Home


FREQUENCY RESPONSE OF IF TRANSFORMERS.

Using Function Generator and DSO to Display the Frequency Response of IF transformers and Filters.

Almost from the invention of radio, engineers have wanted an easy way to see the frequency response of tuned amplifiers. If you attended any sort of electronics educational institution prior to about 1985 you probably performed a rather tedious lab experiment in which you set a signal generator to a frequency and wrote down the frequency and voltage in a chart and then moved on to the next frequency to obtain another data point. Then you took the data sheet back to your dorm and sat down with a piece of graph paper and a French curve to graph the data. If you didn't make any procedural errors such as reading the wrong scale on the voltmeter you would get a nice looking curve.

While automatic frequency response plotters became available in the 1960s their cost was prohibitive for all but the largest universities and they certainly would never allowed mere students to use them in lab experiments. These devices were so far beyond the means of hams and other hobbyists that we couldn't even afford to read the catalog.

The development of two products changed all of that. They were the digital sweep function generator and the digital storage oscilloscope. While the introductory price of these items was still in 5 figures the prices slowly came down through the 1990s and the first decade of the 21st century. While there are Chinese made versions of these instruments available for around 50 dollars, as we will see, they are so lacking in features as to be next to useless.

The General Setup.

Figure 1 General Setup for Sweeping and IF Transformer.

For a Verbal Description click here.

Don't Sweep It Too Fast.

The maximum sweep rate is limited by the highest Q element inside the filter. When the frequency sweeps across the peak of a high Q element the high Q means that the ringing of the filter will come to the correct amplitude slowly. It's sort of like a flywheel. It takes a finite amount of time to get a flywheel spinning and it will continue to spin for a while after the driving source is removed. It's the same for a tuned circuit. If you instantaneously apply a signal to an IF transformer the amplitude of the signal on the secondary will take a finite amount of time to rise to its maximum value. Similarly, when the input signal is shut off the transformer's tuned circuits will ring for a little while after the signal has gone. The effects of this are that the peak will be shifted to the right of its actual frequency and the skirts of the curve will appear to rise and fall more slowly than they actually do.

For the IF transformers in an AM broadcast receiver this time is quite fast. If it were too slow the higher frequencies in the modulation would be lost. Sweep rates in the low to mid audio band can be used for these transformers. When Collins mechanical filters or crystal filters need to be swept the rate must be kept below 10 sweeps per second. This is too low to be easily used with an analog oscilloscope. The response time of a filter is determined by its overall bandwidth but when it is swept the Q of individual resonators in the filter determines how it responds to the sweeping frequency.

All three oscilloscope screenshots below were made using a Siglent SDG 5082 generator and a Siglent SDS 2202X DSO. The frequency scale is 10 kHz/div.

Figure 2 Sweep Rate of 1 MHz/sec.

Figure 3 Sweep Rate of 10 MHz/sec.

Figure 4 Sweep Rate of 100 MHz/sec.

The rate of 1 MHz/sec seems to be good. Rates lower than this give no improvement in the display. At first glance the rate of 10 MHz/sec looks as good but look closely at the rounded top of the yellow trace. There is just a little bit of sag on the right side as compared to the 1 MHz/sec rate. The 100 MHz/sec rate looks really bad. Obviously this rate is beyond the limit.

Setting Up For Sweeping, Practical Considerations.

The times 10 probes that go with any modern oscilloscope have low capacitance but it is still enough to detune an IF transformer. After you make all of the connections the alignment must be retouched a bit. Don't forget to ground the shield can of the transformer. Turn off the sweep function and set the generator to 455 kHz. Carefully adjust both tuning adjustments for maximum deflection on the scope. Adjust the two alternately to be sure to account for any interaction between the two adjustments.

I have collected several function generators, I suppose for situations like this. I will now go through them giving detailed setup instructions and showing the results. Some are better than others. I will start with the worst and work up to the best. The generators are,

  • FeelTech FY3200S. eBay, $50.
  • Siglent SDG 810. Circuit Specialists, $282.
  • Siglent SDG 1025. Circuit Specialists, $319.
  • Siglent SDG 5082. Circuit Specialists, $599.

Note: The last two digits of Siglent model numbers are the upper frequency limit of the generator in MHz.

Oscilloscope Setup, For all Generators.

Y axis, Ch1 and Ch2.

Vertical Range, 500 mV/div.
Vertical Position, Center of Screen.
Input selection, DC or AC.

Trigger.

Source, Ext/5.
Mode, Normal.
Level, Zero.
Slope, Positive.
Coupling, DC.

Sweep.

Time, Depends on generator settings.
Position, Center of screen.

FeelTech FY3200S

Below you will find setup instructions in English.

  1. Press the Ch2 button and use the left and right arrow keys and the knob to set the frequency to 0.00096kHz.
  2. Press the WAVE button once and the word SINE in the display will change to "SQUAR" (square).
  3. Press the PARAM button twice and change the AMPL (amplitude) to 5.0 volts.
  4. Press the Ch1 button.
  5. Press the SWEP (sweep) button twice and set the BEGIN FREQ (beginning frequency) to 405.00000 kHz.
  6. Press the SWEP button again and set the END FREQ (end frequency) to 505.00000kHz.
  7. Press the SWEP button again and set the TIME to 1s (1 second).
  8. Press the SWEP button again and the word STOP will appear in the low right of the display.
  9. Press in on the knob and the word STOP will change to RUN.

If you have connected the IF transformer correctly and set up your scope right you will see this pattern on the screen.

Figure 5 Sweep of IF Transformer With eBay Generator.

