Unity Gain Buffer as a Current Source

Question

enter image description here enter image description hereI keep having issues with creating a unity gain buffer amplifier to isolate a voltage source. My voltage source is very limited in current ~1-2 mA and I need that signal to drive a transformer at 10 VAC. I thought it would be as simple as using an op-amp, but I am finding out that op-amps do not like to source more than a few milliamps. My transformer draws about 20 mA at 10 V.

I thought I could use a BJT or two to help source this current, but I'm finding out they do like my 15V rails because all they do is get hot and output nothing. I have them in a push pull configuration. They are darlingtons rated for much more current than what I'm drawing. I am going to be using an NI cRIO module to create a 10 VAC sinewave and all I need is to find a way to buffer the current output, so that I can drive this unloaded transformer.

Someone mentioned the LT1210 but I seem to be missing something since I lose half of my input voltage. ignore the .01 Resistors. I have them there to measure current through that trace.

• jonk: Does that relay truly have a 10 VAC RMS coil? And you want this activated by a low current output compliance DC voltage of some (unstated) kind? Also, let me know if you have the kind of cash to buy NI modules just to generate the 10 VAC you need for a relay coil. I'm kind of, well, shocked. Almost any other solution would be cheaper. (Unless you have them freely laying about.) Maybe you should provide us with some more detail about the surrounding context regarding what you are really trying to achieve.
• badgerengr: There is no relay... I'm driving a set of transformers to simulate 3 phase voltage. They will not be loaded. The NI module only has a max of 10Vac output which is why I need the transformers. To drive an unloaded coil of the transformers I need 20mA of sourced ac current from some type of amplifier circuit that I don't know where to begin designing since op amps typically can't source that much current.
• jonk: Why not get a mains transformer with a 10 VAC secondary and use the DC signal, with appropriate thresholding and a simple BJT switch, to activate a DC relay that then passes along the secondary AC to you load?
• badgerengr: Because I need to create 3 phase high voltage (>100VAC) out of basically nothing. I have a signal generator that can create a 3 phase low voltage (<10VAC). I need to get this low voltage to become high voltage somehow without any current from the source. So I was thinking I pull this current from an op amp so that I can drive my transformer. The setup works with a new op amp but since its pulling 20mA it quickly heats up my op amps and burns them out. The ones that can source more than 20mA on digikey are over 150ドル and there must be a cheaper option.
• Elliot Alderson: What makes you say that BJTs can make sine waves? What do you think is inside your op amp?

Answer 1

You have to use a power amplifier or maybe a line driver. These type of opamps can supply more current than the standard opamp. (You can find some e.g. at TI website under Amplifiers/Line drivers or Amplifiers/Power Opamps).

If you are sure that you don't need more than 20mA, you might also find some standard opamp. Take care, that the higher the output current, the lower the output swing - check that in the datasheet.

Extra tip, because I failed to do this a few times: You also need to take care of the self-heating of the amplifiers. It is usually the sum of the intrinsic heating (quiescent current times supply voltage) and the load dependent heating (output current times the voltage drop on the opamp). Multiply this power with the thermal resistance and check whether you are in the operating junction temperature limits.

Edit: as pointed out by electrogas also check out the TI AN-31 OpAmp Circuit Collection. You find the current sink on page 21. enter image description here

• badgerengr: You're confident there are op amps that can pull that much power from the rail to my output load? As much as I can tell the op amps on digikey cannot source that much current. The ones that can source that much current can't handle the 10VAC out I'm trying to get. The ones on there that can are are upwards of 200ドル which I cannot afford. I know for a fact that the ones I have accidentally burnt up can output 20mA. But what I cannot tell is how much source current they can pull.
• electrogas: There are ways to add a couple bjt's to the output of an op-amp (and put them inside the feedback loop) to boost the current. Look for circuits for audio amplifiers. Otherwise, an opamp that I like to use for high current (about an amp or so) is the LT1210. There is also the LT1010 unity gain buffer. Also there are audio amplifiers that can pump out the current, LM3876 as I recall, also newer ones that are class D and very efficient.
• G. B.: Yes, I'm using the ALM2402 with 500mA, 12V. I have also used the OPA548 with 30V and 1.5A output current (although with fan cooling and shutdown pin) a few years ago.
• badgerengr: Those are quite expensive compared to audio amplifier ICs which are capable however they are built for a different purpose. Would you have any worries if I used one of those?
• G. B.: Not at all, just check the characteristics. Max. output load, output current, output swing, etc. Maybe you are better off with a differential output amp to drive your coil.

Answer 2

My power supply is some weird Chinese thing that is switching voltage at 2MHz. This switching voltage is near a full volt +&-. This is the root cause of all my worries. The LM1875 can handle this switching much better than my other op amps I was using. This still causes problems because the sensor I am trying to test is a CT sense which uses the voltage drop across a .1 Ohm resistor. At this switching voltage it would immediately ruin my sensing controller. Luckily I never actually tried to run it, I've only been using test resistors and a multi meter. My original problem with the transformer issue was solved fairly quickly once I found a suitable heat sink and changed my design from a buffer to a simple unity gain inverting amp. I believe it was this switching frequency causing me issues with that first design.

Answer 3

So you are driving a transformer from an opamp. The transformer only requires 20mA which the opamp should be able to provide, yet it overheats and fries.

First suspect is DC offset. If the transformer requires 20mA AC current but the opamp has a DC offset on the output, and the transformer winding has a low DC resistance, then a lot of current may flow. So you should measure DC output voltage.

The fix is to add a coupling capacitor in series with the transformer. You can use two electrolytic caps in series back to back to make a non-polar electrolytic cap. Capacitor value should be large enough to pass your signal frequency.

You can also add a resistor in series with the opamp output, maybe 50 ohms. At 20mA it won't drop too much voltage and it will protect the opamp.

If this is not a DC offset problem, then you need a more powerful opamp. Since you need 3 channels, the cheapest and simplest high current opamp option will be LM1875, but you can also use a 5.1 home theater amplifier.

You will probably need to add a few ohms resistor or a cap in series with the transformer if the primary coil resistance is too low. This is again to avoid high DC current due to DC offset, which will make your amplifier hot even without signal and may trip its protection circuit.

Now, that said, it's quite weird that your opamp fries with only 20mA current. Maybe it's oscillating (check with a scope). In this case, add supply decoupling caps and a resistor on the output.

You can also parallel several opamps to increase current. Add a 50R resistor on the output of each to make sure their different DC offsets don't cause extra DC current.

• badgerengr: I never thought to check if there was any dc current being drawn into the transformer. I just assumed there wouldn't be since the high side is theoretically open 800kohms and I was driving the buffer with a sine wave function generator. But a simple high pass filter should do the trick then?
• bobflux: Is there a DC current? If so then just add a cap in series, that should do the trick