Well, I hadn't thought of the problem of bandwidth. A DN2540 at low (for it) voltages has too high input capacitance too keep it from having a drooping high frequency response, even with all that NFB around it. LTspice seems to think it will be -3dB at 20kHz if used like an 'anode follower'. So there goes that idea.
Now... What if I try that with an enhancement mode IRF510 or similar?
--
Nope. I don't think so. I can't get that to work with a single +48V supply.
OK, so... How does one make a little amplifier from transistors that:
1) Has high input impedance (>100k) and low output impedance (<1k)?
2) Has 6dB gain?
3) Has frequency response +0dB/-3dB from 2Hz to 200kHz into a 10k ohm load?
4) Has low harmonic distortion (<0.05% at 1Vrms out into 10k ohm load, preferably all 2nd harmonic)?
5) Sinks enough current to not slew limit at high frequencies into 2nF or worse cable capacitance?
Is an opamp the only sensible way to accomplish this set of goals?
--
Last edited:
In a word, 'No'. But it is one of the most *design-efficient*, in that minimal design time / effort is required, and despite that, it will most likely work as intended.
There are some reasonably simple single-stage designs -- if you can tolerate signal inversion. But they probably wouldn't use your '510's without a bit of head scratching.
So, have you decided that the idea of 6dB gain is too appealing to pass up?
THD of 0,05% @ 1 Vrms with a 10k ohm load is a modest and achievable goal. There is no solution I know of that will have even mostly, let alone all 2nd harmonic. But you can get the THD low enough that you won't mind the small amounts of the other higher-numbered ones.
What on earth kind / length of cable are you driving that has 2000pF of capacitance?! The slew rate will take care of itself if you can get to 200kHz only 3 dB down.
But plan to design for similar sinking and sourcing current capability for best performance.If you REALLY need to drive 2nF cables, most op-amps won't like it without help.
You might take a look in the Headphones section. There are some excellent, and not-too-complicated, amplifiers capable of meeting all 5 of your specs.
I'd still like to talk you out of the 48V supply. But even if you lower it, a TO-92 package cannot dissipate the necessary power to serve as an output device. This includes the CCS: Properly set up, it will have to get rid of about the same amount of heat.
Cheers
Last edited:
Well, I doubt I actually have 2nF of cable capacitance to deal with, but perhaps 1nF. I have a pair of active monitors fed from a so-called passive preamp (volume control/input switchbox). The cables from the passive control box are about 10 feet long. At 100pF/ft, that would be 1000pF, correct? Granted, 100pF/foot is not the lowest capacitance cable available, but it's not unusual for home audio shielded cables to measure around that figure.
I picked that 2000pF number as an achievable goal plus some headroom.
As I understand it, bandwidth per se is not a guarantee of freedom from slew limiting into a particular load capacitance. That's more a matter of the driving circuit being able to dump enough current into the load to charge the load capacitance at high frequencies.
From what I understand. I = C * 2 * pi * F * Vpeak
Therefore, for the circuit to drive 3V peak up to 40kHz into a 2,000pF load without slew limiting, the amplifier will need to be able to sink 1.5mA current into that load. I figure I'd go for at least double that, or 3mA. So yes, it does appear that an OPA2134 with +/-15V rails would be able to handle this with no problem.
I suppose there many JFETs that can handle this as well. I have some J309 and J310.
There are so many possibilities. I guess I just need to pick one, any one.
I can be talked out of that +48VDC supply. A 120V to 12V-0-12V 1A transformer is cheap and easy to get.
--
100p/ft is incredibly large. Fat core with nano-thin insulation.
30pFd/foot is a good guess for most construction.
100pF/meter is also good, in fact essentially the same.
So 10 foot cables are 300pFd.
OK. Then 10V peak in 300pFd at 20kHz is 0.4mA peak.
I'd want 2 or 3mA in the active device.
This is not a telephone system. There we may provision 40mA to drive a couple miles of cable to 4kHz, or to drive a 1000 foot line to 15kHz.
Thanks PRR.
I googled 'typical capacitance of RCA cable' and got all sorts of crazy stuff from a variety of forums, etc. I suppose people are out there measuring the capacitance of their cables using cheap LCR meters and posting their findings on the innerwebs. Not accurate?
