Author Topic: "Left brain going right brain" amplifier  (Read 4866 times)

Offline Ampdog

Re: "Left brain going right brain" amplifier
« Reply #30 on: February 11, 2014, 12:55:06 AM »
Jim,

Most interesting thought patterns. I am turning this around in my brain (which will not be for much longer now seeing the hour ....) so no immediate comments. I am experiencing a warning signal about practicality, but as you said on paper for now.

We are experiencing thinking problems because we argue from a voltage based point-of-view. As mentioned there is very much a current feed logic (but that is immediately compromised by loudspeaker control in the l.f. region - still, has been done). One has to consider GearSlave's point, as well as how often this will be needed (back to my aversion of funny impedance loudspeakers; not to repeat). In other words, also bringing into the mix how often this feature of high-current-low-voltage will be required during a typical performance  (.... if you meant to include those loudspeakers that only dip at certain frequencies, as I see most do).  This again brings in the large-reservoir-capacitor factor also mentioned.

For those not in the know, a comparison of current feed vs. the more common voltage feed was experimentally done long ago by Dunlavy at a university - and the listeners preferred the current feed by a small margin! As someone said: It is amp-turns which drive the loudspeaker!  But that is quite another story.
Audio must be the only branch of engineering where lack of basics' knowledge is considered a superior form of wisdom. (Anon)

Offline JimGore

Re: "Left brain going right brain" amplifier
« Reply #31 on: February 11, 2014, 10:09:51 AM »
I suppose this idea of mine is not just out of the box thinking, it's also out of the house thinking as it has left most of you feeling dazed and confused  :roll:

It is truly such a foreign concept?  To me it makes so much sense that I can't understand why it isn't done all the time.

Offline ingenieus

Re: "Left brain going right brain" amplifier
« Reply #32 on: February 11, 2014, 01:46:42 PM »
The answer is this:  We start with the *JIM* amplifier as a base.  We allow it to measure the connected speaker's minimum impedance once during the power-on sequence, then adjust the internal gain of the *JIM* amplifier such that it will provide full power into that minimum load without any clipping.

OK I get it. To measure the impedance, you need to measure the voltage and current from 20Hz to 20kHz. More or less continuously, otherwise a peak or dip in the loudspeaker impedance curve might be missed. In other words a sine wave generator is needed. It needs to be low distortion otherwise the harmonics will mean that the measurements are not accurate. Sweep through the frequency range, measuring voltage and current along the way. Calculate impedance, store values. Get minimum value. Calculate gain needed. Adjust resistance in feedback network to suit. Not easy but it can be done. A micro processor would be useful.

The thing is - all this extra hardware needed costs money. If you put that money into a bigger PSU you can build a *ROM* amp and there is no need for the complicatidrivyon.

The hard core audiophiles will not look on this kindly. For one, to measure current you need to add a resistor or Hall sensor in series between the output and speaker. You can switch it out with a relay after the measurement but that relay is not going to be cheap.

Here is one way to measure current (with a Hall sensor) and voltage. The two are then fed into an analog multiplier that drives a power meter. What you propose is a variation on this.


Offline valvesound

  • Trade Count: (+5)
  • AVForums Grandmaster
  • *****
  • Posts: 1,406
  • Total likes: 0
  • A EUPHORIC EXPERIENCE
Re: "Left brain going right brain" amplifier
« Reply #33 on: February 11, 2014, 02:21:33 PM »
I suppose this idea of mine is not just out of the box thinking, it's also out of the house thinking as it has left most of you feeling dazed and confused  :roll:

It is truly such a foreign concept?  To me it makes so much sense that I can't understand why it isn't done all the time.


I think it is more a case WHY? rather than understanding HOW?  :tongue:
Please contact me if you are interested in any of my products to transform your enjoyment and experience of high end stereo

Offline JimGore

Re: "Left brain going right brain" amplifier
« Reply #34 on: February 11, 2014, 02:55:18 PM »
Quote
OK I get it. To measure the impedance, you need to measure the voltage and current from 20Hz to 20kHz. More or less continuously, otherwise a peak or dip in the loudspeaker impedance curve might be missed. In other words a sine wave generator is needed. It needs to be low distortion otherwise the harmonics will mean that the measurements are not accurate. Sweep through the frequency range, measuring voltage and current along the way. Calculate impedance, store values. Get minimum value. Calculate gain needed. Adjust resistance in feedback network to suit. Not easy but it can be done. A micro processor would be useful.

