G
Guest
Guest
I plan to make a simple output buffer for a function generator and would like to use an opamp designed to drive low impedence at 10V or so. What are some good candidates?
Help Support GroupDIY Audio Forum:
5534
I use a 5534 and a BD139/140 pair for the output of my function generator.
But often you want e.g. 600 Ohms of output impedance from your generator anyway, and in that case a 5534 alone will do fine..
But often you want e.g. 600 Ohms of output impedance from your generator anyway, and in that case a 5534 alone will do fine..
bcarso
Well-known member
What is the nature of your function generator to begin with?
Most of the off-the-shelf Far East-manufactured generators these days already have a pretty hefty discrete buffer amp and a 50 ohm output Z.
Most of the off-the-shelf Far East-manufactured generators these days already have a pretty hefty discrete buffer amp and a 50 ohm output Z.
DrFrankencopter
Well-known member
Buf634 will drive heavy loads no problem. I'm using them at work to drive hydraulic servo valves. They put out 100ma without even getting hot.
Cheers,
Kris
Cheers,
Kris
Svart
Well-known member
I am still unclear about what defines "low" and "high" impedence.. I have always thought in terms of current sourcing and current sinking.
I have never seen a straight answer to certain questions like:
when we say "low" impedence I assume this means a low number value which is easier to drive, but when we talk about transformers, we usually talk about the AC impedence through the coil itself. So when we talk about impedence in solid state circuits, that value is generally measured to ground since we cannot measure "through" it. This would make a low number value the opposite, a higher value for our parts to fight against. Of course if you look at it another way, low impedence means that we aren't wasting our current as heat or flux in a trafo circuit and therefor have to sink more current on the return, but then again in solid state the receiving circuit is the part that demands the levels of drive we feed it and it either gets enough, starves or melts the driver by demanding too much.
so which is it? Are we really supposed to measure all nodes to ground and reference them that way? or is it truly different for the type of circuit that you are testing?
:guinness:
I have never seen a straight answer to certain questions like:
when we say "low" impedence I assume this means a low number value which is easier to drive, but when we talk about transformers, we usually talk about the AC impedence through the coil itself. So when we talk about impedence in solid state circuits, that value is generally measured to ground since we cannot measure "through" it. This would make a low number value the opposite, a higher value for our parts to fight against. Of course if you look at it another way, low impedence means that we aren't wasting our current as heat or flux in a trafo circuit and therefor have to sink more current on the return, but then again in solid state the receiving circuit is the part that demands the levels of drive we feed it and it either gets enough, starves or melts the driver by demanding too much.
so which is it? Are we really supposed to measure all nodes to ground and reference them that way? or is it truly different for the type of circuit that you are testing?
:guinness:
[quote author="Svart"]when we say "low" impedence I assume this means a low number value which is easier to drive,[/quote]
No, a "low" number is hard to drive. If you connect a low impedance (like a 2 ohm speaker) to the output of a 5534, you would practically short-circuit it. A high impedance (like a 1M tube amp input), would almost not load the 5534 at all, so you could get a full voltage swing from the 5534.
Best regards,
Mikkel C. Simonsen
No, a "low" number is hard to drive. If you connect a low impedance (like a 2 ohm speaker) to the output of a 5534, you would practically short-circuit it. A high impedance (like a 1M tube amp input), would almost not load the 5534 at all, so you could get a full voltage swing from the 5534.
Best regards,
Mikkel C. Simonsen
Svart
Well-known member
I see what you are saying. So it is a matter of current sourcing and the ability of the driver to handle the current demand of the driven part?
So when we shunt some resistence to ground, lets say 10k in an IC based circuit, are we in fact just loading(pulling down) the output of the driver and the input of the receiver side down to either some level or to counter act something?
:guinness:
Sorry, when i think of Impedence or Resistance, i think of a measurement. I think of current sourcing/sinking and the associated levels when I am dealing with things of this nature.
So when we shunt some resistence to ground, lets say 10k in an IC based circuit, are we in fact just loading(pulling down) the output of the driver and the input of the receiver side down to either some level or to counter act something?
:guinness:
Sorry, when i think of Impedence or Resistance, i think of a measurement. I think of current sourcing/sinking and the associated levels when I am dealing with things of this nature.
