Voice Coil Rail
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Reposted because I was getting strange 403 errors..
I couldn't measure across the resistor because the test instruments all share the same ground (I think - will have to play further).
Still - this works, and the circuit is officially magic.
I put in some gain to show things up better:
IC2 is a differential amplifier with a gain of 100 which is making a voltage (red) proportional to the voltage across the 1 Ohm resistor. Its peak amplitude is about 5V so that's 50mV. I don't have time to play with it right now.
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
The Vertical scale got lost - it's 3V per large square
No it's not shareware, but I think there are many which are. I can share it with students. ...
With +/- 12v supplies it works with the variable res up to 50 Ohms, output (red) about 20Vpk-pk.
I couldn't measure across the resistor because the test instruments all share the same ground (I think - will have to play further).
Still - this works, and the circuit is officially magic.
I put in some gain to show things up better:
IC2 is a differential amplifier with a gain of 100 which is making a voltage (red) proportional to the voltage across the 1 Ohm resistor. Its peak amplitude is about 5V so that's 50mV. I don't have time to play with it right now.
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
The Vertical scale got lost - it's 3V per large square
No it's not shareware, but I think there are many which are. I can share it with students. ...
With +/- 12v supplies it works with the variable res up to 50 Ohms, output (red) about 20Vpk-pk.
Last edited by ChrisR on Mon May 01, 2017 6:35 pm, edited 1 time in total.
Chris R
-
- Posts: 3417
- Joined: Sat Nov 20, 2010 10:40 am
- Location: Santa Clara, CA, USA
- Contact:
OK, this is great. The voltage swing is there at cross over! Like I said before, this swing MIGHT not affect us (page 8 with capital letter).ChrisR wrote:
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
Since my scope is dead, I'd be really interested in the region (between time 300 and 400) where the green line drops, I'd like to see (if you could) what happens to the blue and pink at that point :-) This is just out of curiosity.ChrisR wrote:I do feel for Mike
Here. a sawtooth generator input, with R4 giving the output a lot to do, (these transistors would need heatsinks). It also magnifies the step in the green trace to make the function obvious.
The op amp is a crummy old 741, and the transistors not high gain.
No glitch.
The 1M resistors simply allow a non-coincident trace.
Peter,mjkzz wrote:OK, this is great. The voltage swing is there at cross over! Like I said before, this swing MIGHT not affect us (page 8 with capital letter).ChrisR wrote:
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
Sure there is some voltage swing and the voltage swing/(load current) is exactly 1K because of the resistor around the zero crossing, then the transistors begin to help out when the current demand gets around +-0.8ma. The circuit still remains closed loop with no dead zone and their is no cross over distortion...just look at the output waveform, the red trace....do you see any hint of cross over or any other type of distortion? It's perfectly linear as it should be!! Look at all the commotion with the green and blue traces, yet the output red trace shows no sign of any of this!!
The circuit just works as it should.
Best,
Mike
ChrisR wrote:Reposted because I was getting strange 403 errors..
I couldn't measure across the resistor because the test instruments all share the same ground (I think - will have to play further).
Still - this works, and the circuit is officially magic.
I put in some gain to show things up better:
IC2 is a differential amplifier with a gain of 100 which is making a voltage (red) proportional to the voltage across the 1 Ohm resistor. Its peak amplitude is about 5V so that's 50mV. I don't have time to play with it right now.
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
The Vertical scale got lost - it's 3V per large square
No it's not shareware, but I think there are many which are. I can share it with students. ...
With +/- 12v supplies it works with the variable res up to 50 Ohms, output (red) about 20Vpk-pk.
Chris,
I was wondering what all the commotion going on was with the blue and green traces, thought it was the op amp model or simulator acting up, but then I reread your post and realized you were varying VR1 around showing the effects of the load changing...clever!!
The output Red trace shows no hint of this varying as it should. Negative feedback current mode works well as you have shown.
So if you classify this circuit as "officially magic", so be it. But honestly a little clever physics are at play here, maybe some folks will use and benefit from this mode of operation...I know I sure have!!
Best,
Mike
Chris,
If your simulator has a Current Controlled Voltage Source (CCVS) device, then you can use this to measure the load current. Place the current sense side in series with the load and ground one side of the voltage output. Set the gain to whatever you like (10, 100, 1000), the default is 1 amp equals 1 volt. The load current will be the output voltage/gain.
Best,
Mike
If your simulator has a Current Controlled Voltage Source (CCVS) device, then you can use this to measure the load current. Place the current sense side in series with the load and ground one side of the voltage output. Set the gain to whatever you like (10, 100, 1000), the default is 1 amp equals 1 volt. The load current will be the output voltage/gain.
Best,
Mike
OK, Mike, I am just curious about what happens at and around crossing over part, ie, when time scale is bounded around there so we (or I ) can see some details, how circuit responds.mawyatt wrote:Peter,mjkzz wrote:OK, this is great. The voltage swing is there at cross over! Like I said before, this swing MIGHT not affect us (page 8 with capital letter).ChrisR wrote:
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
Sure there is some voltage swing and the voltage swing/(load current) is exactly 1K because of the resistor around the zero crossing, then the transistors begin to help out when the current demand gets around +-0.8ma. The circuit still remains closed loop with no dead zone and their is no cross over distortion...just look at the output waveform, the red trace....do you see any hint of cross over or any other type of distortion? It's perfectly linear as it should be!! Look at all the commotion with the green and blue traces, yet the output red trace shows no sign of any of this!!