The thing is - all this extra hardware needed costs money. If you put that money into a bigger PSU you can build a *ROM* amp and there is no need for the complicatidrivyon.

That's right! (sigh of relief!)

There is already a microprocessor in the mix, so no additional cost is associated on that front ;D

As for the PSU - I am already using a pretty substantial PSU, so any larger will not really be helping me out here as it becomes a discussion between having 1kVa vs 2kVa - pretty pointless for home use.

Quote
The hard core audiophiles will not look on this kindly. For one, to measure current you need to add a resistor or Hall sensor in series between the output and speaker. You can switch it out with a relay after the measurement but that relay is not going to be cheap.

Audiophiles?  Who are they?  I want to build this thing for me and listen to it myself.  If anyone else doesn't like it then it's none of my concern  :flame:

Anyway, I am not sure why "they" will not like it - the hall effect sensor is non invasive.  It is not connected in any way (other than magnetic) to the actual amplifier circuit at any point.

Quote
Here is one way to measure current (with a Hall sensor) and voltage. The two are then fed into an analog multiplier that drives a power meter. What you propose is a variation on this.

Thanks very much!


I think it is more a case WHY? rather than understanding HOW?  :tongue:

Why?  Because I have arms full of amplifiers that work on the *ROM* model.  My main amplification is exactly this, and is to my ears at least the best amplifier I have ever listened to.  To me there is no point in building just another same old amplifier.  It's been done to death.

Also, I would much rather have 230W of power drive where I need it as opposed to 70W with the addition of low impedance drive bragging rights.  Like I said - I already have the low impedance bragging rights kind of amplifiers anyway.

Lastly, how little progress will we have made as a society if people didn't come up with *different* ideas now and then?  Not sure about you, but I am far happier driving my car to work than I would have been riding a horse.  I am sure the designers of the first cars were asked just this same question - not that I am doing anything that interesting tho...

Thanks again for all your help Gentlemen.  Like I said in the opening post of this thread - I often feel the need to do the different thing.

How great will it be (for me at least) if this thing actually works the way I imagine it to?  If it sounds good too (as I am sure it will), then that's even more of a feather in my cap.

Kind Regards,
Ian.

Offline Shonver

  • Trade Count: (+8)
  • AVForums Grandmaster
  • *****
  • Posts: 10,521
  • Total likes: 22
  • Criss-cross rhythms that explode with happiness
    • Elipse (under construction)
Re: "Left brain going right brain" amplifier
« Reply #35 on: February 11, 2014, 04:30:00 PM »
So far we only have "how to measure impedance". It is not yet clear how adjusting gain of the amplifier does anything, or how you will link the measured impedance to selection of gain (maybe talk to us a bit about that).
________________

DON'T PANIC
Capie

Offline JimGore

Re: "Left brain going right brain" amplifier
« Reply #36 on: February 11, 2014, 05:03:49 PM »
So far we only have "how to measure impedance". It is not yet clear how adjusting gain of the amplifier does anything

The only thing the gain adjustment does is ensure maximum unclipped output for the given minimum impedance load.

how you will link the measured impedance to selection of gain (maybe talk to us a bit about that).

What I will most likely do is build the amplifier such that it allows manual adjustment of the gain.  When its in working condition, I will measure the maximum power before clipping into a couple of points using fixed resistors as the load.  Perhaps one measurement every 0.5 ohm or something to that effect will be good enough.  I will make these measurements whilst keeping the input levels constant at a reasonable level such as 1 or 2V RMS.  I will then manually adjust the gain to the point where it just starts clipping, and take note of the resistor values used in order to achieve the required output.

From there I have a table which provides the gain settings for each 0.5 ohm step.  These will show the levels where the amplifier just starts to clip, so I will need to build some margin in - perhaps to the order of 3% drop for each gain ratio.  These figures I can implement into the microprocessor which will then control the gain as needed for the driven load after the microprocessor has read the minimum imp of the connected load.

Offline Shonver

  • Trade Count: (+8)
  • AVForums Grandmaster
  • *****
  • Posts: 10,521
  • Total likes: 22
  • Criss-cross rhythms that explode with happiness
    • Elipse (under construction)
Re: "Left brain going right brain" amplifier
« Reply #37 on: February 11, 2014, 09:53:35 PM »
I don't know by how many dBs you'd have to change the gain to do this, but my recommendation is to leave the gain fixed, and just fiddle with attenuation (volume control). Reason being that changing the closed loop gain of the amplifier also changes its operating conditions; that's introducing a variable that may have unpredictable consequences. To put it another way, it is extra work to ensure that the amplifier's performance is unaffected over the range of gains that you will be adjusting it through.