G
Guest
Guest
I plan to make a couple of generators (inspired by another thread...and my old one died). One 8038 based and maybe one state variable based. I want to use them for general audio testing for both tube and transistor circuits. I think a descent buffer is necessary here. Maybe something like an opamp with transistor drive like Jakob mentioned would fit the bill.
bcarso
Well-known member
[quote author="thomasholley"]I plan to make a couple of generators (inspired by another thread...and my old one died). One 8038 based and maybe one state variable based. I want to use them for general audio testing for both tube and transistor circuits. I think a descent buffer is necessary here. Maybe something like an opamp with transistor drive like Jakob mentioned would fit the bill.[/quote]
Ah yes then you definitely need the buffer, unless the loading on the state variable one is light. I forget what the 8038 is good for but it's not very much.
Ah yes then you definitely need the buffer, unless the loading on the state variable one is light. I forget what the 8038 is good for but it's not very much.
SonsOfThunder
Well-known member
[quote author="bcarso"]I forget what the 8038 is good for but it's not very much.[/quote]
It's good for about 2% THD... all the time...
I would not recommend it for anything except to verify that a ckt will pass a signal.
HTH!
Charlie
It's good for about 2% THD... all the time...
I would not recommend it for anything except to verify that a ckt will pass a signal.
HTH!
Charlie
pstamler
Well-known member
[quote author="SonsOfThunder"][quote author="bcarso"]I forget what the 8038 is good for but it's not very much.[/quote]
It's good for about 2% THD... all the time...
I would not recommend it for anything except to verify that a ckt will pass a signal.[/quote]
It's good enough to measure frequency responses if you use an averaging meter rather than peak-reading. And it makes a fairly nice triangle wave.
Peace,
Paul
It's good for about 2% THD... all the time...
I would not recommend it for anything except to verify that a ckt will pass a signal.[/quote]
It's good enough to measure frequency responses if you use an averaging meter rather than peak-reading. And it makes a fairly nice triangle wave.
Peace,
Paul
If you has in mind a real function generator (not only sine, but pulse, triangle and so on output signals), it is mistake - to use the common type operational amp as 5534.
The choice of the output IC depends on both generators' highest frequency and of output signal's shape.
The 5534 may be used only for the sine (audio) generator, but even in this case this OpAmp is not best - its Iout is only 20mA.
For this case 1/2 of NJM4556A or 1/2 of M5216 (Iout=70mA) is more preferable.
For real function generator output stage you must calculate the bandwidth
depending on acceptable distortion for given signal's shape and highest frequency.
For this case the suitable OpAmps are: THS4631 (slew rate=900V/µS; BW=210 MHz; Iout=95mA), 1/2 of LT1364 (slew rate=1000V/µS; BW=70 MHz; Iout=60mA) or 1/2 of AD8397 (slew rate=53V/µS; BW=69 MHz; Iout=70mA).
Regards,
The choice of the output IC depends on both generators' highest frequency and of output signal's shape.
The 5534 may be used only for the sine (audio) generator, but even in this case this OpAmp is not best - its Iout is only 20mA.
For this case 1/2 of NJM4556A or 1/2 of M5216 (Iout=70mA) is more preferable.
For real function generator output stage you must calculate the bandwidth
depending on acceptable distortion for given signal's shape and highest frequency.
For this case the suitable OpAmps are: THS4631 (slew rate=900V/µS; BW=210 MHz; Iout=95mA), 1/2 of LT1364 (slew rate=1000V/µS; BW=70 MHz; Iout=60mA) or 1/2 of AD8397 (slew rate=53V/µS; BW=69 MHz; Iout=70mA).
Regards,
G
Guest
Guest
I need a function generator that does square, triangle and sine. I think (hope) the 8038 will do a descent square and triangle and a somewhat functional sine. A state variable should do a relatively good sine I think, but no square or triangle. Is there a better solution?
I think a buffer is necessary to keep the generator independent from the load. I appreciate all the help here. Thanks.
I think a buffer is necessary to keep the generator independent from the load. I appreciate all the help here. Thanks.
bcarso
Well-known member
You can make a sine wave gen with a bandpass filter and a comparator (or simply an open-loop op amp), which supplies the square wave. Then the tricky part is integrating the square wave to get the triangle, because a simple open-loop system will have the amplitude proportional to period.
Actually, there is a tricky filter design that gets you a pretty good triangle from a square wave, but it has to be adjusted for each frequency.
Many are the ways.
Actually, there is a tricky filter design that gets you a pretty good triangle from a square wave, but it has to be adjusted for each frequency.
Many are the ways.