The circuit just works as it should.
Best,
Mike
I understand PERFECTLY that red line stays the same even there are "commotion" when it is NOT at the crossing part, it is because the BJT is operating in active region ([edit]notice the green line stays above or below the blue line, that is because the BJT is turned on and is operating in active region till crossing point[/edit]), just like I said before, from +2ma to +1ma, you will not have any change, and as you said, it is because of feedback. However, around the point where the BJT must be turned on to supply additional current, IF it takes the circuit a little longer to reach its supposed value because of the relative large voltage swing, then, even though eventual current reaches right value, this additional time can be considered a distortion. The question is, does it matter to our application? As we can see from the graph, it is probably can be ignored. But, this probably is neat-picking, if we shorten the timing around those points, ie, magnifying the graph, do we see it? THAT IS MY CURIOSITY.
Do not take me wrong, as I said many times before, I never said your design does not work.
Last edited by mjkzz on Mon May 01, 2017 8:50 pm, edited 4 times in total.
Peter,
Maybe Chris can run a DC sweep around zero at +- 5ma, this should show any misbehavior. I'll bet you a beer and a bracket (I need a couple simple brackets to mount your KH precision focus rails to a Thor sliding clamp) it's perfectly linear in that region (within the confines of the op amp OLG), and everywhere else too
Best,
Mike
Maybe Chris can run a DC sweep around zero at +- 5ma, this should show any misbehavior. I'll bet you a beer and a bracket (I need a couple simple brackets to mount your KH precision focus rails to a Thor sliding clamp) it's perfectly linear in that region (within the confines of the op amp OLG), and everywhere else too
Best,
Mike
hahaha, I edited my response to you. Here, let me set one example: if time it takes is 0.Xms for regions where BJT is active to move from +n ma to +n-1 ma and IF it take 0.Yms where Y > X, to move from 1ma to 0, then it is a distortion but it could be so insignficant (as I said before on page 9) and it appear to be TRUE according to Chris's graphs for the time scale.mawyatt wrote:Peter,
Maybe Chris can run a DC sweep around zero at +- 5ma, this should show any misbehavior. I'll bet you a beer and a bracket (I need a couple simple brackets to mount your KH precision focus rails to a Thor sliding clamp) it's perfectly linear in that region (within the confines of the op amp OLG), and everywhere else too
Best,
Mike
Again, Mike, do not get me wrong, I am NOT trying to say your design is wrong. If my scope is still functioning, I could have solved this question my self by observing my own circut (which is uni-directional design) circuit regarding cross over part (with my design, it is from 0 to first bit)
PM me if you need those brackets.
-
- Posts: 3417
- Joined: Sat Nov 20, 2010 10:40 am
- Location: Santa Clara, CA, USA
- Contact:
Chris...after looking at your circuit and output, I realized that my original concern about crossover issues are not fully dispelled here. By having such a large input signal, and such a small load, any small irregularities in the crossover region could be masked. I also wonder if the 741 opamp in your simulation is idealized according to a behavioral model of gain, BW, etc or does it have subcircuits representing the actual schematic of a 741.ChrisR wrote:
The formerly straight lines are all wobbly because while it was running I was rapidly changing the value of the variable 10 ohm variable resistor from end to end.
If it is a simulated 741, then to verify that issues don't occur in the crossover region, would you mind simulating with a much smaller input waveform magnitude, and a larger load resistor? The 4" speaker I tested has 8 ohms resistance, and a sensitivity of about 1um/250uA. What I'd like to see is a sim with R1=8 ohms, and input amplitude of 0.1V peak. this will drive a +/-2mA signal through R1, with the Q1/Q2 pair turning on in the middle of the voltage swing.
If the 741 is behaviorally-modeled, then the simulation would of course yield a perfect result regardless of level or drive, so does not answer the crossover issue. I assume this is not the case but want to be sure.
Maybe a DC sweep between -0.75ma and +0.75m, that should be sufficient. :-)mawyatt wrote:Peter,
Maybe Chris can run a DC sweep around zero at +- 5ma, this should show any misbehavior. I'll bet you a beer and a bracket (I need a couple simple brackets to mount your KH precision focus rails to a Thor sliding clamp) it's perfectly linear in that region (within the confines of the op amp OLG), and everywhere else too
Best,
Mike
-
- Posts: 3417
- Joined: Sat Nov 20, 2010 10:40 am
- Location: Santa Clara, CA, USA
- Contact:
I did not see that earlier request.mjkzz wrote:Maybe a DC sweep between -0.75ma and +0.75m, that should be sufficient. :-)mawyatt wrote:Peter,
Maybe Chris can run a DC sweep around zero at +- 5ma, this should show any misbehavior. I'll bet you a beer and a bracket (I need a couple simple brackets to mount your KH precision focus rails to a Thor sliding clamp) it's perfectly linear in that region (within the confines of the op amp OLG), and everywhere else too
Best,
Mike
Just as important as the range, is ensuring the 741 is a real opamp and not a behavioral model.