Let's split the difference and say 50pF/ft. So 10 ft = 500pF. That would mean only 1mA could easily drive 10 feet of cable. What is the input capacitance likely to be of a powered speaker like a JBL LSR305? Perhaps 100pF? More? I don't know.
So let's say the total load might be 1000pF. Just to be super-safe.
Also, let's double our uppermost frequency just to be super-safe, so 40kHz instead of 20kHz.
So the target is the load C is 1000pF, the highest frequency to be swung is 40kHz, and the highest voltage is (an admittedly whopping) 3V peak. That comes to only 0.75mA.
So I guess 4mA is *plenty*.
The OPA2134 wins. (As expected.)
I suppose a vacuum tube like a 6DJ8 could do well enough, and it would glow.
An IRF510 with 10mA Ids and a +24VDC supply doesn't seem too crazy either. That would be intermediate between the opamp and the triode, in that it's more silly and wasteful than the opamp, but not as silly and wasteful as the triode.
--
I googled 'typical capacitance of RCA cable' and got all sorts of crazy stuff from a variety of forums, etc. I suppose people are out there measuring the capacitance of their cables using cheap LCR meters and posting their findings on the innerwebs. Not accurate?
Let's split the difference and say 50pF/ft. So 10 ft = 500pF. That would mean only 1mA could easily drive 10 feet of cable. What is the input capacitance likely to be of a powered speaker like a JBL LSR305? Perhaps 100pF? More? I don't know.
So let's say the total load might be 1000pF. Just to be super-safe.
Also, let's double our uppermost frequency just to be super-safe, so 40kHz instead of 20kHz.
So the target is the load C is 1000pF, the highest frequency to be swung is 40kHz, and the highest voltage is (an admittedly whopping) 3V peak. That comes to only 0.75mA.
So I guess 4mA is *plenty*.
The OPA2134 wins. (As expected.)
I suppose a vacuum tube like a 6DJ8 could do well enough, and it would glow.
An IRF510 with 10mA Ids and a +24VDC supply doesn't seem too crazy either. That would be intermediate between the opamp and the triode, in that it's more silly and wasteful than the opamp, but not as silly and wasteful as the triode.
--
Last edited:
Cheap PVC-insulated cables can have a very high capacitance, far above the 100 pF/m for polyethylene-insulated 50 ohm coaxial cables:
Tasker C118 stereo unbalanced cable: 280 pF/m (85.344 pF/ft), https://www.tasker.it/db_files/products/e0c92a1af8.pdf
Tasker C113 mono unbalanced cable is even 460 pF/m (140.208 pF/ft): https://www.tasker.it/db_files/products/747740f1da.pdf
Tasker C118 stereo unbalanced cable: 280 pF/m (85.344 pF/ft), https://www.tasker.it/db_files/products/e0c92a1af8.pdf
Tasker C113 mono unbalanced cable is even 460 pF/m (140.208 pF/ft): https://www.tasker.it/db_files/products/747740f1da.pdf
Also you can try to make symetric follower with irf510 and irf9510 ,using tl431 to bias , it will consume less current , source and sink resistance much lower than 560 ohms initial idea. But input capacitance will double , that's a drawback .
But opamp still wins in terms of distortions and easier to build I think.
But opamp still wins in terms of distortions and easier to build I think.
So the very worst case could be about 150pF/foot (rounding up, for simplicity).
1500pF for 10 feet (3 meters).
Add whatever input capacitance (and stray capacitances) there might be at the amplifier input to get an extreme-worst-case of 2000pF. That is what I started with. I'll stick to that as one of the design criteria for a line level buffer/cable driver.
To swing 2.83V peak at 40kHz into a 2nF load will require the driving circuit to sink 1.42mA.
4mA would be fine, and that's apparently what you'll get from an OPA2134 inverting amplifier set to 6dB gain run off a +/-15VDC supply.
10mA would be way overkill, but not off-the-charts-insane-overkill.
Yes, I admit, going for 35mA Ids is silly. Wasteful, not necessary.
--
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.