But let me not discourage you. This may be a journey you must make, for better or worse.
________________

DON'T PANIC
Capie

Offline scrarfussi

Re: "Left brain going right brain" amplifier
« Reply #38 on: February 12, 2014, 07:22:52 AM »
sounds like a interesting idea . Just some things i am trying to figure out .  I  understand what it is you are trying to achieve  but maybe after 3 pages i am wrong and still slightly confused . pardon my slowness extreme amateur
firstly  you want a amplifier that delivers maximum at that given load . meaning you wanna sense the load and have the amplifier adjust its output accordingly  .
but then since most speakers the impedance varies depending on the frequency . i maybe wrong with this assumption but i would assume that if it goes below 8 ohms on some frequencies it goes  a bit higher on other frequencies as well like 10 ohms .
So basically you are going to have a circuit monitoring these changes and adjusting power supply accordingly to give you that kind of instantaneous power ?
 
suum cuique pulchrum est


"Then Said my heart here will i take my rest."

Offline valvesound

  • Trade Count: (+5)
  • AVForums Grandmaster
  • *****
  • Posts: 1,406
  • Total likes: 0
  • A EUPHORIC EXPERIENCE
Re: "Left brain going right brain" amplifier
« Reply #39 on: February 12, 2014, 08:25:29 AM »
I don't know by how many dBs you'd have to change the gain to do this, but my recommendation is to leave the gain fixed, and just fiddle with attenuation (volume control). Reason being that changing the closed loop gain of the amplifier also changes its operating conditions; that's introducing a variable that may have unpredictable consequences. To put it another way, it is extra work to ensure that the amplifier's performance is unaffected over the range of gains that you will be adjusting it through.

But let me not discourage you. This may be a journey you must make, for better or worse.


I agree with Shonver. Two things that will change with gain adjustment is the bandwidth and the distortion figure of the amplifier. Maybe the bandwidth is high enough and the distortion low enough to not notice (hear) it, but vary it will.
Please contact me if you are interested in any of my products to transform your enjoyment and experience of high end stereo

Offline JimGore

Re: "Left brain going right brain" amplifier
« Reply #40 on: February 12, 2014, 10:44:20 AM »
sounds like a interesting idea . Just some things i am trying to figure out .  I  understand what it is you are trying to achieve  but maybe after 3 pages i am wrong and still slightly confused . pardon my slowness extreme amateur
firstly  you want a amplifier that delivers maximum at that given load . meaning you wanna sense the load and have the amplifier adjust its output accordingly  .

That's correct.

Quote
but then since most speakers the impedance varies depending on the frequency . i maybe wrong with this assumption but i would assume that if it goes below 8 ohms on some frequencies it goes  a bit higher on other frequencies as well like 10 ohms.
So basically you are going to have a circuit monitoring these changes and adjusting power supply accordingly to give you that kind of instantaneous power ?

This part of your question is somewhat frustrating to me, but I will answer it one more time (it has already been answered at least 3 or 4 times by me in this thread).

I am going to measure the MINIMUM impedance of the speaker ONCE during the power-up sequence of the amplifier.  After I find the MINIMUM, the amplifier will adjust itself to provide maximum power into that load.  This will NOT happen at any other time or during music playback.


I agree with Shonver. Two things that will change with gain adjustment is the bandwidth and the distortion figure of the amplifier. Maybe the bandwidth is high enough and the distortion low enough to not notice (hear) it, but vary it will.

Understood.  I was planning on designing around this - ensuring the amplifier is stable at minimum and maximum gain settings shouldn't be too hard.  Anything in-between will then be within the parameters so should be fine.

I did however have some other discussions with clever guys yesterday, and they have given me some other ideas around the implementation that may make things easier and cause less stability concern.  I will simulate some of these and see what I come up with.

Cheers,
Ian.

Offline ingenieus

Re: "Left brain going right brain" amplifier
« Reply #41 on: February 12, 2014, 12:54:13 PM »
Changing the resistance in the feedback network will also change the low-frequency roll-off point, since it is standard/best practice to have a great big capacitor in there. If you leave out the capacitor, you may run into DC offset problems.