The first thing you will notice is the quantitization of the trace. This is unavoidable in digital devices however because we are analog creachers digital equipment for our use has been designed with small enough steps so we perceive it as if it were analog. The frequency steps in this generator are so large as to make it little more than a toy.

The trace of the frequency response will not appear in the center of the screen. If it does, go out and buy a power ball ticket. While the frequency of the square wave on channel 2 is the same as the sweep repetition frequency they are not locked together in phase. Their relative phase depends on when you start the sweep. You can repeatedly press the knob to stop and start the sweep until you get it where you want it.

The consequence of this is that the screen has no absolute calibration. If you pre tuned the transformer to 455kHz as described above you can have some confidence that the peak of the response is very close to 455 kHz. The relative calibration of 10 kHz per division is called into question by the need to set the frequency of the synchronizing wave on channel 2 to 0.96 Hz. This does not give me high confidence in the relative calibration either. The extra 4% might be accounted for as retrace time in an analog generator, a digital frequency generator should not require very much time to reset to the starting frequency. The scope picture below puts an end to all of this speculation.

Figure 6 Picture Showing Error in Relative Frequency Calibration.

The discontinuities in the purple trace at about 2 and a half horizontal divisions and exactly 13 divisions indicate the retrace points. The narrow spike makes them easy to find. The spike on the right exactly covers up the graticule line and the one on the left falls almost 10.4 divisions from it. And there's your 4% error. Looks like it's true, you get what you pay for.

Siglent SDG 810.

This procedure will be the same for any of the 800 series. The least expensive is the SDG 805 at $248 from Circuit Specialists.

Figure 7 Picture of Siglent SDG 810 Sweep Function Generator.
Screen contains a mistake. Don't use it for setup.

To make this generator work well you need a second function generator. Here is how to hook them up.

Figure 8 Using Second Function Generator To Trigger Sweep Gen and Scope.

For a verbal description click here

The second generator could be the eBay job or something home brew. All it needs to do is generate a square wave in the frequency range of 0.1 to 10 Hz and have either a digital user interface or a fine Vernier for setting the frequency. I am going to use the one built into my Siglent scope. I have never used it in anger before. I actually ordered the model without the function generator but due to a backordering mix up I was sent the one with the generator. Setup is easier if you can enter or set the period of the output wave rather than the frequency.

The sweep time and the period of the triggering wave must be the same. The sweep time is determined by the sweep time you set on your scope and the number of horizontal divisions on your scope's screen. If you haven't noticed my scope has 14 horizontal divisions. I have seen DSOs with 12 divisions as well as the traditional 10. The sweep time on the sweep generator is equal to the time it takes for your scope to complete one horizontal sweep for the time/div setting you have chosen to use. I am using a setting of 10 ms/div which gives a total sweep time of 140 ms.

I will not give button by button operating instructions because I assume you own one of these and are familiar with it. Also the Siglent gear comes with manuals written in good English. A picture of the screen will go a long way.

Figure 9 Close up of SDG 810 screen showing set up.

If you have the output impedance set to 50 ohms as I do you get twice the output value you set. If you have it set to Hi Z you will get what you set.

Scope Sweep, 10 ms/div.
SDG 810 settings.
Amplitude (actual output) 10 Vpp.
Sweep Time 140 ms. Set correctly for your scope.
Center Frequency, 455 kHz.
Frequency Span, 140 kHz. (Adjust for your scope to give a scale of 10 kHz/div). Source, External.

Figure 10 Resulting Screen for SDG 810 and additional function generator.

Siglent SDG 1025, SDG 5082, and Other Two Channel Function Generators.

Once again I won't go into the details of setting up each generator. If you own one of the 1000 or 5000 series you will already know how to operate it. If you own a two channel generator of a different brand I don't know how to operate it so I can't tell you how.

For these two generators the scope sweep remains on10 ms/div but you must change the trigger slope to negative.

Scope Settings.
Scope Sweep, 10 ms/div.
Trigger slope, Negative.
SDG 1000 or 5000 Settings.
Channel 1.
Wave, Sine.
Amplitude (actual output) 10 Vpp.
Sweep Time 140 ms. Set correctly for your scope.
Center Frequency, 455 kHz.
Frequency Span, 140 kHz. (Adjust for your scope to give a scale of 10 kHz/div). Channel 2.
Wave, Square.
Amplitude 5 Vpp.
Period, 140 ms. Set correctly for your scope.

Both of these generators use the setup shown in figure 1. The difference between these two generators and the one from eBay is that the phase relationship between the sweep control and the channel 2 square wave output is fixed and will be the same every time you set it up.

Figure 11 Picture Showing Frequency Sweep Retrace Point.

The two models of generator are not consistent with each other but each one is consistent with itself. This is no problem because the horizontal position control can be set for very good frequency accuracy.

In the picture above look at the purple trace between the vertical grid lines one and two. There is a discontinuity and a tiny spike where the frequency sweep arrived at the maximum value and was reset to the minimum frequency to begin another sweep. The retrace seems to take place in the width of one pixel. The sweep retrace on your DSO is just as fast. As the position control is moved to the left the retrace point will move off the left edge of the screen and will appear at the right. Depending on whether your scope has over scan or not you may be able to see the retrace artifact at both edges of the screen or you may not be able to see it at all. The point is that when the retrace occurs on the outside grid lines the calibration places 455 kHz at the exact center of the screen and each horizontal division is 10 kHz.

Conclusion.

This article can hardly be considered a comprehensive survey of sweep generators. I suppose that some of you think I am rich but I'm not that rich. If you have a sweep generator and you can't make it perform as shown here I may or may not be able to help you. That's the way it goes with long distance help. But if you want help contact me through the Fun With Tubes email list and I'll see what I can do.


This page last updated Thursday, October 25, 2018.


Home