's the issue: I cannot get the unity-gain amp to play loud (I have 93 dB speakers for my main listening). Excessive input is causing the input to overload, creating a nasty distortion at high levels. The preamp has to be cranked up really hard and therefore is giving distortion adding to the problem. In addition for some reason the protection circuits are getting activated at high input levels creating the spiking sounds we all know too well.
I have tried the circuit again without the gainfaking resistor, instead setting a gain of 10 through 5.6 K/56 K resistors with inverting input. Sounds better than the other amp, but not as good as unity gain configuration. I could add another preamp, but that would defeat the purpose, sort of...
I have also managed to destroy my first IC. Five to go. A collision between two prototypes in open air with P2P wiring, some kind of short, and lo and behold, one's dead.
Anyway I am going to use inverting for all the chips now. And probably a small tagboard for the components maybe...
Yes I have used 4.7K. I of course used a very crude way of getting to the value, wheras you have been more scientific. Unfortunately I did not get a chance to complete too many tests on the inverting clone as it died within 10 minutes of commissioning. It will take me an hour to rig up another, and I don't have the energy to pull out a new chip. So the chip hitherto used for the non-inverting clone shall be rewired to make an inverting gain amp.
Built the circuit as per Keui's schematic (using TDA 2030, not the LM series), with the resistor values at 5.6 (- to GND), 4.7 (+ to ground) and 56 K (feedback and input).
How does it sound? In a word, brilliant. Deep and rich bass and controlled and clean top end, the non-inverting clone sounds like cough syrup now. Sweet at first taste, but...
Now here's the issue: I cannot get the unity-gain amp to play loud (I have 93 dB speakers for my main listening). Excessive input is causing the input to overload, creating a nasty distortion at high levels. The preamp has to be cranked up really hard and therefore is giving distortion adding to the problem. In addition for some reason the protection circuits are getting activated at high input levels creating the spiking sounds we all know too well.
I have tried the circuit again without the gainfaking resistor, instead setting a gain of 10 through 5.6 K/56 K resistors with inverting input. Sounds better than the other amp, but not as good as unity gain configuration. I could add another preamp, but that would defeat the purpose, sort of...
I have also managed to destroy my first IC. Five to go. A collision between two prototypes in open air with P2P wiring, some kind of short, and lo and behold, one's dead.
Anyway I am going to use inverting for all the chips now. And probably a small tagboard for the components maybe...
Yes I have used 4.7K. I of course used a very crude way of getting to the value, wheras you have been more scientific. Unfortunately I did not get a chance to complete too many tests on the inverting clone as it died within 10 minutes of commissioning. It will take me an hour to rig up another, and I don't have the energy to pull out a new chip. So the chip hitherto used for the non-inverting clone shall be rewired to make an inverting gain amp.
JAZZ2250 said:
sangram,
If I understand Kuei's impliment correctly, 18k resistor comes from the following reason: there are three resistors (two 220ks and one 22k) which are in parallel connection at amp's inverting input. So, the amp's inverting input "sees" the sum of these resistors which comes with 18.333k (18k in short). For impedance matching btwn inverting and non-inverting inputs, 18k resistor is placed btwn non-inverting input and ground.
If you want to use 56k resistors, replace 220k and 22k resistors with 56k and 5.6k, respectively. The resistor at non-inverting input will be 56k//56k//5.6k = 4.666k (4.7k in short).
Please correct my calculation, if there's something wrong. I'm still in a learning stage of diy audio. So, there could be something wrong.
There seems to be a deal of confusion over this circuit configuration.
In fact this is a conventional technique for getting a signal gain less than the minimum stable loop gain with an op-amp.
I have expanded the math in the attached diagram. In essence the configuration provides an above unity loop gain (in this case approx. 11) but at the same time attenuates the input signal in exactly the same proportion, hence preserving the signal gain.
You can see this effect in the way the effect of R3 cancels out in the signal gain equation below.
mc2
In fact this is a conventional technique for getting a signal gain less than the minimum stable loop gain with an op-amp.
I have expanded the math in the attached diagram. In essence the configuration provides an above unity loop gain (in this case approx. 11) but at the same time attenuates the input signal in exactly the same proportion, hence preserving the signal gain.
You can see this effect in the way the effect of R3 cancels out in the signal gain equation below.
mc2
Attachments
Re: simple follower
They would be instable at low gains, and unity gain.
It is same with NE5534 (not NE5532 which is unity gain stable)
If you use NE5534 OP at lower gain than 3-5,
it can begin to oscillate.
When designers make OP or Amp-IC, they put in some compensation cap
but only so much capacitance that the IC will be stable,
at the application/gain that they intend to use it for.
Amplifier IC:s are made by the industry for special purposes, mostly.
They are aimed for a special user. A part of the market
that are suppose to buy the product.
To much capacitance is bad, and lower the speed and performance
of that IC.
So if they make a power amp which should work at like gain=20
which is the normal, they put in a compensation cap
into the chip that makes sure it is stable at gain=20.
That means that if you use it for lower gain, say 10,
it could be unstable.
--------------------------
The trick with OPamp NE5534, is to use an external extra cap
between pin 5 and 8. Those pins have been made accessable in NE5534.
A second hand solution is to put a comp cap parallell
with the feedback resistor. From output to inverted input.
This will slow down the OP and make it stable at Gains below 3-5
--------------------------
It would be possible to use some external compensation
for 5-pin Power OP amps like TDA2050 LM3886 etc.
