EMRFD Message Archive 3583

Message Date From Subject
3583 2009-09-29 12:31:45 joop_l RF current transformer
The last days I have built some current transformers. The aim is to be able to measure crystal current in an oscillator. Or perhaps look at the collector current of some transistor. They were wound 1:10 and 1:5 to hook them up to a 50 Ohm scope input.

But I am struggling a bit to find the optimal transformer. I measured different ones as a voltage transformer using two back to back.

The first attempt was a 1:10 transformer on a 4A11 13mm toroid (similar to #43 material). Upper frequency was 3.3MHz (-1dB) and 7.2MHz (-3dB).

Second came an unknown binocular core of #61 type material, also 1:10. Frequency range was 0.99 - 7.5MHz (-1dB) and 0.43 - 16.6MHz (-3dB).

The last one was a rewound MiniCircuits TC6-1-75 binocular core as 1:5 transformer. Frequency range was 0.125 - 31.5MHz (-1dB) and 0.053 - 69MHz (-3dB). This core looks promising, but is so tiny I wonder if I could fit a 1:10 set of windings. Which would be nice because of the lower input impedance of the current probe (1:5 gives 2 Ohm whereas 1:10 gives 0.5 Ohm).

Would it be possible to improve on the above or is this sort of what one can expect?

Joop - pe1cqp
3584 2009-09-29 18:02:39 Phil Sittner Re: RF current transformer

Jim Williams of Linear Technology did a video on this very subject (measuring crystal currents) and it's presented as a news article on the company website. You might take a look and see how he approached the problem. As I recall he was concerned with clock accuracy and how excessive currents affected the accuracy.

Best of luck

Phil, kd6rm
----- Original Message -----
3585 2009-09-29 18:45:04 w7zoi Re: RF current transformer
Hi Joop and gang,

I recall having done this measurement many years ago with a 5th overtone crystal at either 100 or 125 MHz. While I should have done the experiment you did with the back to back transformers, I don't recall that I did. Rather, I just attached the output of a generator to a 50 Ohm resistor through a piece of coax. I inserted the single tune winding in series with the crystal and saw results that made sense. So I then hung the core on the crystal lead and did my my measurement. I recall that the crystal current was well above the level that folks told us was appropriate for good aging. But it provided excellent phase noise. Eventually we turned the oscillator into a VCXO and phase locked it to a lower frequency, very stable source, which gave us the best of both worlds.

Another thought would be to use an AD8307. Drop it down on a small board with a 1 Ohm or so input resistor that is differentially attached to the log amp input port. Then put the board containing the chip next to the crystal with the 1 Ohm in series. If the levels are too high, you could put attenuation at the input to the 8307. This is akin to a return loss bridge design that floated around the internet where the bridge circuitry and a 50 Ohm termination resistance were tied directly to an AD8307. I suspect it was a great performer. The hard part of any return loss bridge is the transformer. As I recall, that bridge design was by a W/K/N 9, but I'm not sure of that.

Anyway, have fun. As usual, it appears that you are
3587 2009-09-30 13:06:24 joop_l Re: RF current transformer
3588 2009-09-30 15:06:24 joop_l Re: RF current transformer
Thank you for the supporting words.

The 50 Ohm series resistor hooked up to a generator seems valid too. I intended to do a double check with that when I would settle on the final transformer type.
In Jim Williams video he uses a similar method to check/calibrate his 32K crystal measurment setup. Especially important since he uses the transformer below its flat response range.

But operating at 100 to 125 MHz would already be dreaming with the type of transformers I made up to now.
The last thing I did was somewhat force a 0.3mm wire through the little space left on the tiny MiniCircuits transformer. So I kept the 2x7 factory windings this time. The 0.3mm wire is thicker than the 0.15mm that was used before. It seems this helps to reduce the transformer loss. With this I have a 1:7 transformer, which would give about 1 Ohm impedance and 7mV/mA on the scope. This is a bit awkward for easy calculus, but I guess accuracy and frequency range matters more.
At this point it has a range of 0.082 - 26.8MHz (-1dB) and 0.037 - 58.9MHz (-3dB). But from the former 1:5 version loss has gone down a bit from 0.5dB to 0.37dB for the 1:7. Probably due to the thicker (and a bit shorter) single primary wire.
Too bad it was not a 2x5 winding core. Then I could have used both of the winding pairs. A 1:14 would be possible of course, but that might be going a bit too far. Still I guess it is a relatively simple thing to test.

After looking at the 4A11 material specs that were used in the toroid version, it seems this is probably not the right material to start with. It is already pretty lossy at 10MHz. Perhaps I should try a better suited toroid. Especially when looking at the Tektronix CTx specs. Those seem to be based on toroids.

Your log-amp idea might be interesting for the higher frequencies. A challenge might be to keep the response curve flat at these -50 to -60 dBm levels. Or perhaps creating a reference chart. The Tek probe impedances seem to go up in impedance around 100MHz. So having 1 Ohm pure resistive might be pretty useful actually. I still have an AD8307 on the shelf that needs a proper project. This might just be the one.
Thanks for the idea. For crystal current it seems well suited. For seeing the shape of a collector current the transformer still seems the way to go.

3590 2009-10-04 08:24:19 joop_l Re: RF current transformer
Hi guys,

For those who are interested, I uploaded an article about the progress of my RF current probe. It turned out better then I hoped.
The article is in my files section:

Joop - pe1cqp
3603 2009-10-06 01:43:22 Kerry Re: RF current transformer
W7ZOI wrote;

"...... a return loss bridge design that floated around the internet where the bridge circuitry and a 50 Ohm termination resistance were tied directly to an AD8307. I suspect it was a great performer".

It certainly is a good performer; the performance is only limited by the AD8307.