Si8635 RF Isolator Output

[Sparky49]

Hi

I've built a circuit using a Si8635 RF Isolator. The datasheet (https://www.silabs.com/documents/public/data-sheets/si863x-datasheet.pdf?fbclid=IwAR3oMAyNjynhvtSsHQ5jLYvBNxjVzWNcFvqIUMIj__LGdztuyqhBs6HhPPo) says that the output high Voltage should be ~5V, and the output Z is 50 ohm.

However, when I hook it up to my scope (HP54542A) I measure ~5V only when in high Z input, and ~2.5V in 50 ohm mode. This sounds like they have spec'd the 5V high on the output of the source and not including any series impedance... I have measured all supply voltages etc and they are fine. The data itself comes out ok, it's just the amplitude is very wrong.

My question is, where have I misinterpreted the datasheet? I imagined that I could simply connect the output to my 50 ohm input on the scope and see the 5V high voltage.

Thanks in advance.

[Zero999]

No, 50Ω means its output stage has an output impedance of 50Ω. This means if the output is connected to a 50Ω load, the voltage will fall by half, as you've created a potential divider, with the source and load resistance.

In other words, this is exactly what you should expect to happen.

[SiliconWizard]

Second that.
Its outputs have an output impedance of 50 ohm. If you load them with 50 ohm, the voltage output will be divided by 2 obviously. To compensate, many typical signal generators will multiply their output amplitude by 2 when you set them for a 50 ohm load. This is not the point of this IC, and there would be no way it could with a 5V power supply (unless it had some kind of internal step up or something, which is absolutely not the point of these isolators).

Note: these isolators are "high-speed" (150Mbps) but they are not "RF" isolators per se. They are not meant to be loaded by 50 ohm loads, nor meant to directly drive coax. cables for instance (if that was the application you had in mind).

[Zero999]

A quick schematic showing the 50Ω resistance of the output stage, connected to the load.

If a 1V drop is acceptable, connect a complementary emitter follower to the output.

[Sparky49]

So when I read the sepc for, say, a function generator it might be 10Vpp max, 50 ohm output. But that is spec'd for the output from the device, after the resistor in the output stage. Here the spec is 5V output high, but before the internal 50 ohm resistor. Why the difference in spec? Why would you spec the output voltage for a system that isn't terminated into the same impedance?

Thanks again.

[Sparky49]

Second that.
Note: these isolators are "high-speed" (150Mbps) but they are not "RF" isolators per se. They are not meant to be loaded by 50 ohm loads, nor meant to directly drive coax. cables for instance (if that was the application you had in mind).

Interesting thanks. I'm using them to isolate a transmission line in a noisy environment from the control devices. What should I be using to drive the coax (3-4 meters, up to 100Mbps)?

Cheers.

[SiliconWizard]

Well, I don't know exactly what your project is and whether using coax cables is mandatory or not.

If you're free to use whatever solution to transmit a 100Mbps isolated digital signal over 3-4m, I would suggest using LVDS transceivers as a front-end to/from the Si8635 isolators, and then use a cable with twisted pairs. If you have several digital signals to transmit, using an ethernet cable would then be handy and a relatively low-cost solution to this, as it has 4 twisted pairs.

[Yansi]

Have a look at ICs for driving coax cables for SDI video.  These can blast some hundred(s) of mbps down a coax line.

[Zero999]

So when I read the sepc for, say, a function generator it might be 10Vpp max, 50 ohm output. But that is spec'd for the output from the device, after the resistor in the output stage. Here the spec is 5V output high, but before the internal 50 ohm resistor. Why the difference in spec? Why would you spec the output voltage for a system that isn't terminated into the same impedance?

Thanks again.

Unless otherwise stated, the output voltage is specified with no load. The output impedance is the internal resistance of the output stage. In some devices, it's not consistent and can change with the power supply voltage (look at the specification for a CMOS logic ICs, such as the CD4011 or 74HC00) or frequency, as is the case with an audio amplifier.

There aren't different specifications and there is no requirement that the output is terminated into the same impedance, unless stated otherwise. In an ideal world, the terminating impedance should be as high as possible to avoid voltage drops. The output voltage can be calculated using Ohm's law.

Work out the current:
I_OUT =  V_SUPPLY /(R_OUT + R_LOAD).

Now V_OUT is simply equal to the supply voltage minus the voltage dropped inside the output resistance.
V_OUT = V_SUPPLY - R_OUT * I_OUT.

Signal generators typically have a 50Ω or greater output impedance to protect the output from damage. It limits the short circuit current and any voltage spikes when driving an unterminated piece of co-axial cable.

[Sparky49]

Thanks everyone, especially Zero999 - I will bear in mind the spec for no load condition in the future.