Hi John, I'd like to use the OPA192 + BUF634A in a composite amplifier configured for precisely programmable current output (cf. figure 41 in the datasheet). For the control voltage, I'm using an 18-bit bipolar DAC with its output between -4.096 V and +4.096 V.
The output current flows through a DUT with a variable resistance (down to as little as 10 milli-Ohms). I'd like to have as much overhead as possible. In your experience, what would be a realistic upper voltage limit for the bipolar rails, given that: I'd like to use the SOIC-8 package of the BUF634A; and the programmed current output can vary between ±100 mA (sustained over very long periods); I'm effectively using a 40 Ohm resistor to set the current.
And how much would this change if I would use the upcoming SOIC package with the exposed thermal pad?
It looks like preproduction samples of the PowerPad BUF634A are now available (XBUF634AIDDAT). Is there an anticipated launch date for the new package?
I also just noticed the new OPA1637 in John's signature, which looks very interesting! John, do you happen to know when the spice model will be available?
I also just noticed the new OPA1637 in John's signature, which looks very interesting! John, do you happen to know when the spice model will be available?
OPA1637: SPICE model release, and full product release should be in the next few weeks.
BUF634A other packages: I believe these are also planned to release in 2nd quarter this year.
Well it should, considering the OPA1656 consumes 4.5x more power supply current than the OPA1637. OPA1656 is also an op amp (1 output) where the OPA1637 is a fully-differential amplifier (2 outputs). That means the 1637 also has to burn power supply current on an additional output stage and a common-mode feedback loop.
I'm not really sure what this comparison proves? They are two VERY different products, with very different intended applications.
John: The current TINA-TI model for the BUF634A appears to be wide-band only. It would be nice if the BW pin was supported - or at least if one could select between the two extremes of the BW options.
Well... It doesn't have a UX4 socket for a 300B on top (with options for 2A3 and 45, of course), so there's Audio Fail #1...
I don't think the electrons will care much whether the IC is labeled "Audio" or not.
Tom
Well... It doesn't have a UX4 socket for a 300B on top (with options for 2A3 and 45, of course), so there's Audio Fail #1...
I don't think the electrons will care much whether the IC is labeled "Audio" or not.
Tom
John: The current TINA-TI model for the BUF634A appears to be wide-band only. It would be nice if the BW pin was supported - or at least if one could select between the two extremes of the BW options.
Well... It doesn't have a UX4 socket for a 300B on top (with options for 2A3 and 45, of course), so there's Audio Fail #1...
I don't think the electrons will care much whether the IC is labeled "Audio" or not.
Tom
Yep, the team has been working on the variable bandwidth simulation model for quite a while. I can check-in to see where they are at with it.
If I make an audio-specific version, can I charge more for it?
If I make an audio-specific version, can I charge more for it?
Mbe if you autograph each one!
It's very encouraging to see the work TI is devoting to high end audio.
Yep. Anybody can make NAND gates. Precision circuits is a whole different matter. Keep at it, John.
Tom
In the datasheet of the BUF634A, FIg.11~14 show distortion figures of the BUF634A (on its own I presume) as a function of frequency,
with 10Vp-p and 1k / 50R load.
Would you also be kind enough to show :
1) H2/H3 as a function of output amplitude (say from 0.5Vrms to 5Vrms) at 1kHz & 10kHz into 1kohm respectively;
2) H2/H3 as a function of load impedance (say from 32R to 1k) at 1Vrms, 1kHz & 10kHz respectively ?
And one more question if I may :
Fig.15 shows the relationship between bandwidth and the bandwidth adjustment resistor.
3) Could you also kindly show the quiescent current as a function of the same resistor values ?
3a) And how much of that quiescent current belongs to the output devices ?
(i.e. how much Class-A output current is available at what resistor value)
Many thanks in advance,
Patrick
with 10Vp-p and 1k / 50R load.
