Well now that I think about that maybe I was used to everything being NPN as far as transistors are concerned.... I don't really work with transistors much. But in the industrial automation stuff, all the sensors in north america are generally PNP. NPN is usually found in stuff from Asia and sometimes Europe.
Apparently I'm missing some of the resistor values I thought I grabbed from my friend (they might have fell out of the bag in my car when I was taking them home, but I might not find them), which might make things a bit harder.
So, I'd like to be able to make a SPICE model of the system, but I've never worked with SPICE before. I've got LTSpice, but it doesn't have models for the opamps... I can get models of them from TI, but they're Orcad PSpice models, and after a bit of research those are not easily converted. I did try downloading a less than legal copy but it seems to complain that it's not compatible with my version of windows.
I was thinking that since I should be using the TLC272 for the amp stage, and I've got transistors for a buffer, I should be able to use the TL074 as the vground, shouldn't I? All the TL074 would need to drive would be the transistor gates, which it should be capable of, and the rest of the current would be provided by the transistors, right?
Apparently I'm missing some of the resistor values I thought I grabbed from my friend (they might have fell out of the bag in my car when I was taking them home, but I might not find them), which might make things a bit harder.
So, I'd like to be able to make a SPICE model of the system, but I've never worked with SPICE before. I've got LTSpice, but it doesn't have models for the opamps... I can get models of them from TI, but they're Orcad PSpice models, and after a bit of research those are not easily converted. I did try downloading a less than legal copy but it seems to complain that it's not compatible with my version of windows.
I was thinking that since I should be using the TLC272 for the amp stage, and I've got transistors for a buffer, I should be able to use the TL074 as the vground, shouldn't I? All the TL074 would need to drive would be the transistor gates, which it should be capable of, and the rest of the current would be provided by the transistors, right?
I don't have money to buy a different supply, and bipolar supplies are not as common as single line supplies.
When I bought my supply I was planning on using 18 volts rail to rail and split it. I couldn't find a 18volt supply that was cheap and fit the specs I was looking for, so I bought the supply I have now, which is actually a bipolar supply (it supplies +15, -15, and +3.3).
I don't have any money for a different supply, and I'd rather not let this one be wasted, so I'm planning on using the +15 output and split it.
I'm not afraid of making things a bit more complex than they need to be since board space isn't a big deal. By far the most important part of this for me is the learning experience. Having a working headphone amp is more of a nice side effect.
The next amp I make will focus on more quality, but that probably won't be until I have a steady income, which is probably still a while away.
Earlier in the thread Godfrey mentioned that it had better current output than the TL074 (I have the DIP14, not DIP8 chip). There is nothing on the spec sheet that says what it's maximum current output is, but it does mention the maximum power dissipation vs air temperature at 30 volts rail to rail with no signal/load. Judging by that it can handle somewhere around 170mw dissipation at 25 celsius. I'm not sure how a signal affects that, but assuming signal is negligible it should handle about 170/15= 11.3ma... Meaning that since signal is not negligible it should be less than that. (assuming my math here is correct...)
So far the design I have on paper is this:
I'm using this as a reference for the vground and buffer stage, which means my design might be wrong, if my understanding of things is wrong.
I'm a bit confused about how exactly the opamp works here. It's got the feedback like a negative feedback amp, but instead of the inverting input being connected to ground like wikipedia shows, it's got the signal, which makes it look more like a differential amplifier... In any case, from what I can tell the opamp is the voltage amplifier (and if it follows the rules of negative feedback, it has a gain of 5.45), and the transistors as a voltage follower to buffer the low current capabilities of the opamp.
What I'm trying to do is create a virtual ground with an opamp and use the same voltage follower setup to increase the current carrying capacity of the virtual ground circuit.
Then what I plan is I'll create something similar to what godfrey posted on the first page with two TLC272s paralleled for the voltage amplifier stage. The 272 can handle about 30ma in either direction, so a total of 60ma per channel or 120ma total, which is just under what the transistors should be able to handle (150ma).
