Function generator with a changeable output impedance

[SergeantFreeman]

Hello there,
I have the following problem: I want to use a function generator to power a resonant circuit. It´s just a simple LRC series resonant circuit, nothing special.
The problem is that the output impedance of the function generator kills my quality factor (-->picture). I aim to get the best possible quality factor out of the circuit.
If I simulate the circuit without the 50 Ohms output impedance I get a nice Q-Factor. But in my measurements the 50 Ohms of additional series resistance destroys everything. The SimSmith Simulation is close to the measurement because I can tune all of my manually measured Q-Factors of the components.
Is there a way to power the circuit without destroying the Q-Factor?





Thanks!

[TimFox]

In olden days, when driving a filter that needed a high source impedance, we used a pentode with its high plate resistance, absorbing the device capacitances into the filter.
Today, you can use the similarly high output impedance of a common-emitter BJT, or a common-source FET.
For testing, one just uses an appropriate high or medium-high resistance resistor after the 50 ohm generator.
Similarly, if you need a low source impedance, you should drive it from a "follower", but it may be difficult to obtain much less than 50 ohms at very high frequencies.
To test, one just uses a voltage divider whose total resistance is 50 ohms to load the generator.

L-C filter design must include the source and load impedances.
In general, a parallel L-C circuit works with relatively high impedances, and a series L-C circuit works with relatively low impedances.
(High and low means with respect to the reactances of the L's and C's at the frequency of interest.)

[macboy]

A function generator output is specifically destined to drive either a 50 ohm terminated input or a high impedance. If you want low impedance to drive a load like your filter, you need to connect to an amplifier.

It looks like you are using frequencies up to about 200 kHz. I'd try a simple opamp voltage follower. Most opamps should have low enough closed loop output impedance at those frequencies but it doesn't hurt to check the datasheet of the device you are intending to use.

[mawyatt]

Be careful about typical GP op-amp closed loop output impedance at 200KHz. Most GP op-amps only have 20~30dB available small signal gain at 200KHz, so not much excess gain to drive down the open loop output impedance, and much less available gain for larger signal swings. Also, the closed loop output impedance will "look" inductive with rising impedance vs frequency, so resonances are likely with capacitive loads.

Probably best to use a higher performance op-amp like those utilized in AWG outputs for these reasons, or a buffer type like the BUF634 with or without an input amplifier.

Best,

[bdunham7]

You can buy AWG amplifiers with much lower output impedances.

https://www.ebay.com/itm/114230957723

[schmitt trigger]

I had an almost identical issue, which I resolved with a LT1210 in a TO220 package
It has the bandwidth and current capacity you require.

[mawyatt]

You can buy AWG amplifiers with much lower output impedances.

https://www.ebay.com/itm/114230957723

Yeah, these Juntek are not bad and reasonable.

https://www.eevblog.com/forum/testgear/awg-buffer-amp/msg3368650/#msg3368650

Or DIY.

https://www.eevblog.com/forum/testgear/hv-buffer-amp-for-awg/msg3623953/#msg3623953

https://www.eevblog.com/forum/testgear/cheap-diy-high-powervoltage-awg-buffer-amp/msg3449784/#msg3449784

Even designed some (PCBs also) that use the LM3866, but haven't had time to build up yet.

Best,

[radiolistener]

if your circuit needs impedance which is about 4-9 times higher or lower than your source, you can use transformer to transform impedance.

For example, if you're needs transformer with 1:4 impedance ratio, it means 1:sqrt(4) = 1:2 turns/voltage ratio. Just make sure that primary coil has enough inductance (enough windings) for your frequency and select proper transformer core material and size with taking into account your working frequency and power...

Transformer gives you noise free impedance transformation and don't require power supply, it helps you to keep noise figure as low as possible

For impedance ratio higher than 1:8 transformer working bandwidth will be too narrow due to physical limitations, but if you're using it at fixed frequency you can make transformer with higher impedance ratio.

For example if your signal generator has 50 Ω output and you're using transformer with 9:1 impedance ratio (3:1 turns ratio), then your output after transformer will be 50 / 9 = 5.6 Ω

The same, if you're needs high impedance, just reverse transformer to get 1:9 impedance ratio and your output after transformer will be 50 * 9 = 450 Ω

[SergeantFreeman]

Thanks for the great ideas!

Esp. the transformer sounds really interesting. Do you know good literature about that?

[magic]

Be careful about typical GP op-amp closed loop output impedance at 200KHz. Most GP op-amps only have 20~30dB available small signal gain at 200KHz, so not much excess gain to drive down the open loop output impedance, and much less available gain for larger signal swings. Also, the closed loop output impedance will "look" inductive with rising impedance vs frequency, so resonances are likely with capacitive loads.

Rail to rail outputs can be even stranger.
Flat closed loop output impedance up to 1~100kHz (depending on type) due to the open loop impedance dip and then falling at 20 and 40 dB per decade.

[Zero999]

If a much lower output voltage is acceptable, then you could add an attenuator. This can be as simple as connecting a low value resistor to the output. For example, if you connect a 1R resistor, the output will go down by 1/51, but the output resistance will be the equivalent of a 1R resistor and 50R resistor connected in parallel.

[Terry Bites]

Have a look at. adsl/ dsl line drivers have <0.1R output impedances at 200kKz.

eg EL1511 OPA2670 THS6182 LT6300 etc