But at the expense of some speed and performance.
And My opinion is that this remaining operation, would not be so low
that you can hear any difference.
--------------------------
/halo - loves NE5534
more than NE5532 (which has somewhat lower prestanda, because of more internal compensation capacitance)
but instead is DUAL OP and stable at Gain=1
JAZZ2250 said:
As Kuei Yang Wang saysKuei Yang Wang said:
They would be instable at low gains, and unity gain.
It is same with NE5534 (not NE5532 which is unity gain stable)
If you use NE5534 OP at lower gain than 3-5,
it can begin to oscillate.
When designers make OP or Amp-IC, they put in some compensation cap
but only so much capacitance that the IC will be stable,
at the application/gain that they intend to use it for.
Amplifier IC:s are made by the industry for special purposes, mostly.
They are aimed for a special user. A part of the market
that are suppose to buy the product.
To much capacitance is bad, and lower the speed and performance
of that IC.
So if they make a power amp which should work at like gain=20
which is the normal, they put in a compensation cap
into the chip that makes sure it is stable at gain=20.
That means that if you use it for lower gain, say 10,
it could be unstable.
--------------------------
The trick with OPamp NE5534, is to use an external extra cap
between pin 5 and 8. Those pins have been made accessable in NE5534.
A second hand solution is to put a comp cap parallell
with the feedback resistor. From output to inverted input.
This will slow down the OP and make it stable at Gains below 3-5
--------------------------
It would be possible to use some external compensation
for 5-pin Power OP amps like TDA2050 LM3886 etc.
But at the expense of some speed and performance.
And My opinion is that this remaining operation, would not be so low
that you can hear any difference.
--------------------------
/halo - loves NE5534
more than NE5532 (which has somewhat lower prestanda, because of more internal compensation capacitance)
but instead is DUAL OP and stable at Gain=1
Nielsio said:
Hi Niels
Just want to reinforce KYW's (= u-know-hoo) comments on keeping PS Earth & Signal Earth separate. They are separate but also tied together. Sounds confusing but is in practice quite simple. You may find this wiring diagram of practical interest, even if the components values etc are not the same, the Earthing is.
An externally hosted image should be here but it was not working when we last tested it.
"G" is both Chassis Ground & Power Supply Ground (that's how I do it).
"S" is Signal Ground and is kept away from "G".
There is more here:
http://members.ozemail.com.au/~lisaras/ps.htm
Joe
Peter Daniel said:
Strictly yes. But does anyone care about absolute phase?
On the other hand + also indicates which of the outputs are 'hot'. Then if you want to manintain absolute phase, then reverse speaker leads.
Incidentally, in many three-way speaker systems the midrange driver is wired out of phase in order to sum correctly at the crossover(s). Should the midrange - where most of the music info is derived - or the bass deternmine IF the speaker terminals should be + or - ????
I use + to indicate 'hot' as it's safer to let the end user know.
Joe
PS: I'm not saying absolute phase is inaudible, merely most don't want to care about it as it can drive you nuts. Like reversing speaker leads nearly everytime you change disks... no wonder some of think we're nuts.
We ARE!
Joe Rasmussen said:
2 or 3 weeks ago, I was experimenting with a proper and reversed connection from IGC to the speakers
and there was a noticable difference.
When amp's output was connected to + on a speaker, the sound seemed to be smoother and less bright, but at the expense of air at the very top end. With amp's output connected to - terminal on a speaker, there was much more air and ambience and this was indeed a connection I preffer (and also correct one).
Peter Daniel said:
No arguments on that score. What we need is phase switches to go with the gainclones with no more additional active circuitry... any ideas?
Just to prove I am absolute phase conscious: Anyone familiar with Vacuum State Electronics RTP Phono Preamp, and I've made a number of those, has separate phase switches for Left & Right and no added circuit as it's fully differential.
You can see the separate Phase Switches:
An externally hosted image should be here but it was not working when we last tested it.
Wouldn't it have been nice if everything in the recording chain was kept absolute phase? Or if any inverting stage was balanced by another inverting stage through the process. And ALL instruments in the same recording as well, so that we have the percussion (cone moves IN) and horns (cone moves OUT). Perhaps it is this leading edge we can hear?
So we can avoid this happening: "Oh, if I connect my speakers this way the percussion sounds great but the horns not quite right, then I reverse leads and the horns sounds great but the drums lose their magic."
Like I said, it's enough to make you go nuts!
Betcha, after a while you just want to forget it and go on playing music rather than equipment. Still, phase switches would be nice. Yeah I know, mea culpa.
Joe
Joe Rasmussen said:No arguments on that score. What we need is phase switches to go with the gainclones with no more additional active circuitry... any ideas?
Sure.
Use a line input transformer. Something like the Jensen JT-11P4-1 (which actually has a wee bit of voltage gain). Seeing as the primary is inherently balanced, you can readily change polarity with a DPDT switch. You also get galvanic isolation, the ability to readily accept balanced or unbalanced inputs, and over 90dB of common-mode rejection even from unbalanced sources.
se
Steve Eddy said:
Yep, you can do it that way... but then I have friends who would avoid line level (or low level) transformers like the plague. Me, I am not that judgmental.
BUT is it worth it, adding a major component like that, just so you can change absolute phase? Only the individual can answer that!
But using it for balanced (I prefer using the word 'differential', even if they are not excactly the same thing), don't worry, I'm a convert.
Joe