Is there anything terribly wrong with this design? Am I misunderstanding something here? This design is based on the assumption that I actually understand what's going on, so if I'm missing something important about how this works it's because it's been so long since I've worked with any AC stuff.
When I bought my supply I was planning on using 18 volts rail to rail and split it. I couldn't find a 18volt supply that was cheap and fit the specs I was looking for, so I bought the supply I have now, which is actually a bipolar supply (it supplies +15, -15, and +3.3).
I don't have any money for a different supply, and I'd rather not let this one be wasted, so I'm planning on using the +15 output and split it.
I'm not afraid of making things a bit more complex than they need to be since board space isn't a big deal. By far the most important part of this for me is the learning experience. Having a working headphone amp is more of a nice side effect.
The next amp I make will focus on more quality, but that probably won't be until I have a steady income, which is probably still a while away.
Earlier in the thread Godfrey mentioned that it had better current output than the TL074 (I have the DIP14, not DIP8 chip). There is nothing on the spec sheet that says what it's maximum current output is, but it does mention the maximum power dissipation vs air temperature at 30 volts rail to rail with no signal/load. Judging by that it can handle somewhere around 170mw dissipation at 25 celsius. I'm not sure how a signal affects that, but assuming signal is negligible it should handle about 170/15= 11.3ma... Meaning that since signal is not negligible it should be less than that. (assuming my math here is correct...)
So far the design I have on paper is this:
- Use a TL074 as a rail splitter.
- Use my transistors to buffer the new virtual ground to allow for larger current draw.
- Each channel has 2 parallel TLC272 for a total of 4. That adds up to 120 ma, just under what the buffer should be able to supply.
I'm using this as a reference for the vground and buffer stage, which means my design might be wrong, if my understanding of things is wrong.
I'm a bit confused about how exactly the opamp works here. It's got the feedback like a negative feedback amp, but instead of the inverting input being connected to ground like wikipedia shows, it's got the signal, which makes it look more like a differential amplifier... In any case, from what I can tell the opamp is the voltage amplifier (and if it follows the rules of negative feedback, it has a gain of 5.45), and the transistors as a voltage follower to buffer the low current capabilities of the opamp.
What I'm trying to do is create a virtual ground with an opamp and use the same voltage follower setup to increase the current carrying capacity of the virtual ground circuit.
Then what I plan is I'll create something similar to what godfrey posted on the first page with two TLC272s paralleled for the voltage amplifier stage. The 272 can handle about 30ma in either direction, so a total of 60ma per channel or 120ma total, which is just under what the transistors should be able to handle (150ma).
Is there anything terribly wrong with this design? Am I misunderstanding something here? This design is based on the assumption that I actually understand what's going on, so if I'm missing something important about how this works it's because it's been so long since I've worked with any AC stuff.
Either of your op amps will accept a ±15V supply, so IMO a virtual ground serves no purpose, other than to needlessly limit the output swing.
The circuit shown is a basic op amp configuration. R1 & R2 create the virtual ground reference applied to the noninverting input.
From your second-to-last paragraph, I get the impression that there's some confusion about base current and emitter current for the transistors. I admit I don't have the specs memorized nor have I checked their datasheets. 150mA sounds about right for the emitter current, so 60mA would be way plenty for the base current.
TBH I'm not comfortable with this statement. If you want "things a bit more complex than they need to be" you should probably get advice from a degreed engineer. They're trained in that sort of thing.
The circuit shown is a basic op amp configuration. R1 & R2 create the virtual ground reference applied to the noninverting input.
From your second-to-last paragraph, I get the impression that there's some confusion about base current and emitter current for the transistors. I admit I don't have the specs memorized nor have I checked their datasheets. 150mA sounds about right for the emitter current, so 60mA would be way plenty for the base current.
I wasn't using 60mA as the base current.
The TLC272 is rated at a maximum of + or - 30mA output. If I have 2 in parallel that should be + or - 60mA total.
Multiply that by 2 channels and the virtual ground circuit is going to need to sink or source up to 120mA at the most, though probably a fair bit less than that in operation.