Would you also be kind enough to show :
1) H2/H3 as a function of output amplitude (say from 0.5Vrms to 5Vrms) at 1kHz & 10kHz into 1kohm respectively;
2) H2/H3 as a function of load impedance (say from 32R to 1k) at 1Vrms, 1kHz & 10kHz respectively ?
And one more question if I may :
Fig.15 shows the relationship between bandwidth and the bandwidth adjustment resistor.
3) Could you also kindly show the quiescent current as a function of the same resistor values ?
3a) And how much of that quiescent current belongs to the output devices ?
(i.e. how much Class-A output current is available at what resistor value)
Many thanks in advance,
Patrick
The old BUF634 had a 10 Ohm internal output resistor making it easy to build a composite amplifier with several BUF634 in parallel without external output resistors on each BUF634.
But what about the new BUF634A, does it still have an internal 10 Ohm output resistor for easy paralleling? It is being marketed as a drop in replacement for BUF634 so it should be the same as the old BUF634 in this regard, but nothing is mentioned in the datasheet.
But what about the new BUF634A, does it still have an internal 10 Ohm output resistor for easy paralleling? It is being marketed as a drop in replacement for BUF634 so it should be the same as the old BUF634 in this regard, but nothing is mentioned in the datasheet.
Last edited:
That 10ohm resistor was intended to provide some protection (against HF oscillations) in capacitive loads. It would not allow to connect BUF634's in parallel without balancing resistor, unless you are using the BUF634 in open loop (a stupid thing to do, IMO). Otherwise, the 10ohm resistor is in the feedback loop and at low frequency it's effect on the output impedance is practically zero.
So you need external resistors for paralleling both BUF634 and BUF634A, anyway.
So you need external resistors for paralleling both BUF634 and BUF634A, anyway.
You are wrong about needing external output resistors on the old BUF634.
Application note :
https://www.ti.com/lit/pdf/sboa065
Figure 2.
Quote from application note :
The output resistance of the
BUF634 is about 10Ω. Therefore, series output resistors for
decoupling the individual buffers are no longer necessary.
Does the new BUF634A have the same 10 Ohm output resistance?
Application note :
https://www.ti.com/lit/pdf/sboa065
Figure 2.
Quote from application note :
The output resistance of the
BUF634 is about 10Ω. Therefore, series output resistors for
decoupling the individual buffers are no longer necessary.
Does the new BUF634A have the same 10 Ohm output resistance?
Last edited:
You need to read further.
What they say is that a resistor is not required for stability purposes, which is correct.
Then, they calculate the sink/source excess current, depending on the op amp offset, which is again correct, but please note the setup has only x2 noninverting gain. The offset in the calculation got multiplied by the closed loop gain, and so does the excess current. What they do not say is the effect of no series resistor (outside the feedback loop) on the distortion peformance. I'll let you figure it out by yourself, and I guess you won't be happy.
About the BUF634A, I have no idea.
What they say is that a resistor is not required for stability purposes, which is correct.
Then, they calculate the sink/source excess current, depending on the op amp offset, which is again correct, but please note the setup has only x2 noninverting gain. The offset in the calculation got multiplied by the closed loop gain, and so does the excess current. What they do not say is the effect of no series resistor (outside the feedback loop) on the distortion peformance. I'll let you figure it out by yourself, and I guess you won't be happy.
About the BUF634A, I have no idea.
Should have read the BUF634A datasheet more carefully.
BUF634A in WB Mode:
Zo, output impedance, DC,10 mA, 5 Ω.
So the outpt resistance is mentioned and is half of the old BUF634 so it should be possible to parallel more of them within the feedback loop of an opamp without external output resistors on each buffer output.
BUF634A in WB Mode:
Zo, output impedance, DC,10 mA, 5 Ω.
So the outpt resistance is mentioned and is half of the old BUF634 so it should be possible to parallel more of them within the feedback loop of an opamp without external output resistors on each buffer output.