Since the gain on those transistors is nominal at 100, maximum around 300, 360 or something, my TL074 should be supplying a very small base current in order to satisfy the needs of the transistors.
As for the comment about complexity, I just meant I didn't mind having to work out things like this. but your right, if I really wanted to make it more complex I'd get a hold of a proper engineer
I kinda figured that the noninverting input was a voltage divider, but my op-amp understanding is pretty basic, so looking at the inverting input has me a bit confused on exactly what is happening. Wikipedia's page shows the inverting input on a negative feedback amplifier as being tied to the ground, rather than a signal source. I understand that the relationship of the resistors creates a gain in a negative feedback amplifier circuit but I'm not sure how that interacts with the signal.
If I could see a simple signal representation of how that configuration worked and what the output was in relation to the input I'd probably understand pretty well. I'm one of those people where if I understand how something actually works, everything just makes sense, but if I don't, i'm completely lost because I don't rely on just memorizing random crap.
The TLC272 is rated at a maximum of + or - 30mA output. If I have 2 in parallel that should be + or - 60mA total.
Multiply that by 2 channels and the virtual ground circuit is going to need to sink or source up to 120mA at the most, though probably a fair bit less than that in operation.
Since the gain on those transistors is nominal at 100, maximum around 300, 360 or something, my TL074 should be supplying a very small base current in order to satisfy the needs of the transistors.
As for the comment about complexity, I just meant I didn't mind having to work out things like this. but your right, if I really wanted to make it more complex I'd get a hold of a proper engineer
I kinda figured that the noninverting input was a voltage divider, but my op-amp understanding is pretty basic, so looking at the inverting input has me a bit confused on exactly what is happening. Wikipedia's page shows the inverting input on a negative feedback amplifier as being tied to the ground, rather than a signal source. I understand that the relationship of the resistors creates a gain in a negative feedback amplifier circuit but I'm not sure how that interacts with the signal.
If I could see a simple signal representation of how that configuration worked and what the output was in relation to the input I'd probably understand pretty well. I'm one of those people where if I understand how something actually works, everything just makes sense, but if I don't, i'm completely lost because I don't rely on just memorizing random crap.
OK I was confusing the various current outputs. In the circuit shown, the virtual ground is the reference for the signal. Since the + input is at vground, the op amp will put the - input at the same potential, and that op amp DC output makes the cap before the speaker necessary. It sounds like the Wikipedia circuit you mention is a noninverting configuration.
Avoiding the complexity, the resistors only interact with the signal by allowing a certain portion of the 180° output (remember it's the inverting input) to cancel the signal input. Thereby controlling the gain.
The signal at the load will be a replica of the input, with some crossover distortion added. This will be a short "null" no output signal condition while neither transistor is conducting. Sorta like this exaggerated triangle wave.... ^-v-^-v. As explained on that page, the author relied on the slew rate of the op amp to minimize that null period.
Avoiding the complexity, the resistors only interact with the signal by allowing a certain portion of the 180° output (remember it's the inverting input) to cancel the signal input. Thereby controlling the gain.
The signal at the load will be a replica of the input, with some crossover distortion added. This will be a short "null" no output signal condition while neither transistor is conducting. Sorta like this exaggerated triangle wave.... ^-v-^-v. As explained on that page, the author relied on the slew rate of the op amp to minimize that null period.
Just came across this mini tutorial:
http://www.analog.com/static/imported-files/tutorials/MT-032.pdf
It does a very good job of explaining things.
The more I think about things the more I think my design will work, unless I've managed to overlook some issue. Sometime in the near future I'll throw together a full schematic showing my design. I might need some help figuring out resistor values and such.
Also, I have a dual ganged audio taper (I think) 10k pot for volume control.
http://www.analog.com/static/imported-files/tutorials/MT-032.pdf
It does a very good job of explaining things.
The more I think about things the more I think my design will work, unless I've managed to overlook some issue. Sometime in the near future I'll throw together a full schematic showing my design. I might need some help figuring out resistor values and such.
Also, I have a dual ganged audio taper (I think) 10k pot for volume control.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.